Open System Services Management and Operations Guide Abstract This guide describes how to manage and operate the HP NonStop™ operating system Open System Services (OSS) environment. Product Version OSS Monitor T8622H02, T8622G12 Supported Release Version Updates (RVUs) This guide supports G06.27 and all subsequent G-series RVUs and H06.06 and all subsequent H-series RVUs until otherwise indicated by its replacement publication.
Document History Part Number Product Version Published 527191-002 OSS Monitor T8622H01, T8622G11 July 2005 527191-003 OSS Monitor T8622H01, T8622G12 September 2005 527191-004 OSS Monitor T8622H02, T8622G12 May 2006
Open System Services Management and Operations Guide Glossary Index What’s New in This Guide xiii Guide Information xiii New and Changed Information Figures Tables xiii About This Guide xv What This Guide Is About xv Who Should Read This Guide? xv What This Guide Does Not Cover xv What Is in This Guide? xvi Related Reading, Training, and Services Unsupported Utilities xix Acknowledgment xx Notation Conventions xx xvii 1.
2. Operating the OSS Environment Contents 2.
4. Managing Servers Contents 4.
4. Managing Servers (continued) Contents 4. Managing Servers (continued) Reconfiguring a Server 4-46 Reconfiguring an OSS Name Server 4-46 Reconfiguring the OSS Message-Queue Server 4-47 Reconfiguring the OSS Sockets Local Server 4-48 Reconfiguring a Network Services Server 4-49 Removing a Server 4-49 Removing an OSS Name Server 4-49 Removing a Network Services Server 4-50 Troubleshooting a Server 4-50 5.
5. Managing Filesets (continued) Contents 5. Managing Filesets (continued) Renaming a Fileset 5-34 Updating Existing Fileset Configurations 5-35 Removing Older Configuration Files 5-35 Moving a Directory Hierarchy to Its Own Fileset 5-36 Cleaning Up a Fileset 5-37 Troubleshooting Filesets 5-39 Managing and Repairing Fileset Catalog Files 5-40 Upgrading OSS Catalog Files 5-40 Moving and Removing OSS Catalog Files 5-41 6.
. Managing Security Contents 8.
10. Managing OSS Devices Contents 10. Managing OSS Devices The Scope of OSS Device Management 10-1 Device Access 10-1 Managing Printers in the OSS Environment 10-1 Specifying a Default Printer 10-2 Using the /etc/printcap or printcap File 10-3 11. Managing Problems Problem-Reporting Procedures 11-1 Gathering Version Information About OSS Files 11-1 12.
A. Messages Contents A. Messages OSS EasySetup Utility Messages A-2 CVT Messages A-3 CVT Warning Message A-3 CVT Error Messages A-3 FSCK Messages A-6 FSCK Consistent-Fileset Messages A-7 FSCK Inconsistency and Error Messages A-7 OSS Monitor Messages A-27 Unnumbered Messages A-27 Numbered Messages A-35 OSSTTY Subsystem Messages A-58 Startup Messages A-58 B. Manually Setting Up an OSS Environment C.
Glossary Contents Glossary Index Figures Figure 1-1. Figure 1-2. Figure 1-3. Figure 1-4. Figure 1-5. Figure 1-6. Figure 1-7. Figure 2-1. Figure 2-2. Figure 2-3. Figure 2-4. Figure 2-5. Figure 3-1. Figure 3-2. Figure 3-3. Figure 3-4. Figure 4-1. Figure 4-2. Figure 4-3. Figure 4-4. Figure 4-5. Figure 4-6. Figure 5-1. Figure 5-2. Figure 5-3. Figure 5-4. Figure 6-1. Figure 6-2. Figure 6-3.
Figures (continued) Contents Figures (continued) Figure 6-4. Figure 6-5. Figure 6-6. Figure 7-1. Figure 7-2. Figure 7-3. Figure 8-1. Figure 8-2. Figure 9-1. Figure 10-1. Figure 12-1. Figure 12-2. Figure 12-3. Figure 12-4. Figure 12-5. Figure C-1. Figure C-2. Figure C-3. Figure C-4.
Tables (continued) Contents Tables (continued) Table E-2. Table E-3. Size Limits for Files Created on H06.
Contents Open System Services Management and Operations Guide —527191-004 xii
What’s New in This Guide Guide Information Open System Services Management and Operations Guide Abstract This guide describes how to manage and operate the HP NonStop™ operating system Open System Services (OSS) environment. Product Version OSS Monitor T8622H02, T8622G12 Supported Release Version Updates (RVUs) This guide supports G06.27 and all subsequent G-series RVUs and H06.06 and all subsequent H-series RVUs until otherwise indicated by its replacement publication.
What’s New in This Guide • • • New and Changed Information In Appendix A, Messages, messages that are described in the Operator Messages Manual were removed. For descriptions of these messages, see the Operator Messages Manual. In Appendix D, Falling Back to a Previous Release Version Update, information about falling back to release version updates (RVUs) prior to G06.12 has been removed. In Appendix E, Environment Limits, a description of file size formats and file size limits has been added.
About This Guide The HP NonStop operating system Open System Services (OSS) environment enables users on HP NonStop servers to integrate an operating system similar to the UNIX operating system into their work environment. You manage and operate the OSS environment primarily from the Guardian environment. What This Guide Is About This guide describes how to manage and operate the OSS environment. It describes only what is unique to the OSS environment, its management, and its operation.
What Is in This Guide? About This Guide What Is in This Guide? This guide contains information and procedures for managing and operating the OSS environment. It is divided into the following sections: • • • • • • • • • • • • Section 1, Introducing Open System Services, presents an overview of how to manage and operate the OSS environment. It includes a task table that refers you to the appropriate documentation for specific kinds of tasks.
Related Reading, Training, and Services About This Guide • • • • • Appendix A, Messages, describes OSS EasySetup, CVT, FSCK, OSS messagequeue server, OSS Monitor, and OSS subsystem messages. Appendix B, Manually Setting Up an OSS Environment, summarizes the procedures to configure and start a new OSS environment. Appendix C, OSS Management Utilities, describes the OSSTTY and OSS EasySetup product utilities and files.
Additional Reading About This Guide • • • • • SCF Reference Manual for G-Series RVUs SCF Reference Manual for H-Series RVUs Security Management Guide Software Internationalization Guide The courses NonStop S-Series Configuration and Change Management and NonStop S-Series Production Management Additional Reading Problem reporting often involves using the Event Management Service (EMS) and the Guardian VPROC utility.
Unsupported Utilities About This Guide • • • • Spooler Utilities Reference Manual System Generation Manual for G-Series RVUs Telserv Manual WAN Subsystem Configuration and Management Manual UNIX System Administration • • • Frisch, Æleen. Essential System Administration, Second Edition. O’Reilly & Associates, Inc. Sebastopol: 1995 Nemeth, Evi, Garth Snyder, Scott Seebass, and Trent R. Hein. UNIX® System Administration Handbook, Second Edition. Prentice-Hall.
Acknowledgment About This Guide Acknowledgment The Portable Archive Interchange (pax) utility software was developed by Mark H. Colburn and is sponsored by the USENIX Association. Copyright (c) 1989 Mark H. Colburn. All rights reserved. The pax utility, distributed to HP free of charge, is used to support the tar, cpio, and pax user interfaces furnished with the OSS environment and is documented in this guide and the Open System Services Shell and Utilities Reference Manual.
General Syntax Notation About This Guide italic computer type. Italic computer type letters within text indicate C and Open System Services (OSS) variable items that you supply. Items not enclosed in brackets are required. For example: pathname [ ] Brackets. Brackets enclose optional syntax items. For example: TERM [\system-name.]$terminal-name INT[ERRUPTS] A group of items enclosed in brackets is a list from which you can choose one item or none.
Notation for Messages About This Guide Quotation marks around a symbol such as a bracket or brace indicate the symbol is a required character that you must enter as shown. For example: "[" repetition-constant-list "]" Item Spacing. Spaces shown between items are required unless one of the items is a punctuation symbol such as a parenthesis or a comma. For example: CALL STEPMOM ( process-id ) ; If there is no space between two items, spaces are not permitted.
Change Bar Notation About This Guide A group of items enclosed in brackets is a list of all possible items that can be displayed, of which one or none might actually be displayed. The items in the list might be arranged either vertically, with aligned brackets on each side of the list, or horizontally, enclosed in a pair of brackets and separated by vertical lines. For example: proc-name trapped [ in SQL | in SQL file system ] { } Braces.
About This Guide Change Bar Notation Open System Services Management and Operations Guide— 527191-004 xxiv
1 Introducing Open System Services The Open System Services (OSS) environment provides a user and programming interface similar to that of the UNIX operating system. The OSS environment combines the benefits of the UNIX operating system with the features of the HP NonStop operating system. Open System Services differs from the UNIX operating system in that almost all management and operations activities are performed through Guardian environment commands.
Management Tools Introducing Open System Services Figure 1-1. The Operating System Environments OSS Environment OSS Shell and Utilities Guardian Environment Guardian products, subsystems, and command interpreters OSS Application Program Interface Guardian Application Program Interface OSS File System Guardian File System NonStop Operating System VST011.
Management and Operations Tasks Introducing Open System Services Management and Operations Tasks The only OSS management and operations tasks that must occur within the OSS environment are performing backups, configuring printer aliases, and defining default profiles; the rest can occur from the Guardian environment. Table 1-1 lists common OSS management and operations tasks, along with the location of information on performing those tasks.
Management and Operations Tasks Introducing Open System Services Table 1-1.
OSS File System Concepts Introducing Open System Services Table 1-1.
OSS Files Introducing Open System Services • • • OSS pathnames, which have underlying Guardian filenames. The mapping between OSS pathnames and Guardian filenames is known as filename resolution, and it is done by an OSS name server. The /G directory, which contains filesets for local files in the Guardian namespace, and the /E directory, which contains files on other nodes in the network. The /dev directory being used in a special way. Figure 1-2.
Introducing Open System Services The /G Directory The /G Directory The /G directory provides OSS names for Guardian files on your local NonStop S-series or NonStop NS-series node. Each Guardian filename has a corresponding OSS name of the form /G/volume/subvolume/fileID, where volume, subvolume, and fileID are case-insensitive; for example, $SYSTEM.SYS00.CONFLIST becomes /G/system/sys00/conflist. The /G directory itself is reserved for HP use. You cannot put anything in this directory.
The /dev Directory Introducing Open System Services Figure 1-3 shows how the /E directory on your local node allows remote file access through an Expand network. Figure 1-3. Pathname Resolution for Remote File Access Through the Guardian Expand Network Local Node \NODE1 OSS Environment Guardian Environment Remote Node \NODE2 OSS Environment Guardian Environment OSS pathname OSS pathname Guardian filename /E /node2 /usr /donl /test /usr /donl /test \NODE2 .$VOL2 .ZYQ00011 .
Introducing Open System Services Components to Be Managed Components to Be Managed You need to manage the software described in the following subsections: • • • • Input/Output Utilities on page 1-9 OSS Security on page 1-9 OSS File-System Components on page 1-10 Interprocess Communication Facilities on page 1-11 Input/Output Utilities Input/output utilities include: • • OSSTTY Servers on page 1-9 Terminal Helper Servers on page 1-9 OSSTTY Servers Guardian administrative applications or OSS applicati
OSS File-System Components Introducing Open System Services OSS File-System Components The software components that work together to provide access to an OSS file are shown in Figure 1-4 on page 1-10. Many of these components are configured and managed using the interfaces described in this guide. Figure 1-4.
Introducing Open System Services Interprocess Communication Facilities Open System Services System Calls Reference Manual and in the Open System Services Library Calls Reference Manual. When the application opens an OSS file, an OSS name server locates the correct file within the correct fileset, based upon configuration information you have supplied to the OSS Monitor for that fileset’s catalog. An OSS file manager process named $ZFMnn runs in each processor (nn indicates the processor number).
Introducing Open System Services Interprocess Communication Facilities OSS IPC facilities include the facilities described in the following subsections: • • • • OSS Shared Memory and Semaphores on page 1-14 OSS Message Queues on page 1-15 Pipes and FIFOs on page 1-15 OSS Sockets on page 1-15 Figure 1-5 on page 1-13 shows these facilities with their approximate equivalents in the Guardian environment.
Interprocess Communication Facilities Introducing Open System Services Figure 1-5.
Interprocess Communication Facilities Introducing Open System Services Figure 1-5.
Introducing Open System Services Interprocess Communication Facilities OSS shell commands allow you to manage OSS message queues, OSS shared memory, and OSS semaphores; these commands are discussed further in Section 2, Operating the OSS Environment. OSS Message Queues Message queues are linked lists of messages used by programmers to pass data from one process to another. In the OSS environment, the OSS message-queue server process named $ZMSGQ manages the message queues.
Introducing Open System Services • Interprocess Communication Facilities Multiple transport-provider processes can be used to create separately addressed IP subnetworks within a node When AF_INET sockets are used with NonStop Parallel Library TCP/IP: • • • No transport-agent process or transport-provider process is involved in data routing. A transport-agent process must still be started in each processor to initialize the OSS sockets interface for that processor.
Introducing Open System Services • • Interprocess Communication Facilities HP provides the OSS sockets local server, $ZPLS, as a transport-provider process for OSS AF_UNIX sockets. You manage the OSS sockets local server through the SCF module of the OSS Monitor, as described in Section 4, Managing Servers. HP provides processes for each of its TCP/IP implementations as transportprovider processes for OSS and Guardian AF_INET and AF_INET6 sockets.
Interprocess Communication Facilities Introducing Open System Services Figure 1-6. OSS AF_INET Sockets Servers for NonStop TCP/IP Node A Processor 0 Processor 1 Process 1 Process 2 $ZTA00 $ZTA01 $ZTC0 Node B Processor 0 Processor 1 $ZTC0 $ZTC1 $ZTA00 $ZTA01 Process 3 Process 4 VST012.VSD The application processes (Processes 1 through 4) communicate with each other through the transport-provider ($ZTCn) processes, with data being carried back and forth by the transport-agent processes.
Introducing Open System Services Interprocess Communication Facilities For example, given the situation pictured in Figure 1-6 on page 1-18, suppose Process 2 sends a message to Process 4. The following occurs: 1. The message is forwarded by $ZTA01 (the transport-agent process on processor 1 of Node A, where Process 2 is running) to $ZTC0 (the transportprovider process that supports the IP address used by Process 2). 2.
Interprocess Communication Facilities Introducing Open System Services Similarly, when Process 3 sends a message to Process 7, the message is forwarded to $ZTA01, then to $ZPLS, then to $ZTA02, and then to Process 7. Figure 1-7. OSS AF_UNIX Sockets Servers Node A, Processor 2 Node A, Processor 1 Process 3 Process 4 $ZTA01 $ZPNS Process 7 $ZTA02 $ZPLS VST019.
2 Operating the OSS Environment You operate the Open System Services environment from the Guardian environment. As a system operator, when you are operating the OSS environment, there is little you need to do differently from operating the Guardian environment. This section describes what you do to operate the OSS environment that is unique to the OSS environment.
Operating the OSS Environment Automatic Startup Service provides the most availability for OSS systems and requires the least user intervention if a failure occurs. • • Using the STARTOSS utility if your system was initially configured by using the OSSSETUP utility, and the OSSINFIL file has been properly maintained. See STARTOSS Utility on page C-14 for more information. Using individual commands described under Managing the OSS Subsystem on page 2-6 and in other sections of this guide.
Operating the OSS Environment Possible Ways to Stop the OSS File System Possible Ways to Stop the OSS File System You can stop the OSS file system by unmounting the root fileset, which effectively stops the OSS environment. This step can be performed either by: • • Using the STOPOSS utility. STOPOSS might also stop processes you want to continue running; see the STOPOSS Utility on page C-16 for considerations when using this command.
Operating the OSS Environment Manually Stopping the OSS File System and the OSS Environment however, the most commonly used signal can be sent by entering the OSS shell command: kill PID1 PID2 PID3 ... where PID1, PID2, and PID3 are OSS process IDs displayed by the ps command.
Operating the OSS Environment Manually Stopping the OSS File System and the OSS Environment Sending Warnings to Users Use the OSS shell wall command or an OSS shell script to warn users of the OSS shell about a shutdown. Follow your site’s broadcast message procedures to warn users of Guardian environment processes that might be using OSS files.
Operating the OSS Environment Manually Restarting the OSS File System and the OSS Environment For information about shell scripts, see the Open System Services User’s Guide. For information about the wall command, see the wall(1) reference page either online or in the Open System Services Shell and Utilities Reference Manual. You should also follow your site’s broadcast message procedures to warn users who log in through TACL to use OSS files from Guardian environment processes.
Operating the OSS Environment Starting the OSS Monitor messages recorded using the Event Management Service (EMS) through either of the following: • • The OSS shell logger command Program calls that use the syslog() function and related functions For more information about the logger command, see the logger(1) reference page either online or in the Open System Services Shell and Utilities Reference Manual.
Operating the OSS Environment Starting the OSS Monitor OSSSETUP utility, you can start $ZSMP by using the STARTOSS utility; see STARTOSS Utility on page C-14 for more information. Alternatively, you can perform these actions yourself. Starting the OSS Monitor as a Normal Process 1. Log in as the super ID. 2.
Operating the OSS Environment Starting the OSS Monitor Starting the OSS Monitor as a Persistent Process 1. At a TACL prompt, enter: SCF ASSUME PROCESS $ZZKRN 2. Start the $ZSMP process as a generic process so that the OSS Monitor can communicate with it without a delay during restart. a. At an SCF prompt, enter the following to add $ZSMP to the NonStop Kernel subsystem configuration as a generic process: ADD PROCESS $ZZKRN.#ZSMP, & PROGRAM $SYSTEM.SYSTEM.
Operating the OSS Environment Starting the OSS Monitor 4. At an SCF prompt, to start the OSS Monitor, enter: START PROCESS $ZZKRN.#ZPMON 5. At an SCF prompt, to verify that the OSS Monitor is running, enter: STATUS PROCESS $ZZKRN.#ZPMON 6. If the OSS Monitor process is not running, check the Event Management Service (EMS) log for related event messages. See Appendix A, Messages, for explanation of any console messages received and possible corrective actions.
Starting the OSS Monitor Operating the OSS Environment Table 2-1. Currently Used TACL PARAMs for the OSS Monitor (page 1 of 2) PARAM Description AUTOSTART { AUTO | MANUAL } Specifies whether the automatic startup service should be used for any stopped servers or filesets that are configured to use that service. This PARAM can provide the initial configuration value for the AUTOSTART attribute of the subsystem.
Starting the OSS Monitor Operating the OSS Environment Table 2-1. Currently Used TACL PARAMs for the OSS Monitor (page 2 of 2) PARAM Description OSS^NAMESERVER^TIMEOUT seconds Defines the amount of time that the OSS Monitor waits for a response from an OSS name server. This PARAM can provide the initial configuration value for the IOTIMEOUT attribute of the subsystem.
Operating the OSS Environment Starting the OSS Monitor Monitor can no longer start an OSS name server or the FSCK utility. A valid terminal device must exist when the OSS Monitor starts these programs. One possible solution to the problem is to select a permanent terminal such as an asynchronous terminal or a terminal-simulator process (such as $ZHOME) and then specify that device as the OSS Monitor home terminal.
Operating the OSS Environment Starting the OSS Monitor Table 2-3. Wildcard Characters in OSS Monitor Commands Characters Uses and Examples * An asterisk represents any number of characters (including zero) in an entity name. Use this character to save keystrokes when: • • Entering a single name. For example, you can enter RO* for ROOT. Naming many entities at once. For example, /us* matches /user1, /us, and /usr. ? A question mark represents any single character.
Stopping the OSS Monitor Operating the OSS Environment • For example, you can create the alias DF for the STATUS FILESET command to simulate the UNIX df command by doing the following at SCF prompts: 1. Use the SCF ASSUME command to access the OSS Monitor: ASSUME PROCESS $ZPMON 2. Issue the SCF ALIAS command for the desired alias: ALIAS DF STATUS FILESET 3.
Obtaining Information About the OSS Subsystem Operating the OSS Environment For example, if you enter LISTDEV at an SCF prompt, a display similar to that shown in Figure 2-3 would appear.
Operating the OSS Environment Obtaining Information About the OSS Subsystem Listing the Objects Managed by the OSS Monitor When the OSS Monitor is running, you can use the SCF NAMES command to identify the objects managed by the OSS Monitor.
Operating the OSS Environment Changing the OSS Subsystem Configuration Checking the Version of the OSS Monitor When the OSS Monitor is running, you can use the SCF VERSION command to determine the product-version information for the running copy of the OSS Monitor. For example, if you enter the following SCF commands: ASSUME PROCESS $ZPMON VERSION the system displays basic product-version information for the running OSS subsystem.
Operating the OSS Environment Removing the OSS File System FILESET Command on page 12-20 and ALTER SERVER Command on page 12-28. 3. Use the SCF ALTER SUBSYS, AUTOSTART AUTO command to enable the automatic startup service and initiate automatic startup. See ALTER SUBSYS, ALTER MON, and ALTER PROCESS Commands on page 12-34. Removing the OSS File System This action can be performed in one of these ways: • • Using the procedure described in this subsection. Using the OSSREMOV utility.
Operating the OSS Environment Monitoring OSS Processes 12. Remove all previously saved catalogs by entering the following TACL command: RUN $tsvvol.ZOSS.CVT PURGE SERIAL mmmm IN $old_vol.ZX0nnnnn $tsvvol specifies the disk volume of the installation target subvolume (TSV) containing the CVT utility. See Managing and Repairing Fileset Catalog Files on page 5-40 for more information on the CVT utility. mmmm specifies the FSCK serial number of a previous FSCK diagnose or repair operation.
Operating the OSS Environment Monitoring OSS Processes From the OSS Environment Monitoring OSS Processes From the OSS Environment The OSS shell ps command has a -W flag with many options for monitoring processes from the OSS environment.
Operating the OSS Environment Managing OSS Processes Figure 2-5. TACL STATUS, DETAIL Display for an OSS Process . . . System: \NODE1 June 4, 2001 12:09 Pid: 9,326 ($Z0S5) Primary Priority: 160 Wait State: %001 (LREQ) Userid: 255,44 (SUPER.WEBMASTR) Myterm: $ZTN1.#PT74DAB Program File Name: $ROOT.ZYQ00000.Z00116NL Swap File Name: $ROOT.#0 Process Time: 0:0:0.008 Process Creation Time: February 17, 2001 16:53:25.
Operating the OSS Environment Making OSS Application Processes Persistent with the Kernel Subsystem Making OSS Application Processes Persistent with the Kernel Subsystem Just as the OSS Monitor can be made a persistent process using the SCF interface to the NonStop Kernel subsystem, the same SCF interface provides commands to configure, start, and stop OSS application processes that are persistent.
Operating the OSS Environment Making OSS Application Processes Persistent with the Kernel Subsystem >>out redirects standard output to the file out in the current working directory of the OSS process 2>>err redirects standard error to the file err in the current working directory of the OSS process See the sh(1) reference page either online or in the Open System Services Shell and Utilities Reference Manual for more information about OSS shell redirection.
Operating the OSS Environment Making OSS Application Processes Persistent with the Kernel Subsystem consideration means that an OSS persistent process that runs in more than one processor (an OSS persistent process with a valid CPU attribute) must be started from within a script that uses an OSS shell run command with a -name flag of the form: run -name base_name$ZCPU...
Operating the OSS Environment • Making OSS Application Processes Persistent with the Kernel Subsystem When more than 128 characters are required to correctly start an application or script, a script requiring less than 128 characters in its STARTUPMSG attribute can be run to start the command or script that needs to be made persistent.
Operating the OSS Environment Making OSS Application Processes Persistent with the Kernel Subsystem launches the named process $OSS1 that runs the program (-p) /bin/tail as if it were a login shell (-ls); the tail utility uses the file (-f) named log in the home directory of the user with the user ID OSS.APPS. Because of the -ls flag, the tail command inherits any environment variables defined in either /etc/profile or $HOME/.profile for the user ID OSS.APPS.
Operating the OSS Environment Making OSS Application Processes Persistent with the Kernel Subsystem PROGRAM $SYSTEM.SYSTEM.OSH ASSOCPROC $OSS2, STARTUPMSG "-ls -p /bin/sh startmyapp reload" c. Because the configuration uses a STARTMODE of MANUAL, you must enter the following command at an SCF prompt to start the persistent process: START PROCESS $ZZKRN.#OSSAPP These commands configure and start the persistent process object OSSAPP in processor 2, using the named process $OSH2 to launch the OSH utility.
Operating the OSS Environment Making OSS Application Processes Persistent with the Kernel Subsystem To configure and run myapp in all processors as a persistent process that uses the CONFIGURATION value of reload, enter the following commands: a. From a TACL prompt: WHO which produces output that includes the HOMETERM value you need to use, such as: Home terminal: $ZTN0A.#PT4KH30 ... b. Then, at an SCF prompt, enter: ADD PROCESS OSSAPP, NAME $OSH, HOMETERM $ZTN0A.
Operating the OSS Environment Managing OSS Process Scheduling Managing OSS Process Scheduling A frequently used process might not obtain adequate processor time when it runs with a default priority. If that happens, you can assign a nondefault priority to the process. In the Guardian environment, you would write a CMON process to control the scheduling of other processes. In the OSS environment, you can start processes with nondefault priorities using the OSS nice command.
Operating the OSS Environment Managing OSS Process Scheduling Using an OSS Shell Script Unlike a shell alias, the shell script method works for calls to a program from within another shell script. To use an OSS shell script that changes the default priority of a program, follow these steps: Note. Using shell scripts increases system overhead. Be sure that the performance tradeoff is worthwhile before you use this technique. 1.
Operating the OSS Environment Managing OSS Process Scheduling 4. Post a broadcast message to users, using the technique described in Manually Stopping the OSS File System and the OSS Environment on page 2-3. This procedure can be used to alter the scheduling priority of any process, including those released by HP as part of the OSS environment. For example, the following steps decrease the scheduling priority for the c89 compiler by 15: 1.
Operating the OSS Environment Managing OSS Process Processor Use Managing OSS Process Processor Use You can also use OSS shell scripts or aliases to force specific processes to execute in specific processors to distribute the work load within your NonStop S-series or NonStop NS-series node. Note. Using shell scripts increases system overhead. Be sure that the performance tradeoff is worthwhile before you use this technique. To change the default processor for a specific process: 1.
Operating the OSS Environment Managing OSS Interprocess Communication Facilities 3. Change the PATH environment variable in /etc/profile to something similar to the following: export PATH=/usr/local/script:/bin:/bin/unsupported: /usr/ucb:/usr/bin 4. Post a broadcast message to users. Managing OSS Interprocess Communication Facilities The OSS interprocess communication (IPC) facilities require little management.
Using the cron Process Operating the OSS Environment product with the TACL OSH command. The following subsections discuss these options: • • Using the cron Process on page 2-35 Using the NetBatch Product on page 2-38 Using the cron Process The cron process runs shell commands at specified dates and times. The commands that are to be run and the schedule are placed in entries within files in a specific directory. Commands that are to run only once are placed within files in the at queue.
Operating the OSS Environment Using the cron Process /var/adm/cron/cron.deny Lists the user IDs that are denied access to the crontab file. /var/adm/cron/.proto Contains shell commands required to provide the correct shell environment for at and batch jobs. /var/adm/cron/queuedefs Configures the task queues to be used for the at, batch, and cron commands. HP provides sample files for the at.deny, cron.deny, .proto, and queuedefs files.
Operating the OSS Environment Using the cron Process For more information about the crontab command, see the crontab(1) reference page either online or in the Open System Services Shell and Utilities Reference Manual. The at Command The at command runs OSS shell commands at a time you specify. You can use the at command to read from the standard input file or accept as arguments the names of commands to be run and when the commands are to be run.
Operating the OSS Environment Using the NetBatch Product specify one or more user names, only jobs belonging to those users are displayed. If you do not specify any user names, a list of all jobs submitted is displayed. The atq command allows you to sort the output in chronological order based on the time that the at command was issued. You can also specify which queue you want to have printed.
Operating the OSS Environment • Using the NetBatch Product The general form of the TACL OSH command for batch execution of a program in the OSS environment is: OSH <- >>out_file 2>>err_file -p program_path program_args > indicates that the OSS shell appends normal output to the identified file instead of the standard output (stdout) file. out_file is the OSS pathname of the file to receive normal output.
Operating the OSS Environment Using the NetBatch Product specified, the specified pathname must be a Guardian file identifier and the program file must reside in that subvolume. program_args specifies any arguments expected on a command line by the program_path program. Such arguments would include the name of an input file.
Operating the OSS Environment Using the NetBatch Product If the initial working directory is not the current working directory, relative pathnames are resolved from the initial working directory. When the initial working directory is a Guardian subvolume and only a relative pathname is specified, the specified pathname must be a Guardian file identifier and exception output is saved in a file-code-180 file in that subvolume. Site-written shell programs can also be used instead of the default OSS shell.
Operating the OSS Environment Using the NetBatch Product directory is a Guardian subvolume and only a relative pathname is specified, the specified pathname must be a Guardian file identifier and exception output is saved in a file-code-180 file in that subvolume. script_path specifies the OSS pathname for the OSS shell script file to be run. If the initial working directory is not the current working directory, a relative pathname is resolved from the initial working directory.
Operating the OSS Environment Using the NetBatch Product 2. Enter the following at a TACL prompt: BATCHCOM;SUBMIT JOB, IN \NODE.$SYSTEM.SYSTEM.HOURLY, & EVERY 60 * * * * mon-fri As an example of running an OSS shell script periodically, suppose: • • • You want to record OSS process activity every 10 minutes only on weekdays. You have created an OSS shell script file containing the ps command at /script/processes.sh. You want to record normal output in the OSS file /script/processes.
Operating the OSS Environment Using the NetBatch Product Open System Services Management and Operations Guide— 527191-004 2- 44
3 Understanding the OSS File System To manage the Open System Services (OSS) environment effectively, you must understand the OSS file system. The OSS file system works in the same way as a UNIX file system from the point of view of the user. Files in the OSS environment are organized in a hierarchical tree structure. For further information about using the tree structure, see the Open System Services User’s Guide.
OSS Pathnames Understanding the OSS File System OSS pathnames are logical names; they have no connection to storage devices. The relationship of OSS files to disk volumes is illustrated in Figure 3-1. Figure 3-1. OSS Files and Disk Volumes 1 2 3 5 5 5 2 /sdir1 /sdir1/a /sdir1/a/file1 / (root) 4 ... 4 3 ...
OSS Pathnames Understanding the OSS File System Figure 3-2. Guardian Files and Disk Volumes 1 \MYNODE $OTHVOL 2 3 2 $OURVOL 3 SUBVOL1 4 AFILE1 AFILE2 MYSUBVOL 4 MYFILE1 MYFILE2 \MYNODE.$OURVOL.MYSUBVOL.MYFILE1 3 3 SUBVOL2 4 BFILE1 BFILE2 \MYNODE.$OURVOL.YRSUBVOL.YRFILE1 YRSUBVOL 4 YRFILE1 YRFILE2 Legend 1 2 HP NonStop server node Disk volumes 3 Subvolumes 4 Files VST003.VSD Each pathname for an OSS regular file has an underlying Guardian filename.
OSS Pathnames Understanding the OSS File System define or manage these filesets through the SCF interface and they do not appear in the fileset configuration database. The OSS pathnames of Guardian files can have no more than four elements including the /G, as illustrated in Figure 3-3. These elements, although technically OSS filenames, are subject to the length restrictions for Guardian filenames. Figure 3-3. Guardian Files in the OSS File System / (root) directory /E /G 2 1 ... ... ... ...
Understanding the OSS File System Using Pathnames for Remote Files Using Pathnames for Remote Files The rules described in the preceding subsection can be extended to the files accessible through the /E directory. A file on a remote node that is connected to your node through the Expand network appears in /E when all of the following are true: • • • The remote Expand node has a TOSVERSION of D40 or later. The remote node has an OSS name server running with a device subtype of 5.
Using the Local Root Directory as a Pathname Understanding the OSS File System Such a search would be extremely time-consuming and would return undesired information on /E and /G files. OSS shell commands that perform recursive operations include: chgrp chmod chown cp diff find ls mv pax rm rmdir Other products running in the OSS environment might also have commands with recursive behavior.
Understanding the OSS File System OSS File Components OSS File Components Each OSS data file, or regular file, consists of two components: • • The catalog portion, which is recorded in an OSS name server catalog file. This portion supplies structural information about the file. The data portion, which is stored in a file that is identified by a Guardian filename. This portion contains the data seen by end users of the file.
Understanding the OSS File System Relating OSS Files, Filesets, and Disk Volumes Guardian disk files are referenced using the OSS filenames in the /G directory. Only Guardian disk files that are not on disk volumes administered through SMF are visible in the /G directory. Guardian disk files on optical disks are visible in the /G directory but cannot be read from or written to through the OSS file system. For information on accessing files in the /G directory, see the Open System Services User’s Guide.
Understanding the OSS File System • Fileset Size Considerations Can be moved, copied, or restored to systems running RVUs that do not support OSS files larger than 2 gigabytes. In such cases, if the file has an underlying Guardian file format of Format 2 but a size of less than approximately 2 gigabytes: ° The file is automatically converted to use an underlying Guardian format of Format 1. ° The file becomes subject to the file size limit of approximately 2 gigabytes.
Understanding the OSS File System OSS Configuration Files Fileset size depends mainly on the type and complexity of the application mix running on your system. If you are porting an application from another system, you would have some idea of the application’s requirements and could use that as a basis for estimating the application’s disk-space requirements on the NonStop system. OSS Configuration Files The OSS file system gets configuration information from the ZOSSFSET file and storage-pool files.
Understanding the OSS File System OSS Configuration Files As each new file is created, the fileset’s OSS name server continues along the list of volume names, selecting a new volume with each request. The OSS name server ultimately wraps around to the beginning of the list in a round-robin fashion. Thus, if users write only small files in one disk volume in the list and only large files in another, then one volume can fill up before the others.
OSS Configuration Files Understanding the OSS File System Figure 3-4. Storage Pools and Disk Volumes OSS name server Finding an existing file Finding space for creating a file Disk volumes containing OSS files in fileset DATA5 Disk volume containing storage-pool file Disk volumes on which new OSS files can be created in fileset DATA5 VST018.VSD While OSS filesets can span multiple physical disk volumes, individual files cannot.
4 Managing Servers This section describes how to manage the servers used to provide the Open System Services (OSS) environment. Not all servers are managed through the same interface; read the subsections Introducing the OSS Servers on page 4-1 and Configuration Files on page 4-7 before attempting any operation on an OSS server.
The OSS Name Servers Managing Servers Use of the OSSTTY utility as a server is application-dependent. The OSSTTY server is managed using TACL commands, as described in OSSTTY on page C-1. Servers in other subsystems use OSS name servers for OSS pathname resolution.
The OSS Message-Queue Server Managing Servers The backup server process preserves message-queue data as well as the queues themselves if the primary server process fails. You can control the processor in which the backup server process runs. Figure 4-1.
The OSS Sockets Local Server Managing Servers The OSS Sockets Local Server The OSS sockets local server can run as a single process or as a fault-tolerant process pair. The server uses the default process name $ZPLS. The backup server process preserves socket access if the primary server process fails. You can control the processor in which the backup server process runs.
The Network Services Servers and Tools Managing Servers The Network Services Servers and Tools The following subsections briefly discuss: • • • • • inetd on page 4-5 rshd on page 4-5 rexecd on page 4-5 portmap and RPCINFO on page 4-6 BIND 9 Domain Name Server and Tools on page 4-6 All but RPCINFO are usually demon processes or processes started by demon processes on UNIX systems.
The Network Services Servers and Tools Managing Servers running by themselves, or under the Enterprise Toolkit—NonStop Edition. rexecd is started by the inetd process, which must be running for remote SQL/MX compilations. For information about the behavior of the rexecd process, see the rexecd(8) reference page either online or in the Open System Services Shell and Utilities Reference Manual.
Configuration Files Managing Servers checking the signature, other security-aware domain name servers can verify the integrity and authenticity of DNS data. For more information on this implementation of DNS, see RFCs 1033, 1034, and 1035, the BIND 9 Administrator Reference Manual, and the DNS Configuration and Management Manual. The lightweight resolver utility, lwresd, is available for use with specific application program interface (API) functions.
Configuration Files Used for the OSS Name Servers Managing Servers The ZOSSFSET File The ZOSSFSET file is an Enscribe alternate-key file that contains the fileset configuration database. The ZOSSFSET file contains entries that identify the storagepool file and operating characteristics for each fileset in the OSS file system. Entries in this file can be displayed with the SCF INFO FILESET command.
Configuration Files Used for the OSS Name Servers Managing Servers • When the BUFFERED CREATE (fast-create) attribute is set for a fileset: ° ° ° The fileset cannot use more than one volume. The fileset catalog must reside on that volume. The storage-pool file specification is ignored and the storage pool used for file creation (the creation pool) is the catalog disk volume. The BUFFERED CREATE attribute controls buffering of new file labels by the disk process.
Configuration Files Used for the OSS Name Servers Managing Servers currently contains the following listed initial fileset entries.
Configuration Files Used for the OSS Name Servers Managing Servers Attribute Fileset Fileset Fileset (page 2 of 2) REPORT 2 SUBSYS REPORT specification SUBSYS REPORT specification SUBSYS REPORT specification 1 Beginning with RVU G06.15. 2 Beginning with RVU G06.17. 3 Beginning with RVU G06.17 and until RVU G06.18, the default was -1 (same processor as OSS Monitor). 4 Beginning with RVU G06.24.
Configuration Files Used for the OSS Name Servers Managing Servers Figure 4-2. Relationship Among OSS Configuration Files, Processes, and Disk Volumes SCF $SYSTEM $SYSTEM.ZXOSSMON.ZOSSFSET 1 4 $ZPMON $SYSTEM.ZXOSSMON.ZOSSPOOL Disk names 5 $DATA1 $DATA2 Application 3 3 $DATA1.ZYQ00000.Z0000011 $DATA1 $DATA2.ZYQ00000.Z0000435 $ZPNS 2 $DATA2 $ROOTCAT $ROOTCAT.ZX000000.PXLINK $ROOTCAT.ZX000000.PXLOG $ROOTCAT.ZX000000.
Configuration Files Used for the OSS Name Servers Managing Servers ZOSSFSET and might have embedded the factory-default node number in the ALTFILE attributes. After you change the node number, you must check that the OSS Monitor can open the OSS fileset configuration database the next time the OSS Monitor is started. Enter the following command at a TACL prompt after changing the node number: FUP INFO ZOSSFSET,DETAIL If output includes an ALTFILE name that looks something like: \??.$SYSTEM.ZXOSSMON.
Configuration Files Used for the OSS Name Servers Managing Servers If your system has been upgraded from a G05.00 or later G-series RVU, the OSS Monitor automatically creates an initial ZOSSPARM file containing the information from the old system’s ZPOSPARM. If your system has been upgraded from a G-series RVU preceding G05.00, the OSS Monitor automatically creates an initial ZOSSPARM file containing the information from the old system’s TACL PARAMs. See PARAMs Used by the OSS Monitor on page 2-10.
Configuration Files Used for the OSS Name Servers Managing Servers This file must be maintained by the system manager using the SCF ADD SERVER, ALTER SERVER, and DELETE SERVER commands. ZOSSSERV must be present before the root fileset is first mounted and before the first OSS name server is started. Entries in the ZOSSSERV file must follow these rules: • • • There should always be a #ZPNS entry for the root OSS name server process, $ZPNS.
Configuration Files Used for the OSS Name Servers Managing Servers SERVER: • • • For the ROOT fileset #ZPNS #ZPNS For the HOME fileset #ZPNS #ZPNH For the TEMP fileset #ZPNS #ZPNS SQLTIMEOUT 60 60 TYPE NAME NAME Beginning with RVU G06.
Configuration Files Used for the OSS Name Servers Managing Servers Figure 4-4. OSS Configuration Files, Processes, and Disk Volumes Affected by Altering an OSS Name Server Entry in ZOSSSERV SCF $ZPMON Processor 1 Processor 2 Processor 3 SQLCAT 1 6 6 $SYSTEM 2 5 $SYSTEM.ZXOSSMON.
Configuration Files Used for the OSS Name Servers Managing Servers • • The file identifier ZOSSPOOL should not be used, to avoid conflict with an unreleased HP feature. The file identifier OSSPOOL should not be used, to avoid conflict with the file identifier for the sample storage-pool file installed with the OSS product set.
Configuration Database Files Used for the OSS Message-Queue Server Managing Servers Storage-pool file content: • • • • Preconfigured or all defaults accepted $OSS $OSS $OSS Single pool volume specified disk1 disk1 disk1 Two pool volumes specified disk1, disk2 disk1, disk2 disk2 Three or more pool volumes specified all specified disks all specified disks disk3 where disk1, disk2, and disk3 are volume names specified to OSSSETUP.
Configuration Database Files Used for the OSS Sockets Local Server Managing Servers accepted all defaults, ZOSSSERV contained the following initial OSS message-queue server information: Attribute Single-Enclosure System Multiple-Enclosure System BACKUPCPU 0 3 CPU 1 2 MSGMQB 65535 65535 MAXMQID 32 32 MAXMSG 32 * MAXMQID = 1024 32 * MAXMQID = 1024 MSGMSIZE 32000 32000 SERVER #ZPLS #ZPLS TYPE LOCAL LOCAL Beginning with RVU G06.
Configuration Database Files Used for the OSS Sockets Local Server Managing Servers The ZOSSSERV file is the OSS sockets local server configuration database, and it contains an automatically created entry that identifies the characteristics of the OSS sockets local server process. This file must be edited by the system manager using the SCF ALTER SERVER command. ZOSSSERV need not be configured before the root fileset is first mounted or before the first OSS sockets local server is started.
Configuration Database Files Used for the OSS Sockets Local Server Managing Servers Beginning with RVU G06.
Configuration Database Files Used for the OSS Transport Agent Servers Managing Servers Figure 4-5. OSS Configuration Files, Processes, and Disk Volumes Affected by Altering an OSS Sockets Local Server Entry in ZOSSSERV Processor 1 SCF Processor 2 Processor 3 $ZPMON 5 3 1 $ZPLS 4 3 $SYSTEM $ZPLS 2 $SYSTEM.ZXOSSMON.
Configuration Files for the Network Services Servers and Tools Managing Servers Configuration Files for the Network Services Servers and Tools The following subsections briefly discuss the configuration files for the following programs: • • • • • inetd on page 4-24 rshd on page 4-25 portmap on page 4-26 RPCINFO on page 4-26 BIND 9 Domain Name Server and Tools on page 4-27 The remote execution server, rexecd, does not have a configuration file.
Configuration Files for the Network Services Servers and Tools Managing Servers These OSS files contain the same kind of information as is used by Guardian sockets programs in the Guardian environment. Guardian sockets services are configured using the following files: $SYSTEM.ZTCPIP.HOSTS $SYSTEM.ZTCPIP.IPNODES (for NonStop TCP/IPv6 only) $SYSTEM.ZTCPIP.NETWORKS $SYSTEM.ZTCPIP.PROTOCOL $SYSTEM.ZTCPIP.RESCONF $SYSTEM.ZTCPIP.SERVICES There are two versions of the resolver that you can use on a node.
Configuration Files for the Network Services Servers and Tools Managing Servers /etc/hosts.equiv is bypassed when a remote user attempts to use an rsh command as the super ID. These files are sometimes the target of UNIX system intruders. Take standard precautions for their use on a UNIX system when setting them up for the OSS environment. See the hosts.equiv(4) and .
Configuration Files for the Network Services Servers and Tools Managing Servers Guardian environment and in OSS documentation). The portmapper program has RPC program number 100000. # # rpc version 1.16 # portmapper 100000 2001/04/27 portmap sunrpc The fields of each program-entry line are separated by blanks. Any entry other than the portmapper entry would be a requirement unique to another product requiring the use of RPCINFO.
Adding a Server Managing Servers File Purpose (continued) /etc/dns_secure/ named.conf Sample of /etc/named.conf provided for the secure version of named. /etc/dns923/rndc.conf Sample of /etc/rndc.conf provided for the nonsecure version of rndc. /etc/dns_secure/ rndc.conf Sample of /etc/rndc.conf provided for the secure version of rndc.
Configuring a Server Managing Servers b. Use the STOP FILESET Command to stop the started fileset. 4. Use the START FILESET Command to start (mount) the new or stopped fileset assigned to the new OSS name server. This action automatically starts the new OSS name server. 5. If your site uses the STARTOSS utility and the new OSS name server services a new fileset, you should also add the new fileset name to the OSSINFIL file. See OSSINFIL File on page C-19 for more information.
Configuring the OSS Message-Queue Server Managing Servers Configuring the OSS Message-Queue Server You configure the OSS message-queue server by changing its entries in the Enscribe database ZOSSSERV file. The following procedure assumes that no applications that use message queues have been started yet: 1. To stop the OSS message-queue server, enter this OSS Monitor SCF command: STOP SERVER $ZPMON.#ZMSGQ 2.
Configuring the OSS Transport Agent Servers Managing Servers ° To reassign the process to a new primary or backup processor, change the corresponding processor entry. ° To change behavior associated with use of the automatic startup service, change the corresponding BACKUPCPUOK, MAXWAITTIME, DESIREDSTATE, or AUTORESTART entry. 3. To restart the server, enter this OSS Monitor SCF command: START SERVER $ZPMON.
Configuring Network Services Servers, Tools, and Applications Managing Servers environment by creating symbolic links between the Guardian files and the /etc directory: 1. Check that the files or links do not already exist in the /etc directory by entering the following command at an OSS shell prompt: cd /etc ls -al If resolv.conf, hosts, and inetd.conf appear, ignore the rest of this procedure and see portmap on page 4-32 to configure the portmap process. Note.
Configuring Network Services Servers, Tools, and Applications Managing Servers • The TCP/IP domain name resolution (resolver configuration) file used by portmap. For example: ADD DEFINE =TCPIP^RESOLVER^NAME, FILE ALTRES • The TCP/IP host definition file used by portmap. For example: ADD DEFINE =TCPIP^HOST^FILE, FILE ALTHOST Some declarations are not valid in certain combinations.
Configuring Network Services Servers, Tools, and Applications Managing Servers rexecd To configure the rexecd process: 1. Ensure that the /etc/services file is accessible by entering: cd /etc ls -al If the /etc/services file is not listed, follow the directions on page 4-35 to make it accessible, then continue with Step 2. 2. Add a port specification for the exec service to the /etc/services file. 3.
Managing Servers • Configuring Network Services Servers, Tools, and Applications You must configure network services that such applications might use. You configure the necessary network services by making AF_INET or AF_INET6 sockets configuration files available in the OSS file system. To prevent confusion and conflicts between servers, use and maintain the Guardian version of the AF_INET or AF_INET6 sockets configuration files for both environments.
Starting a Server Managing Servers • • You must keep track of the process names or OSS process IDs (PIDs) that OSS sockets applications use so that those processes running in a specific processor (and possibly all processes used by the same application) can be stopped before you stop an OSS transport agent server. You might want to run multiple TCP/IP processes for scalability or load-leveling.
Starting the OSS Message-Queue Server Managing Servers c. Use the OSS Monitor SCF STATUS FILESET $ZPMON.*, DETAIL command to make sure that all filesets with mount points between / and the mount point directory for the fileset selected in Step 2a are started. See the STATUS FILESET Command on page 12-66 for a description of the output. d. Issue the OSS Monitor SCF START FILESET Command for any unstarted filesets identified in Step 2c. 3.
Starting an OSS Transport Agent Server Managing Servers Starting an OSS Transport Agent Server An OSS transport agent is started automatically when its processor starts. You do not need to start an OSS transport agent server to use an OSS AF_UNIX, OSS AF_INET, or OSS AF_INET6 sockets application in a given processor. To restart a stopped OSS transport agent server: 1. Make sure that you are a member of the super group (255,nnn). 2. Enter an OSS Monitor SCF START SERVER Command.
Obtaining Information About a Server Managing Servers • The following environment variable allows the $ZTC2 process to be used as the inetd transport-provider process: export SOCKET_TRANSPORT_NAME=\$ZTC2 You start the portmap program from a Guardian TACL prompt or using the OSS shell gtacl utility.
Determining Whether a Server Is Running Managing Servers Here is an example: Suppose you want to determine whether a transport-provider process for OSS AF_INET sockets is running. The transport provider is the $ZTCn server, described under OSS Sockets on page 1-15. If you enter the LISTDEV command at an SCF prompt, a display similar to that shown in Figure 4-6 on page 4-40 appears.
Determining the Current Configuration of a Server Managing Servers Checking Servers That Are Administered Through the OSS Monitor You can use the OSS Monitor SCF STATUS SERVER command to determine the current status for a server that is administered through the OSS Monitor. For example, if you enter the following command at an SCF prompt: STATUS SERVER $ZPMON.* you can determine the state of all servers administered through the OSS Monitor. The information displayed is the state of current processes.
Determining Usage and Configuration of Network Services Servers Managing Servers • • • • The user who created the server configuration, when it was created, the user who last modified the configuration, and when it was last modified The automatic startup settings for the server (AUTORESTART, DESIREDSTATE, and PERSISTENCECOUNT attributes) If the server is an OSS name server: ° ° ° The size of the inode cache The size of the link cache The timeout value used for input or output with the SQL catalog serv
Stopping a Server Managing Servers Stopping a Server How and when you stop a server depends on the type of server. • • • • • Stopping a Specific OSS Name Server on page 4-43 Stopping the OSS Message-Queue Server on page 4-44 Stopping the OSS Sockets Local Server on page 4-44 Stopping an OSS Transport Agent Server on page 4-45 Stopping a Network Services Server on page 4-45 Other servers used by OSS applications require separate procedures.
Stopping the OSS Message-Queue Server Managing Servers If the OSS name server becomes unstoppable, do one of the following: • • Repair all filesets involved using the OSS Monitor SCF DIAGNOSE FILESET Command, restart them, then stop them again. Add a new fileset to be managed by the unstoppable OSS name server (see the ADD FILESET Command on page 12-7 for command syntax). Start and stop that fileset; for example, if the new fileset is named DUMMY, enter: START FILESET $ZPMON.DUMMY STOP FILESET $ZPMON.
Stopping an OSS Transport Agent Server Managing Servers 2. Warn your users. You can use a method similar to the one described under Manually Stopping the OSS File System and the OSS Environment on page 2-3. 3. Issue the following OSS Monitor SCF command: STOP SERVER $ZPMON.#ZPLS Stopping an OSS Transport Agent Server If you need to shut down the QIO subsystem, you must first stop the OSS transport agent server. For more information about the QIO subsystem, see the QIO Configuration and Management Manual.
Reconfiguring a Server Managing Servers Reconfiguring a Server How and when you reconfigure a server depends on the type of server. • • • • Reconfiguring an OSS Name Server on page 4-46 Reconfiguring the OSS Message-Queue Server on page 4-47 Reconfiguring the OSS Sockets Local Server on page 4-48 Reconfiguring a Network Services Server on page 4-49 You cannot reconfigure an OSS transport agent server. Other servers used by OSS applications require separate procedures.
Reconfiguring the OSS Message-Queue Server Managing Servers 7. Do one of the following: • If the OSS name server is the server for the root fileset, stop all filesets by entering the following OSS Monitor SCF command: STOP FILESET $ZPMON.* • If the OSS name server is not the server for the root fileset, reassign all filesets it manages to another running OSS name server. Use the OSS Monitor SCF ALTER FILESET Command on each fileset to change the NAMESERVER attribute for the fileset.
Reconfiguring the OSS Sockets Local Server Managing Servers 3. If you changed only one or more of these attributes: AUTORESTART, BACKUPCPUOK, or MAXWAITTIME, you have completed the task. The changes take effect immediately. 4. Otherwise, warn your users. You can use a method similar to the one described under Manually Stopping the OSS File System and the OSS Environment on page 2-3. 5. Use the following OSS Monitor SCF command to stop the server: STOP SERVER $ZPMON.#ZMSGQ 6.
Reconfiguring a Network Services Server Managing Servers Reconfiguring a Network Services Server Most network services servers ignore changes to configuration files while they are running. The BIND 9 domain name server named can be reconfigured using the rndc or nsupdate utility, as described in the DNS Configuration and Management Manual and the nsupdate(8) and rndc(8) reference pages online. Configuration files such as /etc/inetd.conf can be edited while the servers are running.
Removing a Network Services Server Managing Servers 2. Use the following OSS Monitor SCF command to determine which filesets are managed by the OSS name server that you want to remove: INFO FILESET $ZPMON.*, DETAIL See the INFO FILESET Command on page 12-47 for a description of the output. 3. Use the OSS Monitor SCF ALTER FILESET Command on each fileset managed by the OSS name server you want to remove to change its NAMESERVER entry to specify another OSS name server. 4.
Managing Servers Troubleshooting a Server Monitor Messages on page A-27 or by using the SCF HELP facility described in Online Help Facility on page 12-2. 2. When you are unsure of the outcome of a command entry, check the EMS log for the most recent status messages. Look up the status messages in the Operator Messages Manual. For a list of the subsystem IDs to look for in the EMS log, see Appendix A, Messages. 3.
Managing Servers Troubleshooting a Server Open System Services Management and Operations Guide— 527191-004 4- 52
5 Managing Filesets This section describes how to perform operations on Open System Services (OSS) filesets.
Creating a Unique Fileset Managing Filesets • To create the directory /home/henrysp from within the Guardian environment, enter the following commands at a TACL prompt: OSH -p /bin/mkdir /home/henrysp OSH -p /bin/chmod 0777 /home/henrysp Note that the OSS shell chmod command gives read, write, and search (execute) permission to all users of OSS files within the mount-point directory.
Creating a Unique Fileset Managing Filesets in which file creation, deletion, or opens are a relatively common occurrence is a a good candidate for the BUFFERED LOG option. ° A fileset using the BUFFERED CREATE (fast-create) option uses memory cache for as much file creation, open, and close information as possible between updates of the actual fileset catalog file on disk.
Creating a Unique Fileset Managing Filesets NFSTIMEOUT 70, DESIREDSTATE STARTED, FSCKCPU -1, MAXDIRTYINODETIME 10, REPORT $S.#MINE & This command: • • Adds the fileset named MINE at the mount point /mine of the local node and sends repair information to the report file spooler location $S.#MINE Keeps the fileset catalog on disk volume $CAT Figure 5-1. OSS Configuration Files, Processes, and Disk Volumes Involved in Adding a Fileset $ZPMON SCF 5 2 $SYSTEM 1 $SYSTEM.ZXOSSMON.
Creating a Unique Fileset Managing Filesets • • • • • • Decreases the likelihood of needing fileset recovery by decreasing the value used for MAXDIRTYINODETIME from its default value to 10 seconds Makes the fileset automatically restart when necessary if the automatic startup service is used, and uses the processor (specified as -1) used by the disk process for the fileset to perform any fileset repairs needed Specifies that the OSS name server for the fileset retains the results of nonretryable Network
Creating a Storage Pool Managing Filesets Figure 5-2. Starting (Mounting) a Fileset Before mounting the fileset USER1 from the volume $DATA5 onto the mount point /usr/utils/src. After mounting the fileset USER1 onto the mount point /usr/utils/src. / (root) / (root) ... ... /dev /bin /etc /dev /usr /bin /etc /usr ... /include /lib ...
Managing Filesets Starting (Mounting) or Restarting Filesets The name of the storage-pool file used for each fileset appears in that fileset’s database record. The OSS Monitor uses this information to tell the OSS name servers where to put OSS data files. Figure 5-3 on page 5-8 shows the contents of a storage-pool file. This storage-pool file was created during installation by copying the sample EDIT file OSSPOOL from the ZXOSSMON subvolume of the installation volume and modifying it appropriately.
Managing Filesets Automatic Restart of Filesets During OSS Monitor Startup Figure 5-3. Example of a Storage-Pool File == == == == == == == == == File: $SYSTEM.ZXOSSMON.OSSPOOL This file is a sample POOL file. A POOL file defines the disk volumes where OSS files of an OSS FILESET can be created. OSS FILESETs, which are managed by the OSS Monitor ($ZPMON), have an attribute named POOL, which is the name of a Guardian EDIT file that resides in the subvolume $SYSTEM.ZXOSSMON.
Managing Filesets Automatic Restart of Filesets by the Automatic Startup Service entry is made indicating that the fileset is in the STARTED state. When the fileset is stopped, the state of the fileset is changed to STOPPED. When the OSS Monitor is started for the first time after a system load, it checks the desired-state configuration for all filesets to determine which filesets it must start or restart.
Managing Filesets Automatic Restart of Filesets After OSS Name Server Failure The automatic startup service can also restart the fileset a maximum number of times during a 10-minute period. See the ADD FILESET Command on page 12-7 or the ALTER FILESET Command on page 12-20 for more information about this service. Automatic Restart of Filesets After OSS Name Server Failure If an OSS name server fails, the OSS Monitor initiates a recovery procedure similar to that performed during OSS Monitor startup.
Managing Filesets Potential Problems During Automatic Restart of Filesets failed. The OSS Monitor reinitiates the fileset restart sequence when the specified FSCK processor is reloaded. You might not want to wait for the FSCK processor to be reloaded. You can correct this situation manually by changing the FSCKCPU value (See ALTER SUBSYS, ALTER MON, and ALTER PROCESS Commands on page 12-34).
Managing Filesets Auditing a Fileset Auditing a Fileset An important component of a secure file system is the ability to trace the history of security-related operations on objects in the system. OSS security auditing allows you to collect a history of audited operations—that is, an audit trail—on a specified set of auditable objects in the system. OSS security auditing allows you to audit access to objects in the OSS filename space.
Managing Filesets Obtaining Information About a Fileset SCF Commands Used Actions taken (continued) ALTER FILESET When a member of the super group attempts to alter the value of a fileset’s AUDITENABLED attribute, an audit record is generated. The before and after values of the AUDITENABLED attribute are included.
Managing Filesets • • • Reconfiguring a Fileset To make the files in the fileset unavailable to users. If a fileset is stopped while users have files open in that fileset, the open files can be used normally. However, once a file is closed, the file cannot be reopened until the fileset is mounted again. To free memory in the OSS name server for that fileset. To delete the fileset from the OSS file system. If your site uses the STOPOSS utility, that utility stops all filesets.
Managing Filesets • • • • • • • • • Changing the Operating Parameters of a Fileset Fileset input/output fault-tolerance (FTIOMODE and NORMALIOMODE attributes) Fileset catalog buffering (the BUFFERED attribute) Fileset storage-pool storage-pool file (POOL attribute) Maximum number of files and directories (MAXINODES attribute) User access restriction (the READONLY attribute) Network File System (NFS) request timeout or cache (pool) size Catalog file volume (CATALOG attribute) OSS file system mount point (
Managing Filesets Changing the Operating Parameters of a Fileset attribute settings are, from highest to lowest fault tolerance and from lowest to highest performance: UNBUFFEREDCP DP2BUFFEREDCP OSSBUFFEREDCP DP2BUFFERED OSSBUFFERED The behaviors associated with these levels of fault tolerance are shown in Table 5-1. Table 5-1.
Managing Filesets Changing the Operating Parameters of a Fileset Table 5-1.
Managing Filesets Changing OSS File Caching for the Disks of a Fileset Value When to use it (continued) LOG An OSS name server does not write PXLOG records; it checkpoints these records to its backup process. The backup process of an OSS name server keeps these records in its memory and uses them to recover partially completed operations in the event of a failure of an OSS name server primary process.
Managing Filesets Changing OSS File Caching for the Disks of a Fileset If you disable OSS file caching, HP strongly recommends that you disable OSS file caching on all the disk volumes in a fileset. To disable OSS file caching for a specific disk volume, you must disable all filesets mounted on the affected fileset: 1. At a TACL prompt, enter: SCF 2. At the SCF prompt, enter the following SCF command to stop each fileset involved: STOP FILESET $ZPMON.
Managing Filesets Changing OSS File Caching for the Disks of a Fileset 3. At an SCF prompt, enter the following set of commands once for each disk volume in the fileset: STOP DISK diskname ALTER DISK diskname, OSSCACHING ON START DISK diskname diskname is the name of a disk volume that contains OSS files. 4. Restart the affected portion of the OSS file system by entering the following SCF command one or more times: START FILESET $ZPMON.
Managing Filesets Changing the Physical Makeup of a Fileset Disabling OSS file caching changes the fault tolerant behavior of the fileset. Turning it off converts OSSBUFFERED behavior to DP2BUFFERED behavior or OSSBUFFEREDCP behavior to DP2BUFFEREDCP behavior. The caching status of a file can change as opens, closes, and other events occur on the file.
Managing Filesets Changing the Physical Makeup of a Fileset fileset; existing OSS files will not be moved to the new volume, but new OSS files will be created on the new disk volume. Make sure that the new catalog disk volume has adequate space for the fileset. 3. Use the SCF ALTER FILESET Command to enter a new name for the catalog disk volume of the fileset.
Managing Filesets Changing the Physical Makeup of a Fileset 2. Use the SCF INFO FILESET Command to determine the name of the storage-pool file for the fileset. 3. Use a Guardian text editor to convert the entry for the disk volume to a comment within the storage-pool file. Note. You should never delete the entry for a volume name from a storage-pool file once the disk volume contains OSS files; converting such entries to comments provides an easy way to document the volume list for the fileset. 4.
Managing Filesets Checking and Repairing Fileset Integrity 9. If the catalog disk volume for the fileset is on the removed disk volume, use the SCF ALTER FILESET command to enter a new name for the catalog disk volume of the fileset. This change causes current catalog files to be moved to the specified new disk volume; previously saved catalog files (with file identifiers of the form PXINnnnn, PXLInnnn, and PXLOnnnn) are not moved. The subvolume name of the catalog files is not changed. 10.
Managing Filesets FSCK Log File Open System Services was designed to allow customers to configure systems where FSCK should never need to run, even in most multiple-failure situations. Traditional UNIX systems are different: those systems have relied on running fsck whenever there is a system failure.
Managing Filesets FSCK Log File The FSCK log file is created with the access permissions for the user ID of the OSS Monitor process when that process was started. You can specify a Guardian filename for the log file using the REPORT option of the ADD FILESET Command, ALTER FILESET Command, or DIAGNOSE FILESET Command. If you do not specify a filename for the log file, the log file is sent to the spooler location specified for the REPORT attribute of the SUBSYS object.
Managing Filesets FSCK Log File Figure 5-4. FSCK Log File Examples (page 1 of 2) FSCK - T8621G09 - (15DEC2001) OSS Fileset Validate/Repair Utility Copyright Compaq Computer Corporation 1994, 1995, 1996, 2001 August 31,2001 01:14:53 Options Selected are: CATALOG $OSS.ZX000000, REPAIR ALL, OPEN FSCK Serial Number = 0005 Catalog Subvolume is \NODE1.$OSS.
Managing Filesets FSCK Log File Figure 5-4. FSCK Log File Examples (page 2 of 2) FSCK - T8621G09 - (15DEC2001) OSS Fileset Validate/Repair Utility Copyright Compaq Computer Corporation 1994, 1995, 1996, 2001 August 31,2001 01:14:53 Options Selected are: CATALOG $OSS.ZX000000, STATUS, ROOT Catalog Subvolume is \NODE1.$OSS.
Managing Filesets Inconsistencies Checked by FSCK Value Mounted state (continued) 0x6002 The fileset is mounted in BUFFERED LOG mode and is marked clean. 0x8001 The fileset is being recovered. 0x8002 The fileset is mounted in BUFFERED LOG mode. 0xA000 The fileset is mounted in BUFFERED NONE mode. 0xFFFF The fileset is mounted in BUFFERED NONE mode by an OSS name server that is not running the current code version.
Managing Filesets Inconsistencies Checked by FSCK Table 5-2. Inconsistencies Checked by FSCK (page 2 of 4) Inconsistency Type Explanation Bad free-inode list Serious An OSS name server maintains two free-inode lists in each catalog. One is used to list the inodes that can be immediately reused. The other is used to list those inodes that can be reused only after a successful unmount and mount sequence.
Managing Filesets Inconsistencies Checked by FSCK Table 5-2. Inconsistencies Checked by FSCK (page 3 of 4) Inconsistency Type Explanation Loop in directory graph Serious If FSCK detects a loop in the directory graph, it breaks the loop by deleting a PXLINK record in the loop. Such a loop occurs when an inode is linked through other inodes back to itself; this means that the directory referred to by the inode number is its own parent.
Managing Filesets Inconsistencies Checked by FSCK Table 5-2. Inconsistencies Checked by FSCK (page 4 of 4) Inconsistency Type Explanation Missing ZYQ file Minor The ZYQ file corresponding to a regular inode does not exist. Because this inconsistency is usually the result of a failure during file creation or during the removal of the last link to a file, the inode and any links to it are deleted.
Managing Filesets Generated Catalog Files Generated Catalog Files The FSCK utility saves a copy of the existing catalog and creates a new one. Each time that FSCK is run to repair a fileset catalog, a unique four-digit FSCK serial number (FSN) is assigned for that run. The FSN is used to encode the names of the saved files as follows: PXINODE: PXINnnnn PXLINK: PXLInnnn PXLOG: PXLOnnnn where nnnn is the FSN. During the final phase of FSCK, the catalog files are renamed.
Managing Filesets Deleting a Fileset Any other errors occurring during writes to the new catalog files cause FSCK to issue an error message and terminate. • • Any error other than Guardian File Management Error 1 (reaching an end of file) on FSCK’s inode swap file, FSCK issues an error message and terminates. An inconsistency in its internal data structures, it issues an error message and terminates.
Managing Filesets Updating Existing Fileset Configurations Updating Existing Fileset Configurations A fileset created while your site used a previous version of Open System Services can be used unmodified with the current software. However, the contents of the ZPCONFIG and ZPMNTTAB configuration files from release version updates (RVUs) preceding G05.00 no longer affect the current configuration of your filesets. If you had a G-series RVU preceding G05.
Managing Filesets Moving a Directory Hierarchy to Its Own Fileset Upgrading From an RVU Preceding G05.00 Older configuration files can be deleted. Once the G05.00 RVU is installed and the OSS file system is started for the first time, the content of the ZXMNTTAB file is no longer accurate. That file should be deleted as soon as you are sure that you need not fall back to a previous RVU. The ZXCONFIG file can also be deleted.
Managing Filesets Cleaning Up a Fileset b. Add a new OSS name server for the new fileset: ADD SERVER $ZPMON.#ZPN1, CPU 1, BACKUPCPU 0 c. Add the new fileset: ADD FILESET $ZPMON.HOME, DEVICELABEL 000001, & CATALOG $OSS, BUFFERED LOG, POOL POOL, & NAMESERVER #ZPN1, MNTPOINT /home d. Start the new fileset: START FILESET $ZPMON.HOME 6.
Managing Filesets Cleaning Up a Fileset The following example, which uses a fileset called HOME, shows how to eliminate extraneous inode entries in the ROOT fileset. You must follow this same procedure for every fileset within the ROOT hierarchy except the ROOT fileset itself: 1. In SCF, stop the HOME fileset by using the following commands: STOP FILESET $ZPMON.HOME 2. Use pax from an OSS shell prompt to back up the ROOT fileset. The HOME fileset is stopped, so it is not included in this backup.
Managing Filesets Troubleshooting Filesets Troubleshooting Filesets When you have problems managing a fileset, follow this general procedure: 1. Check the messages from the OSS Monitor that are sent to your terminal. If you redirect such messages to a log file, check the log file for its most recent entries. Look up the cause, effect, and recovery information for a message either in OSS Monitor Messages on page A-27 or by using the SCF HELP facility described in Online Help Facility on page 12-2.
Managing Filesets Managing and Repairing Fileset Catalog Files The LastError value reported (9) indicates that confusion exists over the state of the mount point. In this example, the LastError value is the same as the OSS Monitor error message number (E0009). Managing and Repairing Fileset Catalog Files An OSS name server maintains a catalog file for each fileset it manages. These catalog files require maintenance when: • • New OSS features require changes to the format of entries in the catalog.
Managing Filesets Moving and Removing OSS Catalog Files All mounted filesets can be stopped in one operation by using the wildcard character * in place of a fileset name in the SCF STOP command. To run the SCF DIAGNOSE FILESET Command on multiple filesets, use the CPU option to begin the process for each fileset on a different processor.
Managing Filesets Moving and Removing OSS Catalog Files HELP | ? is the HELP command, which provides an overview of the CVT command syntax. This is the default action; that is, if this RUN command is entered without any options, the command is interpreted as if the option HELP had been entered. PURGE SERIAL serialno [ IN subvolume ] is the PURGE command, which purges a catalog saved by FSCK. serialno is the FSCK serial number (FSN) of the saved catalog to be purged.
Managing Filesets Moving and Removing OSS Catalog Files For files2, indicates that the • • New location for the catalog files is the saved catalog with the FSN serialno. Files are to be renamed PXINserialno, PXLIserialno, and PXLOserialno. IN subvolume indicates the name of the subvolume for the corresponding files1 or files2 saved catalog files. If this name is not supplied, the user’s default subvolume is assumed.
Managing Filesets • Moving and Removing OSS Catalog Files To save the current catalog on $VOL.ZX000003 as PXIN0004, PXLI0004, and PXLO0004 on the same subvolume, enter the following command at a TACL prompt: RUN $SYSTEM.ZOSS.CVT RENAME CURRENT IN $VOL.ZX000003 TO & SERIAL 4 IN $VOL.ZX000003 This command is equivalent to the following command sequence: VOLUME $VOL.ZX000003 RUN $SYSTEM.ZOSS.CVT RENAME CURRENT TO SERIAL 4 • To make the current catalog on $VOL.TEMP the current catalog on $VOL.
6 Managing OSS Files To manage files in the OSS file system, you need to: • • • • • • • Obtain specific information about their size, location, and fileset membership, as described in Obtaining Information About OSS Files. Install new or updated product files, as described in Installing New Product Files on page 6-4. Remove obsolete files, as described in Removing Obsolete OSS Files and Directories on page 6-9.
Managing OSS Files Interpreting Guardian Filenames for OSS Files The OSS shell gname and pname utilities allow you to provide a Guardian filename and obtain the OSS pathname for a specific file and vice versa. See the following subsections for more information: • • Using the OSS gname Command on page 6-2 Using the OSS pname Command on page 6-3 You can also use the Guardian VPROC utility or the OSS shell vproc command to determine the product-version information for an OSS file supplied by HP.
Managing OSS Files Using the OSS pname Command Figure 6-1. OSS gname Command Examples $ gname test gname: test --> \NODE1.$VOL.ZYQ00000.Z0000DV3 $ gname tes* gname: test2 --> \NODE1.$VOL.ZYQ00000.Z0000KHP $ gname -s test \NODE1.$VOL.ZYQ00000.Z0000DV3 $ gname /E/node2/usr/test3 gname: /E/node2/usr/test3 --> \NODE2.$DATA.ZYQ00001.Z0000DV2 When more than one OSS pathname matches the possible wildcard expansion, only the last file with a matching pathname is listed.
Managing OSS Files Installing New Product Files display and next to the SECURITY heading in the FUP INFO, DETAIL display. For information about interpreting the OSS permissions, see the Open System Services User’s Guide. Figure 6-3. FUP INFO Displays for OSS Files 3> fup info $VOL.ZYQ00000.Z0000DV4 CODE EOF LAST MODIF $VOL.ZYQ00000 Z0000DV4 OSS 142 10:55 OWNER RWEP TYPE REC BL 254,254 -rw-rw-rw- 4> fup info $VOL.ZYQ00000.Z0000DV4, detail $VOL.ZYQ00000.
Managing OSS Files Installing New Product Files This set of archive files is large—on the order of tens of megabytes. It is, however, a temporary group of files. Unless DSM/SCM maintains the OSS files on your node, this set of files can be deleted after OSS installation is complete; deleting them when DSM/SCM maintains OSS files can slow down subsequent installations or upgrades because DSM/SCM will replace them.
Managing OSS Files Using COPYOSS Table 6-1.
Managing OSS Files Using COPYOSS From the archive files, COPYOSS copies the OSS product files into the OSS file system, placing them in locations where you would find them on a typical UNIX system—for example, into directories such as /bin, /usr/ucb, and /usr/include. This command completely installs the basic OSS product set and all other products that use the same installation subvolume. • To load only the Java servlet product files, enter these commands from a TACL prompt: $tsvvol.ZOSSUTL.
Managing OSS Files Using PINSTALL archive file. You can use the -cvf flags of the PINSTALL command to display the table of contents for the pax archive to determine if the named file is actually an empty directory.
Managing OSS Files Removing Obsolete OSS Files and Directories and should be ignored. Using PINSTALL on such files would overwrite current files with obsolete ones. • • • PINSTALL is installed as a part of the basic OSS product set and is available even if the subvolume containing COPYOSS has been deleted from your system.
Managing OSS Files • Updating the whatis Database Files DSM/SCM does not install and maintain OSS product files in your OSS file system, you must manually remove obsolete files after installing any new release version, RVU, or software product revision (SPR) and before using the merge_whatis command. You manually remove obsolete files by entering the following OSS shell commands: cd /etc/install_obsolete Pcleanup -r source Note that the command Pcleanup starts with an uppercase letter.
Managing OSS Files Backing Up and Restoring OSS Files You create or update a whatis database file by merging whatis database fragment files using the merge_whatis utility. Each whatis database fragment file has an OSS filename with the form whatis.piece, where piece varies according to the product containing the reference page files.
Managing OSS Files • • Considerations The pax utility does not support labeled tapes and imposes requirements for unlabeled tapes. The pax utility does not support filenames longer than 100 characters in archives using its default USTAR format. The pax utility can be used when any of the following conditions exists: • • • The system that will be restored (target system) might not have the same node name or node number as the system being backed up (source system).
Managing OSS Files Considerations Guardian Tape Devices and the pax Utility The pax utility uses Guardian tape devices to read and write tape archives. You cannot use pax on labeled tapes. If you need to backup OSS files to labeled tapes or restore OSS files from labeled tapes, use Backup and Restore 2.0. You are also restricted when using pax to single write operations on unlabeled tape.
Managing OSS Files Backing Up the OSS Environment Using a Version of Backup/Restore If the request is canceled or the archive cannot be opened, the pax utility issues the following message to the originating terminal: filename cannot be opened, Guardian file system error: n where n is a Guardian file-system error number. A single archive can span more than one reel.
Managing OSS Files Backing Up User Files information relevant to OSS file access. You can do this in either of the following ways: ° Using Safeguard: SAFECOM DISPLAY AS COMMANDS ON LOG OSSGRPS INFO GROUP * LOG OSSUSRS INFO USERS *.* LOG OSSALIAS INFO ALIAS * LOG Edit these log files to create OBEY files for Safeguard so that the attributes of each file-sharing group, user ID, and user ID alias can be recreated if necessary.
Managing OSS Files Backing Up User Files entire file-system backup. You could write a shell script to do this. For information about shell scripts, see the Open System Services User’s Guide. You can use the find command to produce a list of files that must be backed up and then pipe this list to the backup command pax.
Managing OSS Files Backing Up User Files Backing Up the OSS File Hierarchy of the Current Directory To back up the file hierarchy of the current directory to the tape mounted on Guardian tape device $TAPE, using the blocking factor for 5120 bytes, enter the following OSS shell command: pax -wv -f /G/tape -b 10b . This command has the following form: pax -w -v -f archive_name -b blocksize . -w writes files to the standard output file in the specified archive format (the default format is ustar).
Managing OSS Files Backing Up User Files Backing Up OSS File Hierarchies Using Backup and Restore 2.0 You can specify part or all of multiple directory hierarchies and selectively backup files from within them using Backup and Restore 2.0. For example, if you enter the following at a Backup and Restore 2.0 BR> prompt: BACKUP =MYTAPE, OSS ((/user/bin, /home/sv/myfile, /usr/local/bin) WHERE (EOF > 20000 AND OWNER = dev.
Managing OSS Files Backing Up User Files In the example, this directory is named olddir and is within the current working directory. new-pathname is the relative OSS pathname of the directory to contain the copied files; this specification overrides use of the standard output file. In the example, this directory is named newdir and is the one previously created by the mkdir command within the current working directory. Restoring OSS Files Using Backup and Restore 2.
Managing OSS Files Backing Up User Files -s replstr modifies the file or archive member names named by pattern or file operands according to the substitution expression replstr, using the syntax of the ed utility. For information about the ed utility, see the ed(1) reference page either online or in the Open System Services Shell and Utilities Reference Manual. (The ed utility concepts of “address” and “line” are meaningless in the context of the pax utility and must not be used.
Managing OSS Files Backing Up User Files Creating a pax Backup of OSS Files in the Guardian File System If the output of a pax backup is less than 8 gigabytes (for H06.06 and later RVUs) or less than 2 gigabytes (for RVUs prior to H06.06), you can create the backup in the Guardian file system. If a fileset backup requires more than the size limit of a pax archive file, you can either: • • Back up sections of the fileset (each section must be smaller than the size limit of a pax archive file).
Managing OSS Files Backing Up OSS Files to Other Expand Nodes Figure 6-4 on page 6-22 shows the output of one such DSAP command. The total number of pages in a fileset on one disk is shown in the Total Pages column for its subvolume. To determine the total number of bytes, multiply the total pages by 2048 (the number of bytes/page). Figure 6-4. Output of DSAP command Subvolume Name ... ...
Managing OSS Files OSS Files and Backup/Restore Utilities (T9074) Once the pax archive exists in the Guardian file system, you use the T9074 Guardian BACKUP command to copy the archive file to a remote tape. For information about the BACKUP command, see the Guardian User’s Guide. OSS Files and Backup/Restore Utilities (T9074) OSS files should not be backed up or restored with the T9074 Backup/Restore utilities. The T9074 Guardian BACKUP utility cannot back up OSS files in file mode, only in volume mode.
Managing OSS Files Restoring User Files 9. Restore any other files backed up in Step 2. 10. Start the OSS Monitor. If necessary, use an OBEY file for the kernel subsystem SCF interface to restore its configuration in the persistence monitor and OBEY files or RESTORE to restore the configuration of the OSS environment. 11. Use RESTORE to restore the storage-pool files. The source and target system DEVICELABEL values must match. 12.
Managing OSS Files Restoring User Files Restoring Security Data Used For File Access The security database files contain user ID aliases, file-sharing groups, and initialdirectory information.
Managing OSS Files Restoring User Files Restoring Files From a pax Archive to the Guardian File System Because of the syntactic differences between the names of Guardian and OSS files, the following behaviors can occur when a pax archive member is restored to a Guardian file system: • • An OSS filename might contain characters that are illegal in Guardian filenames. As a result, the archive member cannot be created on the Guardian target and the restore operation fails.
Managing OSS Files Redirecting OSS Standard Files -f archive-name specifies the pathname of the input archive, overriding the default standard input file. Guardian tape devices can be specified with the /G naming convention (for example, /G/tape). -W clobber match-pattern is a HP extension. This flag allows selected files from an archive to be stored on a Guardian target and to overwrite any preexisting Guardian target file with the same name.
Managing OSS Files Redirecting OSS Standard Files also written to the terminal display. Data read or written through these files can be redirected using shell redirection specifications to other processes or to regular (disk) files. On a UNIX system, the pseudo-TTY feature ptty can be used as a source or sink for such data redirection; although the OSS environment does not have a pseudoTTY feature, OSSTTY can do similar data redirection.
Managing OSS Files Redirecting OSS Standard Files Figure 6-5. Redirecting Selected OSS Standard Files Telserv session ard Stand input file /G/ZAAA/#stdout OSS shell OSSTTY Standard output file Sta n da rd err /G/ZAAA/#stderr or OUT $DATA.OSSLOG. EDITFIL TE RM file $VHS VST023.
Managing OSS Files Controlling the Maximum Number of Files Figure 6-6. Redirecting All OSS Standard Files /G/ZTTY/#stdin a St nd d ar p in u $VHS ile tf IN /G/ZTTY/#stdout OSS Application Program Sta n OSSTTY Standard output file da rd err /G/ZTTY/#stderr or OUT $DATA.OSSLOG. EDITFIL TE RM file $VHS Controlling the Maximum Number of Files The fileset catalog determines the number of regular files and directories (the number of inodes) a fileset can contain.
Managing OSS Files Controlling the Maximum Number of Files FILESET Command is used. A fileset that has a only a few inodes (enforced by using a small MAXINODES value) can be recovered faster than a fileset that has a large number of inodes. To change the approximate maximum number of inodes permissible in a fileset that is not started: 1. Use the SCF INFO FILESET Command with the DETAIL option to determine the configured value for the MAXINODES attribute. 2.
Managing OSS Files Controlling the Maximum Number of Files Open System Services Management and Operations Guide— 527191-004 6- 32
7 Managing Terminal Access This section briefly describes the concepts and methods available for providing terminal users with access to the Open System Services (OSS) environment. • • • How Users Gain Access to the OSS Environment on page 7-1 Configuring Telserv Access on page 7-2 Configuring FTP Access on page 7-5 How Users Gain Access to the OSS Environment A user gains access to the OSS environment through a server process.
Managing Terminal Access • Configuring Telserv Access The way the initial working directory is configured for a user definition that is not anonymous. (The FTP user can use an FTP client quote command to gain access to the opposite environment.) See Section 8, Managing Security, for information about configuring initial working directories. Configuring Telserv Access Configuring Telserv access is fully described in the Telserv Manual.
Managing Terminal Access Configuring a Telserv Direct Service Configuring a Telserv Direct Service A Telserv process can be configured to provide login either directly to the OSS shell or directly to the command interpreter code of any executable program accessible through the OSS file system. Logging In to the Shell Directly The services feature of Telserv provides OSS users with the ability to log in directly to the OSS shell, bypassing the TACL prompt.
Managing Terminal Access • • Configuring a Telserv Direct Service HP recommends that you do not configure resilient windows for OSS logins. The default configuration for a service is to not have resilient windows, so simply omit the RESILIENT option when defining the service. Do not rely on the default value for the OSS priority in the SCF ADD SERVICE command; instead, set the priority to a value below 160.
Managing Terminal Access Configuring FTP Access Safeguard user definition, as described in How Users Gain Access to the OSS Environment on page 8-10. The default value used by the OSH program when the PARAM option does not appear places the user in the OSS shell (/bin/sh). Therefore, you must explicitly define a value for the PARAM option to start any program other than the OSS shell.
Managing Terminal Access Configuring FTP Access OSS FTP client users should have the full OSS pathname (/usr/ucb) of the FTP client program added to the PATH environment variable in their .profile files. The default value for PATH does not include the FTP pathname. Alternatively, you can add /usr/ucb to the /etc/profile file for all users (as is done in the /etc/profile.sample file supplied by HP). Section 8, Managing Security, describes how to set up the /etc/profile file.
8 Managing Security This section covers: • • • • Common and Unique Characteristics of OSS and UNIX Security on page 8-1 Managing Users and Groups on page 8-9 OSS Security Auditing on page 8-23 Protecting Your System on page 8-26 Common and Unique Characteristics of OSS and UNIX Security Basic file security is the same for the OSS environment as on a UNIX system. Files are accessed according to a file mode and access permissions, as described in the Open System Services User’s Guide.
Managing Security Administrative Files and Directories modem ttyda or ttydfa ttys0 through ttys9 These files also can be the target of UNIX security intruders. • The OSS environment does not use the following files and directories sometimes found on UNIX systems: ° In /etc: aliases dfs/dfstab exports ftpusers hosts.lpd mail/aliases or sendmail/aliases rc, rc*, or rc?.d shells syslog.conf system ttys tftpdaccess.cf ° In any of various directories: .plan .project sendmail.
Managing Security Administrative Files and Directories Access to OSS auditing logs occurs through the Safeguard audit reduction tool (SAFEART) program. UNIX directories and files such as the following are not provided and therefore do not require monitoring: ° In /var: aculog adm/acct, adm/lastlog, adm/loginlog, adm/messages, adm/pacct, adm/utmp or adm/utmpx, adm/wwtmp or adm/wtmpx spool/atrun or spool/ftp sulog, vold.log, or xferlog ° In /usr: adm/wtmp (FTP login log) etc/rpc.
Managing Security ° Administrative Tools In the user’s home directory: .netrc .rhosts Take normal security precautions with these files and directories. Administrative Tools On many UNIX systems, certain commands and utilities either exist for security administration or have security considerations. Most such commands and utilities are absent from the OSS environment.
Managing Security Administrative Tools rdist rexd rlogin or rlogind routed or gated sendmail share telnet (a TACL TELNET command exists for the Guardian environment) tftpd tip ulimit usermod /usr/etc/exportfs or /usr/etc/showmount vipw • • The GNU suite of utilities The following environment variables: maxuproc nfs_portmon • • The /bin/sh -r restricted shell option For the inetd program: ° ° -t tracing option Service of systat for port 11 Some of these commands and utilities can provide mechanisms
Managing Security Users and Groups The OSS shell supports the IFS environment variable but appropriately clears its settings. The OSS inetd program provides the following default services: time echo discard daytime chargen Users and Groups The OSS environment does not provide: • The following predefined or generic user names (accounts): agent bin daemon demo finger ftp games guest help ingres lp mail maint manager news nobody nuucp .
Managing Security Users and Groups The OSS environment does not provide common UNIX default user names and user IDs unless they are explicitly created by a site administrator. However, equivalent OSS user names and user IDs do exist. For example, the privileges normally associated with the UNIX user name root and the user ID of 0 exist for the OSS user ID (UID) of 65535 (the super ID), which is usually the user SUPER.SUPER.
Managing Security Components of OSS Security Management Components of OSS Security Management Figure 8-1 on page 8-8 illustrates the major components and interfaces of OSS security management. Figure 8-1.
Managing Security Managing Users and Groups Managing Users and Groups This subsection discusses the following user and user-group management concepts that are relevant to the OSS environment: • • • • • Differences Between OSS and UNIX User and User-Group Configuration on page 8-9 How Users Gain Access to the OSS Environment on page 8-10 User and User-Group Attributes on page 8-12 Assigning an Initial Working Directory on page 8-13 Assigning an Initial Program on page 8-17 Hints and suggestions for usin
Managing Security How Users Gain Access to the OSS Environment A NonStop operating system user can have alternate user names, called aliases. Most of the attributes of an alias can differ from those of its underlying user definition. You create an alias with the SAFECOM ADD ALIAS command. A Safeguard user group is either an administrative group or a file-sharing group. An administrative group is used to manage user access; a file-sharing group is used to manage file access.
Managing Security How Users Gain Access to the OSS Environment Both Telserv and the FTP server authenticate the user’s login information against the user definitions configured through the Safeguard subsystem. Note. The EDIT file $SYSTEM.ZTCPIP.FTPUSERS can be used to disallow access to FTP by valid users of other subsystems. When a user name appears in the FTPUSERS file, FTP rejects access without authenticating the user definition.
Managing Security User and User-Group Attributes quote OSS or quote GUARDIAN command to gain access to the other environment. User and User-Group Attributes All NonStop operating system users can use either the Guardian environment or the OSS environment. However, additional attributes should be configured for OSS environment users. In the Guardian environment as well as the OSS environment, users and user groups can have the following attributes: • A user definition can belong to up to 32 user groups.
Managing Security • • • ° Assigning an Initial Working Directory A user group can continue to exist after its last user has been removed from it. There can be up to 65,535 user groups, including 256 administrative groups. There is no limit on the number of file-sharing members (users) in a user group.
Managing Security Assigning an Initial Working Directory volume is the user’s default volume and subvolume is the user’s default subvolume in the Guardian environment. Initial working directories should not be in the Guardian file system (the /G directory), because the OSS environment creates history files in initial working directories. History files can grow indefinitely and cannot be truncated or deleted if the directory is in the /G directory.
Managing Security Assigning an Initial Working Directory or ALTER USER group-name.member-name, INITIAL-DIRECTORY /home/dir group-number,member-number is the structured view of the user ID of an existing user definition. group-name.member-name is the user name of an existing user definition. /home/dir is the initial working directory, expressed as an OSS pathname.
Managing Security Assigning an Initial Working Directory Figure 8-2. TACL Macro to Configure an OSS User (page 1 of 2) ?TACL MACRO == == This unnamed macro adds FTP user logins to the system. == == #FRAME #SET #INLINEPREFIX> INLECHO ON == == SAFECOM /INLINE/ == == == Add a user definition for the user LINDA in user group == SCRIBE. == == This definition provides the user with only Guardian access == and is appropriate for a user who wants initial FTP == access to the Guardian file system.
Managing Security Assigning an Initial Program Figure 8-2.
Managing Security Assigning an Initial Program Assigning an Initial Program Using Safeguard You can assign an initial program to an existing user definition by using the SAFECOM ALTER USER command to change the value of the INITIAL-PROGRAM attribute. You use a SAFECOM command such as the following: ALTER USER group-number,member-number, INITIAL-PROGRAM /code-dir/program or ALTER USER group-name.
Managing Security Hints and Suggestions Assigning an Initial Program Using a TACLCSTM File You can also assign an initial program to a user by either inserting the following entry in the user’s TACLCSTM file or having the user enter the following at a TACL prompt: RUN OSH -ls -prog pathname where pathname is the pathname of an existing OSS program file to be used as the initial program. The -ls specification causes execution behaviors appropriate for a UNIX shell program.
Managing Security Hints and Suggestions This method is preferable to allowing many users to access the super ID logon. The process for licensing the OSS Monitor is: 1. At a TACL prompt, set the volume to $SYSTEM.SYSnn, where nn indicates the current system subvolume. 2. Enter the following commands: FUP LICENSE OSSMON FUP SECURE OSSMON, "N-G-", PROGID This procedure allows anyone in the super group to start OSSMON but also allows OSSMON to run under the super ID.
Managing Security Hints and Suggestions Configuring FTP Users If an FTP server user requests that his or her user definition be configured for initial access to the Guardian environment: 1. Leave that user’s definition with the default null value for its OSS initial working directory. 2. Use the TACLCSTM file method to assign an OSS initial working directory for indirect Telserv access (see Assigning an Initial Working Directory Using a TACLCSTM File on page 8-17).
Managing Security • • Hints and Suggestions An alias must have a valid initial working directory (in the previous example, /user/guest): ° If no valid initial working directory is specified for an anonymous alias, then FTP access for that alias is denied. ° An initial working directory of /E, /G, or a directory in /G is invalid. The aliases anonymous and ftp must be frozen under the Safeguard product, so that those names cannot be used for access through any server process other than FTP.
Managing Security OSS Security Auditing This definition prevents use of the Guardian environment default volume and subvolume as the initial working directory. It also causes use of the OSH command to fail for that user and give the user the suggested pathname as part of an OSH error message. Note that this approach to user and alias definitions can add significantly to logs of error messages or of process failures.
Managing Security • Audit Records for OSS Objects The operation, including whether the operation succeeded or failed, and the details of a defined list of appropriate attributes Audit records are characterized by the following information: • • • • • • • If the object of an operation has a pathname, then either the audit record includes the pathname or the operation is associated with another record that includes the pathname. OSS filenames stored in the audit record are uniquely identifiable.
Managing Security Audit Records for OSS Objects The OSS name server maintains the absolute pathname of the mount point for each fileset that it manages. To ensure that they are generated quickly, all pathnames that are stored in audit records are normalized as follows: • • All dots (.), double dots (..), multiple slashes, and symbolic-link references are resolved. The maximum length of the stored pathname is 1023 bytes.
Managing Security Auditing of OSS Shell Commands For example, assume that a fileset is mounted on /usr/src/projects/mine. The following sequence of calls occur: rename("/usr/src/projects", "/usr/src/tasks"); open("/usr/src/tasks/mine/main.c"); The audit record for the open call might contain /usr/src/projects/mine/main.c (the old pathname) rather than /usr/src/tasks/mine/main.c (the new pathname).
Managing Security • Use of suid Scripts The password for the user ID of a new shell Security is enforced by requiring the user to complete a normal login procedure for the new login name. The new user ID stays in force until the shell exits. The new password stays in force until it is changed again. For super ID users, the shell substitutes a # (number sign) for its usual prompt. You can specify a login shell using /bin/sh with the appropriate environment variables.
Managing Security Preventing Security Problems suid script files have permissions that either start with the digit 4 or have “s” as the owner’s execution permission bit, as shown in the following example: $ ls -l dfile -rwxr-xr-x 1 MANUALS.HENRYP $ chmod 4755 dfile $ ls -l dfile -rwsr-xr-x 1 MANUALS.HENRYP MANUALS 29 Jun 27 15:24 dfile MANUALS 29 Jun 27 15:24 dfile You can detect suid script files by using the find command with the -perm flag, as shown in the following example: $ find . .
Managing Security • • Identifying Attempts to Break Security Do not provide network services servers such as rshd where mechanisms such as the hosts.equiv file can be used to bypass Safeguard validations. For remote procedure call (RPC) applications, encourage your developers to use the application program interface provided by the HP NonStop Distributed Computing Environment (DCE).
Managing Security Identifying Attempts to Break Security Open System Services Management and Operations Guide— 527191-004 8- 30
9 Managing With the Shell The shell is the interactive interface to the Open System Services (OSS) environment. The OSS shell is a UNIX Korn shell. This section describes how to set up the shell to best serve your users. Information about using the shell is in the Open System Services User’s Guide. Reference information is in the sh(1)reference page either online or in the Open System Services Shell and Utilities Reference Manual.
Managing With the Shell Customizing the OSS Shell through either a third-party product or the Safeguard subsystem, run from the Guardian environment, as described in Section 8, Managing Security. Customizing the OSS Shell You can customize the OSS shell by providing a custom default .profile file for your users, setting up the /etc/profile file to meet your needs, and using the localization features in both of these files. Setting Up a Default .
Managing With the Shell Setting Up an /etc/profile File Figure 9-1 shows the /etc/profile.sample file that HP provides. You might want to copy this file to /etc/profile and use it as your system /etc/profile file. (Beginning with RVU G06.14, if your system was ordered preconfigured or your initial OSS configuration was performed using the OSSSETUP utility, this copy was made automatically.) Figure 9-1. Sample /etc/profile File /etc/profile.
Managing With the Shell Setting Up an /etc/profile File UNIX methods so that the OSS environment uses values separate from those of the Guardian environment. For example, the time values used by an OSS shell default to those of the system that the shell runs on. If you maintain an OSS environment for users in a time zone other than that used for your Guardian environment, you can add the TZ environment variable to /etc/profile to make the time zone for your OSS users appropriate to their location.
Managing With the Shell Localizing Software Using the /etc/profile File Instead of a motd Command You can also use the /etc/profile file to send a message to all users when they log in. You can inform them of new features, projected downtime, or any other matter that you think they should know. To do so, include a line in the /etc/profile file such as: echo "message" where message is the message you want to send. You must include the quotation marks.
Managing With the Shell Localizing Software Table 9-1. Localization Environment Variables (page 2 of 2) Variable Meaning and default LC_MESSAGES The language for shell messages. The default value is C/POSIX. LC_MONETARY The monetary format for your shell. The default value is C/POSIX. LC_NUMERIC The numeric format for your shell. The default value is C/POSIX. LC_TIME The time format for your shell. The default value is C/POSIX.
Managing With the Shell Localizing Reference Pages Table 9-2. Locale Names and Filenames (page 2 of 2) Language of locale Filename for locale definition French, Belgium fr_BE.ISO8859-1 French, Canada fr_CA.ISO8859-1 French, France fr_FR.ISO8859-1 French, Switzerland fr_CH.ISO8859-1 German de_DE.ISO8859-1 German, Switzerland de_CH.ISO8859-1 Greek el_GR.ISO8859-7 Icelandic is_IS.ISO8859-1 Italian it_IT.ISO8859-1 Japanese, EUC ja_JP.AJEC Japanese, SJIS ja_JP.
Managing With the Shell Monitoring the OSS Environment With the Shell The directory /usr/EUC/man would contain Japanese translations of reference pages in its subdirectories.
Managing With the Shell Defragmenting Disks The find command searches for files that match criteria you set; for example, it can find all the files in a directory that have not been accessed in a specified number of days. It can then perform an action you specify, such as deleting the files it finds. For detailed information about the find command, see the find(1) reference page either online or in the Open System Services Shell and Utilities Reference Manual.
Managing With the Shell Compressing Files The syntax for the DCOM command to defragment a disk volume is: DCOM volume volume is the name of the disk volume you want to defragment. Compressing Files You can compress files to create more space on a disk. You can use: • • The pack utility, which is a POSIX.2 utility The compress utility, which provides more compression than pack Both utilities compress a file and store the specified file in a compressed form.
Managing With the Shell • • Parsing Command Options With the getopts Command Choose to log into the remote host using a specified user name rather than the local user name Specify an argument to the command you are performing remotely The remote host allows access only if at least one of the following conditions is satisfied: • • The local user ID is not the super ID, and the name of the local host is listed as an equivalent host in the remote /etc/hosts.equiv file.
Managing With the Shell Parsing Command Options With the getopts Command Open System Services Management and Operations Guide— 527191-004 9- 12
10 Managing OSS Devices This section briefly discusses Open System Services (OSS) devices and describes how to manage printers in the OSS environment. The Scope of OSS Device Management The only devices you can manage in the OSS environment are printers. All other devices are managed through the Guardian environment. In the OSS environment, printer and tape-drive definitions are not stored in the /dev directory. Do not modify the contents of the /dev directory.
Managing OSS Devices Specifying a Default Printer Printer management in the OSS environment consists of defining aliases for Guardian spooler-location names. The aliases in the /etc/printcap file provide a systemwide set of definitions. At least one alias should be provided in that file. A user can also define aliases in his or her own printcap file. When you add a new user to the system, consider creating a printcap file in the user’s initial working directory.
Managing OSS Devices Using the /etc/printcap or printcap File /etc/printcap file in Item 4, because all the other items can be omitted or accidentally removed by user actions. Using the /etc/printcap or printcap File Both the /etc/printcap and printcap files contain a list of aliases and the corresponding Guardian spooler-location names for printers. A valid /etc/printcap or printcap file must contain at least one entry to establish the default printer. In the sample file /etc/printcap.
Managing OSS Devices Using the /etc/printcap or printcap File however, an alias specified in a printcap file is not actually used as a default value in a shell command line. For example, assume that /etc/printcap contains default \COMM.$S.#DEFAULT and printcap contains default \FORTY.$S1.#BOOK If the user enters the following OSS shell command: lp file the file file is printed at the Guardian spooler-location named \COMM.$S.
11 Managing Problems Most operational problems are easily resolved by following the recovery recommendations listed in Appendix A, Messages. However, some of the messages indicate that a problem should be reported to HP. This section discusses that possibility. Problem-Reporting Procedures Your site should have a formal procedure for reporting problems detected in its own software or in HP software.
Managing Problems Gathering Version Information About OSS Files When a problem occurs with an executable file, the method you use to obtain productversion information depends on the kind of file involved.The possible methods are: • • For a product with a Subsystem Control Facility (SCF) module, such as the OSS Monitor, you can use the SCF VERSION SUBSYS, VERSION MON, and VERSION PROCESS Commands on page 12-82. For other files, you can use the Guardian VPROC utility.
Managing Problems Gathering Version Information About OSS Files This command produces a display similar to the following: VPROC-T9617D31-(14 APR 95) SYSTEM \NODE Date 18 DEC 1995, 11:21:14 COPYRIGHT TANDEM COMPUTERS INCORPORATED 1986 - 1995 /bin/sh Binder timestamp: Version procedure: Target CPU: AXCEL timestamp: 14MAR95 16:32:52 T8626D30_26MAY95_OSSUTL_AAB UNSPECIFIED 14MAR95 16:33:22 The same information appears if you use the OSS version of the vproc command, as shown in the following example: vproc
Managing Problems Gathering Version Information About OSS Files Open System Services Management and Operations Guide— 527191-004 11- 4
12 Open System Services Monitor The Open System Services (OSS) Monitor enables you to perform operations on filesets, OSS servers, and itself. Those operations are described in detail in Section 5, Managing Filesets, Section 4, Managing Servers, and Section 2, Operating the OSS Environment.
Open System Services Monitor OSS Monitor Features Online Help Facility The OSS PM includes a context-sensitive online help facility similar to that provided in other SCF product modules.
Open System Services Monitor OSS Monitor Features Figure 12-2.
Open System Services Monitor OSS Monitor Features Figure 12-4. SCF Help OSS START FILESET Display The START FILESET command makes an existing fileset available to OSS users. This action is also known as mounting a fileset. Syntax of START FILESET: START [ /OUT / ] FILESET . . . Considerations for START FILESET: . . .
Open System Services Monitor OSS Monitor Features Fault Tolerance and Continuous Availability All servers controlled through the OSS Monitor (except for the $ZTAnn transport agent server) can be run as fault-tolerant process pairs. The OSS Monitor itself can be run as a persistent process, as described in Starting the OSS Monitor as a Persistent Process on page 2-9.
Open System Services Monitor OSS Monitor SCF Command Reference Information If the OSS Monitor encounters a serious internal inconsistency, it attempts to issue the “internal error” message (message E00001). For additional information about OSS messages, see Appendix A, Messages. Subsystem and Process Attributes For the OSS Monitor: • • • • The subsystem ID is 143. The device type is 24. The device subtype is 0. The name of the program file is $SYSTEM.SYSnn.OSSMON.
Open System Services Monitor ADD FILESET Command ADD FILESET Command The ADD FILESET command adds a fileset to the configuration administered through the OSS Monitor. This command affects the contents of the ZOSSFSET file. The syntax of the ADD FILESET command is: ADD [ /OUT filename1/ ] FILESET [$ZPMON.
Open System Services Monitor ADD FILESET Command Only names not currently defined in the ZOSSFSET file are valid values. Fileset names are not case-sensitive. The $ZPMON prefix can be omitted if you have previously specified $ZPMON in an SCF ASSUME command. CATALOG volume specifies the name of the Guardian disk volume to contain the catalog files for the fileset. Disk volume names are not case-sensitive. The volume specified should be mirrored.
Open System Services Monitor ADD FILESET Command ON Safeguard global auditing controls govern which objects are audited. OFF Security-sensitive operations against this fileset are not logged in the Safeguard security audit trail. This is the default value. BUFFERED { CREATE | LOG | NONE } specifies the amount of catalog write buffering used for the fileset. CREATE When a request or transaction requires a write to the PXINODE or PXLINK file, buffer the corresponding write to the PXLOG file.
Open System Services Monitor ADD FILESET Command STOPPED No attempt is made to start the fileset when a system load occurs. No attempt is made to start the fileset during a restart of the OSS environment, unless the fileset has been manually started. This is the default value. DEVICELABEL devicelabel specifies the device label assigned to this fileset. The specified label must not already be in use by another fileset in the database.
Open System Services Monitor OSSBUFFEREDCP ADD FILESET Command Use OSS-buffered input/output with checkpointing. This behavior provides fault tolerance for single failures, with better performance than DP2BUFFEREDCP. OSS filesystem processes and DP2 share responsibility for buffering file data; OSS provides the buffering whenever possible. DP2 checkpoints the file state to its backup process to ensure recovery from single failures. If FTIOMODE is not specified, the default behavior is UNBUFFEREDCP.
Open System Services Monitor ADD FILESET Command NORMALIOMODE { UNBUFFEREDCP | DP2BUFFEREDCP | OSSBUFFEREDCP | DP2BUFFERED | OSSBUFFERED } specifies the input/output buffering and fault tolerance for application file opens that do not use the O_SYNC option: UNBUFFEREDCP Use unbuffered input/output with checkpointing. This behavior provides maximum fault tolerance but with reduced performance. DP2BUFFEREDCP Use disk-process-buffered input/output with checkpointing.
Open System Services Monitor • • • • • • ADD FILESET Command You can use the ADD FILESET command on filesets associated with an OSS name server that is not currently running. You should not add all filesets to one OSS name server. A newly added fileset has no catalog files. The catalog files (PXINODE, PXLINK, and PXLOG) are created the first time the fileset is started. The fileset READONLY attribute cannot be set when a fileset is first added.
Open System Services Monitor • ADD FILESET Command If an FTIOMODE setting of OSSBUFFEREDCP or a NORMALIOMODE setting of OSSBUFFEREDCP or OSSBUFFERED is used, the OSS filesystem buffers the data unless there are file opens from processes in more than one processor and at least one of the file opens has write permission. In that case, DP2BUFFEREDCP behavior occurs instead of OSSBUFFEREDCP behavior or DP2BUFFERED behavior occurs instead of OSSBUFFERED behavior.
Open System Services Monitor • ADD FILESET Command ° Specifies that a saved reply for an NFS request to the fileset USER2 is considered to be obsolete after 60 seconds. ° Specifies that no catalog writes are buffered. To add the fileset USER3 with the storage-pool file ZOSSPOOL using the volume catalog on $DATA2 and the mount point /user3, enter the following command: ADD /OUT $S.#USR/ FILESET $ZPMON.
Open System Services Monitor ADD SERVER Command ADD SERVER Command The ADD SERVER command adds an OSS server to the configuration administered through the OSS Monitor.
Open System Services Monitor ADD SERVER Command CPU primary_processor specifies the processor number of the processor where the primary server process will run. primary_processor is an integer in the range 0 through 15. You can specify the following options in any order: AUTORESTART ntimes specifies the persistence count of the server, where ntimes is an integer value in the range 0 through 10.
Open System Services Monitor ADD SERVER Command BACKUPCPUOK { TRUE | FALSE } specifies whether the server should be started in its configured backup processor when its configured primary processor is unavailable and the server is being restarted automatically. When the server is started in its configured backup processor and its configured primary processor subsequently becomes available, the server automatically switches to its primary processor.
Open System Services Monitor ADD SERVER Command OSS sockets local server, OSS message-queue server, and OSS transport agent servers are started by the START SERVER command. • • • You must be a member of the super group (255, nnn) to use this command. The ADD SERVER command is intended to be used for OSS name servers of filesets other than the root fileset.
Open System Services Monitor ALTER FILESET Command ALTER FILESET Command The ALTER FILESET command changes the configuration of a fileset administered through the OSS Monitor. The information entered in the command is added to or altered in the ZOSSFSET file. The syntax of the ALTER FILESET command is: ALTER [ /OUT filename1/ ] FILESET [$ZPMON.
Open System Services Monitor ALTER FILESET Command The $ZPMON prefix can be omitted if you have previously specified $ZPMON in an SCF ASSUME command. You can specify the following options in any order. You must specify at least one option: AUDITENABLED { ON | OFF } specifies whether the fileset is audited (ON) or not audited (OFF). ON Safeguard global auditing controls govern which objects are audited.
Open System Services Monitor ALTER FILESET Command CATALOG volume specifies the name of the Guardian disk volume to contain the catalog files for the fileset. Any existing catalog files are moved to this disk volume; the subvolume of the catalog files is unchanged. Disk volume names are not case-sensitive.
Open System Services Monitor ALTER FILESET Command FTIOMODE { UNBUFFEREDCP | DP2BUFFEREDCP | OSSBUFFEREDCP } specifies the input/output buffering and fault tolerance for application file opens that use the O_SYNC option: UNBUFFEREDCP Use unbuffered input/output with checkpointing. This behavior provides maximum fault tolerance but with reduced performance. DP2BUFFEREDCP Use disk-process-buffered input/output with checkpointing.
Open System Services Monitor ALTER FILESET Command The pathname described in MNTPOINT must be an existing directory, but the OSS Monitor does not validate this until an attempt is made to start the fileset. Do not include symbolic links in mount-point pathname specifications. Normalization of pathnames for mount points does not include resolution of symbolic links. If the MNTPOINT option is omitted, the mount-point directory for the fileset is not changed.
Open System Services Monitor ALTER FILESET Command DP2BUFFEREDCP Use disk-process-buffered input/output with checkpointing. This behavior provides fault tolerance for single failures, with better performance than UNBUFFEREDCP. DP2 buffers file data and checkpoints the file state to its backup process to ensure recovery from single failures. OSSBUFFEREDCP Use OSS-buffered input/output with checkpointing.
Open System Services Monitor ALTER FILESET Command If the READONLY option is omitted, the write permission granted to users of the fileset is not changed. REPORT filename3 specifies the Guardian filename for the report file to be generated when FSCK performs an automatic recovery for this fileset; filename3 must be a Guardian spooler location. If the specified spooler is unavailable, FSCK uses a file-code-180 file (a C language text file) named $SYSTEM.SYSnn.
Open System Services Monitor ALTER FILESET Command The larger the value specified for seconds2, the more likely that a fileset recovery is needed after a failure, but the faster fileset access becomes because fewer disk writes occur to update the cache from memory. The smaller the value for seconds2, the less likely that a fileset recovery is needed after a failure, but the slower fileset access becomes because more disk writes occur to update the cache from memory.
Open System Services Monitor • ALTER SERVER Command ° Assigns the fileset to the OSS name server process $ZPNS1, which has the server name #ZPNS1. ° Specifies that the OSS name server for the fileset retains the results of nonretryable Network File System (NFS) operations for 60 seconds. ° ° Specifies that files within the fileset are read-only. Causes any previously specified BUFFERED CREATE attribute to be ignored. A fileset to which the READONLY TRUE attribute is assigned is not buffered.
Open System Services Monitor ALTER SERVER Command OUT filename specifies the name of a Guardian output file for informational messages. You can either read this file with a text editor or display it with the FUP COPY command. SERVER server_processname specifies the server being altered. server_processname has the following form: [$ZPMON.]servername servername specifies the name of the server to be altered. Only names currently defined in the ZOSSSERV file are valid values.
Open System Services Monitor ALTER SERVER Command The current persistence count for each server is incremented when that server terminates abnormally or is stopped by something other than SCF. The current persistence count is not incremented because of processor failure. When the AUTORESTART option is omitted from the ALTER SERVER command, the previous value for the persistence count is unchanged.
Open System Services Monitor ALTER SERVER Command If the DESIREDSTATE option is omitted, the previous value for the server is unchanged. INODECACHE size1 specifies the number of entries permitted in the inode cache used by the OSS name server. Possible values are in the range 128 through 500000. The functional upper limit might be smaller than the theoretical maximum, because the upper limit depends upon the amount of memory available when the OSS name server is started.
Open System Services Monitor ALTER SERVER Command MAXWAITTIME seconds2 specifies the maximum number of seconds that the OSS Monitor waits for the server’s primary processor to become available during an automatic restart of the server; seconds2 must be in the range 0 through 32767. A value of 0 means that the OSS Monitor does not wait for the primary processor to become available.
Open System Services Monitor • • • • • • ALTER SERVER Command The primary and backup server processes cannot be configured in the same processor. The ALTER SERVER command can be used only by super-group users (255,nnn). For the OSS message-queue server, a BACKUPCPU value of -1 means the server is to use for its backup process the processor specified by the TACL PARAM BACKUPCPU (see Currently Used TACL PARAMs for the OSS Monitor on page 2-11).
Open System Services Monitor ALTER SUBSYS, ALTER MON, and ALTER PROCESS Commands unavailable for more than 30 seconds during an automatic restart, enter the following command: ALTER /OUT CMDLOG/ SERVER #ZPLS, BACKUPCPUOK TRUE, & MAXWAITTIME 30, AUTORESTART 3, DESIREDSTATE STARTED This command also sends informational messages to the file CMDLOG. ALTER SUBSYS, ALTER MON, and ALTER PROCESS Commands The ALTER SUBSYS, ALTER MON, and ALTER PROCESS commands all change the configuration of the OSS Monitor.
Open System Services Monitor ALTER SUBSYS, ALTER MON, and ALTER PROCESS Commands process_name must be specified unless ASSUME PROCESS $ZPMON or ASSUME $ZPMON was previously specified. You can specify the following options in any order. You must specify at least one option: AUTOSTART { AUTO | MANUAL } specifies whether the automatic startup service is enabled. AUTO All filesets and servers that are configured to use the automatic startup service are automatically restarted after a system load or restart.
Open System Services Monitor ALTER SUBSYS, ALTER MON, and ALTER PROCESS Commands REPORT [ filename2 | NULL ] specifies the target for the FSCK log file and for CVT and FUP output from SCF commands when no report file was configured for use with that fileset or the configured report file for the fileset is unavailable. The specified value must be either NULL, a blank, or the Guardian filename of a spooler location.
Open System Services Monitor CONTROL FILESET Command object configured for the service if that object failed since the last system load and is not currently started. • The AUTOSTART attribute of the subsystem can also be set by starting the OSS Monitor with a PARAM value or RUN command line parameter of AUTOSTART AUTO or AUTOSTART MANUAL. The value entered in the PARAM or command line takes precedence over the previously configured value for the subsystem. The precedence is, in descending order: 1.
Open System Services Monitor CONTROL FILESET Command FILESET [$ZPMON.]filesetname specifies the name of the OSS fileset to be altered. Only the names of currently started filesets defined in the ZOSSFSET file are valid values. Fileset names are not case-sensitive. The $ZPMON prefix can be omitted if you have previously specified $ZPMON in an SCF ASSUME command. You can specify the following options in any order.
Open System Services Monitor CONTROL SERVER Command If you omit the SYNC option, the values in use for the fileset are not updated from the ZOSSFSET file. Considerations • • • • • The CONTROL FILESET command can be used only by super-group users (255,nnn). When you use the SYNC option and the fileset has the READONLY TRUE attribute, the BUFFERED attribute value in the ZOSSFSET file is ignored and the OSS name server uses the BUFFERED NONE attribute.
Open System Services Monitor CONTROL SERVER Command The syntax of the CONTROL SERVER command is: CONTROL [ /OUT filename1/ ] SERVER server_processname { [ , RESET { ALL | MAXINODECACHEUSED | MAXLINKCACHEUSED } ] [ , SYNC ] } OUT filename1 specifies the name of a Guardian output file for informational messages. You can either read this file with a text editor or display it with the FUP COPY command. SERVER server_processname specifies the server being altered.
Open System Services Monitor DELETE FILESET Command The following OSS name server attributes are updated as soon as their ZOSSSERV values are changed with the SCF ALTER SERVER command and do not require use of the CONTROL SERVER command SYNC option: AUTORESTART BACKUPCPUOK MAXWAITTIME If you omit the SYNC option, the values in use for the server are not updated from the ZOSSSERV file. Considerations • • • The CONTROL SERVER command can be used only by super-group users (255,nnn).
Open System Services Monitor DELETE SERVER Command OUT filename specifies the name of a Guardian output file for informational messages. You can either read this file with a text editor or display it with the FUP COPY command. FILESET [$ZPMON.]filesetname specifies the name of the fileset you are deleting. Only names currently defined in the ZOSSFSET file are valid values. The $ZPMON prefix can be omitted if you have previously specified $ZPMON in an SCF ASSUME command.
Open System Services Monitor DIAGNOSE FILESET Command servername specifies the name of the server to delete. Only names currently defined in the ZOSSSERV file are valid values. The first character of the name must be a pound sign (#). Server names are not case-sensitive. The $ZPMON prefix can be omitted if you have previously specified $ZPMON in an SCF ASSUME command. Considerations • • • • You must be in the super group (255, nnn) to use this command.
Open System Services Monitor DIAGNOSE FILESET Command OUT filename1 specifies the name of a Guardian output file for informational messages. You can either read this file with a text editor or display it with the FUP COPY command. FILESET [$ZPMON.]filesetname is the name of the fileset you are diagnosing. Only names currently defined in the ZOSSFSET file are valid values. The $ZPMON prefix can be omitted if you have previously specified $ZPMON in an SCF ASSUME command.
Open System Services Monitor DIAGNOSE FILESET Command STATUS reports any inconsistencies that the OSS name server has detected in the fileset and places the report in the FSCK log file. When the STATUS option is specified, you cannot specify the REPAIR, UPGRADE, or DOWNGRADE option. When the STATUS, REPAIR, UPGRADE, and DOWNGRADE options are all omitted, the default value is REPAIR NONE. REPAIR { ALL | NONE | OPEN | SERIOUS } specifies which inconsistencies to repair in the fileset.
Open System Services Monitor DIAGNOSE FILESET Command When the UPGRADE option is specified, you cannot specify the STATUS, REPAIR, or DOWNGRADE option. DOWNGRADE reformats the catalog for the fileset from a current version to an earlier version. For example, you can downgrade a D40-version catalog to a D30-version catalog to remove support for symbolic links. Caution. When you downgrade a fileset’s catalog to a D30 version, all symbolic links in that fileset are removed.
Open System Services Monitor INFO FILESET Command SCF STOP FILESET command to stop (unmount) the fileset and then enter the following command: DIAGNOSE /OUT CMDLOG/ FILESET $ZPMON.USER1,DETAIL,REPAIR ALL • To stop the integrity check in progress on the fileset USER1 and send informational messages to the file CMDLOG, enter the following command: DIAGNOSE /OUT CMDLOG/ FILESET $ZPMON.
Open System Services Monitor INFO FILESET Command FILESET [$ZPMON.]filesetname is the name of the fileset whose information is to be displayed. Only names currently defined in the ZOSSFSET file are valid values. filesetname can contain wildcard characters. (See Using Wildcard Characters in OSS Monitor Commands: on page 2-13 for the definition of UNIX wildcard characters.) The $ZPMON prefix can be omitted if you have previously specified $ZPMON in an SCF ASSUME command.
Open System Services Monitor INFO FILESET Command directory is the OSS pathname of the mount point for the fileset. This mount point might not appear to be the one most recently specified for the fileset. The value stored and displayed is the result of transforming the specified pathname to its minimum absolute form.
Open System Services Monitor INFO FILESET Command choice indicates the amount of catalog write buffering used for the fileset. Value Meaning CREATE Buffer catalog writes during file creation within the fileset. This is the fast-create option. LOG NONE Buffer only log file activity. Do not buffer anything. kbytes is the number of kilobytes that the OSS name server uses for buffers for Network File System (NFS) operations for the fileset.
Open System Services Monitor INFO FILESET Command time2 is the timestamp for the last time that the fileset configuration was changed, in the form dd mmm yyyy hh:mm:ss.mil. filesetname is the name of the fileset on which this fileset is mounted. servername2 is the server name of the OSS name server for the fileset identified by filesetname. directory is the OSS pathname of the mount point for the fileset. This mount point might not appear to be the one most recently specified mount point for the fileset.
Open System Services Monitor INFO SERVER Command ftiomode is the fault-tolerance and buffering attribute to be used when file opens use O_SYNC in the fileset. The value displayed is one of the keywords described for the FTIOMODE attribute of the ADD FILESET Command on page 12-7. normaliomode is the fault-tolerance and buffering attribute to be used when file opens do not use O_SYNC in the fileset.
Open System Services Monitor INFO SERVER Command OUT filename specifies the name of a Guardian output file for informational messages. You can either read this file with a text editor or display it with the FUP COPY command. SERVER server_processname specifies the server whose information is to be displayed. server_processname has the following form: [$ZPMON.]servername servername specifies the name of the server whose information is to be displayed.
Open System Services Monitor INFO SERVER Command server-devicename is the server name of the server whose information is displayed, shown as a device of the OSS Monitor process. servername is the server name of the server whose information is being displayed. type is the type of the server. Value Meaning LOCAL The server is an OSS sockets local server for AF_UNIX sockets. NAME The server is an OSS name server. MSGQ The server is an OSS message-queue server.
Open System Services Monitor INFO SERVER Command server-devicename is the server name of the server whose information is displayed, shown as a device of the OSS Monitor process. type is the type of the server. Value Meaning LOCAL The server is an OSS sockets local server for AF_UNIX sockets. NAME The server is an OSS name server. MSGQ The server is an OSS message-queue server. TAGENT The server is an OSS transport agent server.
Open System Services Monitor INFO SERVER Command size1 is the number of entries permitted in the inode cache used by the OSS name server. This entry is displayed only for OSS name server processes. size2 is the number of entries permitted in the link cache used by the OSS name server. This entry is displayed only for OSS name server processes. seconds is the number of seconds that the OSS name server waits for a response from SQLCAT to a request. This entry is displayed only for OSS name server processes.
Open System Services Monitor INFO SUBSYS, INFO MON, and INFO PROCESS Commands seconds specifies the number of seconds that the OSS Monitor waits for the server’s primary processor to become available during an automatic restart. Valid values are 0 through 600. state specifies the desired state of the server after the next time the automatic startup service is used. Valid values are STARTED or STOPPED.
Open System Services Monitor INFO SUBSYS, INFO MON, and INFO PROCESS Commands The syntax of the INFO SUBSYS, INFO MON, and INFO PROCESS commands is: INFO [ /OUT filename/ ] object-type [ process ] [ , DETAIL ] OUT filename specifies the name of a Guardian output file for informational messages. You can either read this file with a text editor or display it with the FUP COPY command. object-type is one of the following: MON | PROCESS | SUBSYS MON, PROCESS, and SUBSYS all specify the OSS Monitor.
Open System Services Monitor INFO SUBSYS, INFO MON, and INFO PROCESS Commands process_name is the process name of the subsystem whose information is displayed, $ZPMON. seconds is the number of seconds that the OSS Monitor waits for a response from an OSS name server to an OSS Monitor request.
Open System Services Monitor INFO SUBSYS, INFO MON, and INFO PROCESS Commands no processor is configured for the fileset or the processor configured for the fileset is unavailable. filespec is the Guardian filename or spooler location of the default FSCK log file. volume is the name of the Guardian disk volume that contains the program files for the CVT utility, the OSS Monitor, and other OSS components.
Open System Services Monitor NAMES Command A display such as the following is written to CMDLOG: OSS Detailed Info SUBSYS \NODE1.$ZPMON *IOTimeout............. *FsckCPU............... *Report ............... *ZOSSVol............... CreateBy.............. CreateTime............ AlterBy............... AlterTime............. *AutoStart............. 60 1 $SYSTEM.MONDATA.FSCKOUT $SYSTEM.ZOSS $ZPMON 12 Dec 2001 12:48:28.554 SUPER.SUPER 03 Mar 2002 01:19:11.
Open System Services Monitor NAMES Command SEL [ NOT ] state directs the command to apply only to objects that are in the specified state or, when the NOT field is used, only to objects that are not in the specified state. This option is valid only if an object type is provided. For the SERVER object type, valid values for state are: Value Meaning STARTED STOPPED The server is started. The server is ready to be started.
Open System Services Monitor RENAME FILESET Command Considerations • • • The SEL option applies only to FILESET and SERVER objects. The OSS Monitor version of the NAMES command does not support the SUB option that can appear in other SCF NAMES commands. The OSS Monitor does not have subordinate objects. In the SCF object hierarchy, SUBSYS is the highest of the possible objects for this command.
Open System Services Monitor START FILESET Command The $ZPMON prefix can be omitted if you have previously specified $ZPMON in an SCF ASSUME command. Considerations • • Only super-group users (255,nnn) can use the RENAME FILESET command. You can rename a running fileset, and the new name takes effect immediately. Example To change the name of fileset USER1 to NEWUSER1, enter the following SCF command: RENAME FILESET $ZPMON.USER1, $ZPMON.
Open System Services Monitor START SERVER Command MAXINODES, the OSS name server changes the value of MAXINODES to the minimum of 2200000 or 110% of the number of inodes in use, rounded up to the nearest thousand. The new value for MAXINODES is stored in the ZOSSFSET file. This algorithm allows a fileset to be mounted and a reasonable number of files created in it before its limit is reached and operator intervention is required.
Open System Services Monitor STATUS FILESET Command Considerations • • You can use the START SERVER command to start the OSS sockets local server, the OSS message-queue server, and the OSS transport agent servers. An OSS name server is automatically started when one of the filesets managed by that OSS name server is started. The START SERVER command can be used only by super-group users (255,nnn).
Open System Services Monitor STATUS FILESET Command SEL [ NOT ] state directs the command to apply only to filesets that are in the specified state or, when the NOT field is used, only to filesets that are not in the specified state. Valid values for state are: Value Meaning DIAGNOSING STARTED STOPPED UNKNOWN The fileset is being diagnosed by the FSCK utility. The fileset is started (mounted). The fileset is ready to be started (mounted) or diagnosed. The fileset is in an unknown state.
Open System Services Monitor STATUS FILESET Command The information returned for a detailed report has this format: OSS Detailed Status FILESET fileset-devicename State.................................. state MountTime.............................. time1 LastError.............................. error ErrorDetail............................ error-detail ErrorTime.............................. time2 FsckName............................... volume LastControlSyncTime....................
Open System Services Monitor STATUS FILESET Command time2 is the timestamp for the time when the last error for the fileset was reported, in the form dd mmm yyyy hh:mm:ss.mil. If no error has been reported, this value is the same as the time1 value. status1 indicates whether the configuration of the fileset has changed after its most recent start or synchronization with ZOSSFSET: Value Meaning FALSE The fileset configuration has not changed. TRUE The fileset configuration has changed.
Open System Services Monitor STATUS FILESET Command type indicates the type of buffering performed for the fileset: Value Meaning CREATE When a request or transaction requires a write to the PXINODE or PXLINK file, the corresponding write is buffered to the PXLOG file. Also uses the fast file-creation feature for writing new file labels. When a request or transaction requires a write to the PXINODE or PXLINK file, the corresponding write is buffered to the PXLOG file.
Open System Services Monitor STATUS FILESET Command number3 is the current number of inodes that the OSS name server is using for the fileset, displayed as a decimal number and as a percentage of the maximum number allowed. This field is displayed only if the fileset is in the STARTED state. number4 is the last highwater mark of inode use in the cache for the fileset, displayed as a decimal number and as a percentage of the maximum number allowed.
Open System Services Monitor STATUS FILESET Command The following fields are displayed only when the fileset is in the STARTED state: volume_list2 lists the names of the disk volumes in the fileset’s storage pool that have space available for new file creation. If the fileset is mounted with the READONLY attribute set to TRUE, this field is blank. volume_list3 lists the names of the disk volumes in the fileset’s storage pool that do not have space available for new file creation.
Open System Services Monitor STATUS FILESET Command Considerations • The DETAIL display for an unmounted fileset differs from that of a mounted (started) fileset. If a fileset is not in the STARTED state, its catalog usage information, disk configuration, disk status information, and fileset attributes are not displayed. Examples • To determine the status of the fileset USER1 and send informational messages to the file CMDLOG, enter the following command: STATUS /OUT CMDLOG/ FILESET $ZPMON.
Open System Services Monitor STATUS FILESET Command A display such as the following appears: OSS Detailed Status FILESET \NODE1.$ZPMON.USER1 State.................................. STARTED MountTime.............................. 14 Nov 2004, 14:25:52.396 LastError.............................. 0 ErrorDetail............................ 0 ErrorTime.............................. 22 Nov 2004, 16:25:38.835 FsckName............................... $DATA LastControlSyncTime....................
Open System Services Monitor STATUS SERVER Command STATUS SERVER Command The STATUS SERVER command displays status information about a server administered by the OSS Monitor. The syntax of the STATUS SERVER command is: STATUS [ /OUT filename/ ] SERVER server_processname [ , DETAIL ] [ , SEL [ NOT ] state ] OUT filename specifies the name of a Guardian output file for informational messages. You can either read this file with a text editor or display it with the FUP COPY command.
Open System Services Monitor STATUS SERVER Command SEL [ NOT ] state directs the command to apply only to servers that are in the specified state or, when the NOT field is used, only to servers that are not in the specified state. Valid values for state are: Value Meaning STARTED STOPPED The server is started. The server is ready to be started.
Open System Services Monitor STATUS SERVER Command The information returned for a detailed report has this format: OSS Detailed Status SERVER server-devicename State.................. StartTime.............. LastError.............. ErrorDetail............ ErrorTime.............. LastControlSyncTime.... AlterAfterSyncOrStart.. *CPU.................... *BackupCPU.............. *SQLTimeout............. *InodeCache............. InodeCacheInUse........ MaxInodeCacheUsed...... TimeMaxInodeCacheUsed..
Open System Services Monitor STATUS SERVER Command time2 is the timestamp for the time when the last error for the server was reported, in the form dd mmm yyyy hh:mm:ss.mil. If no error has been reported, this value is the same as the time1 value. The following fields are displayed only when an OSS name server is in the STARTED state: time3 is the timestamp for the time when the OSS name server was last synchronized with ZPOSFSERV, in the form dd mmm yyyy hh:mm:ss.mil.
Open System Services Monitor STATUS SERVER Command time5 is the timestamp for the time when the OSS name server last reached its highwater mark for inode cache use, in the form dd mmm yyyy hh:mm:ss.mil. time6 is the timestamp for the time when the OSS name server last had its highwater mark for inode cache use reset, in the form dd mmm yyyy hh:mm:ss.mil. size4 indicates the number of links that the OSS name server is allowed to cache.
Open System Services Monitor STOP FILESET Command A display similar to the following is returned: OSS Detailed Status SERVER \NODE1.$ZPMON.#ZPNS State....................... STARTED StartTime................... 25 Nov 2002, 10:23:20.123 LastError................... 0 ErrorDetail................. 0 ErrorTime................... 25 Nov 2002, 10:23:20.123 LastControlSyncTime........ AlterAfterSyncOrStart...... *CPU........................ *BackupCPU.................. *SQLTimeout................. *InodeCache.
Open System Services Monitor STOP SERVER Command The $ZPMON prefix can be omitted if you have previously specified $ZPMON in an SCF ASSUME command. Considerations • • • The STOP FILESET command does not close any open OSS files or directories on the affected fileset. The STOP FILESET command can be used only by super-group users (255,nnn). A fileset cannot be stopped while other filesets are mounted on it.
Open System Services Monitor VERSION SUBSYS, VERSION MON, and VERSION PROCESS Commands servername can contain wildcard characters. (See Using Wildcard Characters in OSS Monitor Commands: on page 2-13 for the definition of UNIX wildcard characters.) The $ZPMON prefix can be omitted if you have previously specified $ZPMON in an SCF ASSUME command. Considerations • • The STOP SERVER command can be used to stop the OSS sockets local server, the OSS message-queue server, and the OSS transport agent servers.
Open System Services Monitor VERSION SUBSYS, VERSION MON, and VERSION PROCESS Commands object-type specifies the types of objects whose version information is to be returned. object-type has the following form: MON | PROCESS | SUBSYS MON, PROCESS, and SUBSYS all specify the OSS Monitor. If object-type is omitted, processname must be specified unless ASSUME PROCESS $ZPMON was previously specified. processname specifies the name of the process whose version is to be displayed.
Open System Services Monitor VERSION SUBSYS, VERSION MON, and VERSION PROCESS Commands The format of a detailed VERSION display is: Detailed VERSION process-filename [ SYSTEM system-name ] subsystem-name1 version-information-banner [ GUARDIAN NonStop-OS-version-information-banner SYSTEM system-name subsystem-name1 version-information-banner GUARDIAN NonStop-OS-version-information-banner SCF KERNEL version-information-banner subsystem-name2 PM version-information-banner ] process-filename is the name of
Open System Services Monitor VERSION SUBSYS, VERSION MON, and VERSION PROCESS Commands Examples • The command VERSION $ZPMON produces output such as the following: VERSION SUBSYS \NODE.$ZPMON: OSS Monitor-T8622D46-01SEP98-AAF where the information after the colon (:) is the subsystem name and productversion information associated with $ZPMON. • The command VERSION SUBSYS $ZPMON, DETAIL produces output similar to the following: Detailed VERSION \SIERRA.
Open System Services Monitor VERSION SUBSYS, VERSION MON, and VERSION PROCESS Commands Open System Services Management and Operations Guide— 527191-004 12 -86
A Messages Status, warning, and error messages are sent to the console by: • • • • • • The Open System Services (OSS) EasySetup utilities. These messages are described under OSS EasySetup Utility Messages on page A-2. The Guardian Catalog Volume Tool (CVT) utility. These messages are described under CVT Messages on page A-3. The FSCK fileset integrity checker when the Subsystem Control Facility (SCF) DELETE FILESET or DIAGNOSE FILESET command is used.
Messages OSS EasySetup Utility Messages OSS EasySetup Utility Messages The following console messages can appear on the terminal that issued an EasySetup command. The OSS EasySetup utilities can also generate EMS events, which are issued under the OSS subsystem ID. For information about these messages, see the Operator Messages Manual. WARNING - Event definitions file (filename) not loaded because file could not be found filename indicates the Guardian file that cannot be found, which is usually $SYSTEM.
Messages CVT Messages CVT Messages The following warning and error messages appear on the terminal that issued the CVT command. The Guardian CVT utility does not generate Event Management Service (EMS) events. CVT Warning Message CVT issues the following warning message. Warning:filename - No such File filename is a Guardian filename used in the command. Cause. The named file does not exist. Effect. CVT continues. Recovery. Check the command and correct it as necessary.
Messages CVT Error Messages Cause. The Guardian FILE_OPEN_ procedure call could not be completed. Effect. CVT did not open the indicated file. Recovery. Check the command and correct it as necessary. FILE_PURGE_ Error error on file filename error is a Guardian file-system error value returned by the underlying Guardian FILE_PURGE_ procedure call. See the Guardian Procedure Errors and Messages Manual for information about the indicated error. filename identifies the affected file. Cause.
Messages CVT Error Messages Cause. There is a PXCKSTAT file on one of the subvolumes involved in processing the command. CVT does not move catalogs to, from, or within a subvolume that contains a PXCKSTAT file—nor does it purge a saved catalog from such a subvolume. Effect. CVT terminates. Recovery. Run the FSCK utility against the catalog before attempting the CVT command again. ***Incomplete Command*** Cause. The command entered was incomplete. Effect. CVT terminates. Recovery.
Messages FSCK Messages Effect. CVT terminates. Recovery. List the files in the subvolume to see the FSN values used in their Guardian file IDs. Try the command again with an FSN from those shown in the file IDs. See Generated Catalog Files on page 5-33 for more information about FSN values. ***Invalid Subvolume Name*** - token token identifies the invalid subvolume name. Cause. The value token followed the keyword IN in the command line, but it is not a valid Guardian subvolume name. Effect.
Messages FSCK Consistent-Fileset Messages FSCK Consistent-Fileset Messages If the FSCK utility finds no relevant inconsistencies or errors in a fileset, it issues one of the following messages (listed in alphabetic order). After issuing one of these messages, FSCK clears the mounted flag and the bitmap indicating the types of inconsistencies previously detected by the OSS name server. *** CATALOG AND ZYQ SUBVOLUMES HAVE BEEN PURGED *** Cause. FSCK successfully deleted a fileset. Effect. None. Recovery.
Messages FSCK Inconsistency and Error Messages SERIOUS These messages are issued when FSCK detects a serious inconsistency. A catalog integrity problem exists. WARNING These messages are issued when a recoverable error occurs or when FSCK takes an action (such as purging an empty orphan ZYQ file) that is not reported as an inconsistency. WARNING messages report conditions that FSCK encounters or actions that FSCK takes that do not by themselves indicate problems with the catalog.
Messages FSCK Inconsistency and Error Messages Cause. The FSCK utility encountered a file-management error on an operation (given by operation) on the indicated file because of the error described by errornumber and description. Some common examples of this message are: Message Meaning OPEN Error 11 (Record Not Found) on PXINODE There is no OSS fileset catalog in the specified (or implied) catalog subvolume. The ZOSSFSET file specifies an incorrect volume.
Messages FSCK Inconsistency and Error Messages 8 WARNING 8 - FSCK Run Number nnnn was Interrupted nnnn identifies the FSCK serial number (FSN) of the affected invocation of FSCK. Cause. The indicated invocation of FSCK was interrupted before it finished. Effect. A new FSN is not assigned. If FSCK did not finish creating a new catalog, the existing saved catalog files (PXINnnnn, PXLInnnn, and PXLOnnnn) are used.
Messages FSCK Inconsistency and Error Messages 15 WARNING 15 - Catalog Already Downgraded Cause. The FSCK utility was asked to downgrade a catalog to the previous format, but the catalog is already in that format. Effect. FSCK does not downgrade the catalog. Recovery. Informational message only; no corrective action is needed. 16 WARNING 16 - Dirty Catalog using Fast Create; REPAIR ALL will be performed Cause.
Messages FSCK Inconsistency and Error Messages 99 ERROR 99 - HEAP OVERFLOW Cause. The FSCK utility has exhausted its internal memory. Effect. FSCK terminates abnormally. Recovery. Set the SAVEABEND attribute of the FSCK program file using the nld command. Rerun FSCK to produce a saveabend file.
Messages FSCK Inconsistency and Error Messages If the REPAIR SERIOUS, REPAIR OPEN, or REPAIR ALL option was specified in the command, this is an informational message only; no corrective action is needed. 101 SERIOUS 101 - Duplicate Link ID - Parent:parent, Child:child, Name:link linkname parent indicates the inode number of the parent end of the link. child indicates the inode number of the child end of the link. linkname identifies the affected link. Cause.
Messages FSCK Inconsistency and Error Messages Effect. The indicated record is added to the new PXLINK file. The missing link will be for one of the reserved inodes (/, /G, /E, /lost+found, /dev, /dev/tty, /dev/null, or the superblock). Recovery. If the REPAIR option was not specified or the REPAIR NONE option was specified in the Subsystem Control Facility (SCF) DIAGNOSE FILESET command, the catalog is still in an inconsistent state. Reissue the command with the REPAIR SERIOUS or REPAIR OPEN option.
Messages FSCK Inconsistency and Error Messages catalog is still in an inconsistent state. Reissue the command with the REPAIR SERIOUS or REPAIR OPEN option. If the REPAIR SERIOUS, REPAIR OPEN, or REPAIR ALL option was specified in the command, this is an informational message only; no corrective action is needed. 202 SERIOUS 202 - Corrupt Inode, Inode=inode-number inode-number indicates the affected inode. Cause. The FSCK utility has detected a corrupt record in the PXINODE file. Effect.
Messages FSCK Inconsistency and Error Messages 205 MINOR 205 - Missing ZYQ File, Inode=inode-number inode-number indicates the affected inode. Cause. The indicated regular file inode has no corresponding ZYQ file. Effect. The inode and any links to it are omitted from the new catalog. Recovery.
Messages FSCK Inconsistency and Error Messages Recovery. If the REPAIR option was not specified or the REPAIR NONE option was specified in the Subsystem Control Facility (SCF) DIAGNOSE FILESET command, the catalog is still in an inconsistent state. Reissue the command with the REPAIR ALL option. If the REPAIR SERIOUS, REPAIR OPEN, or REPAIR ALL option was specified in the command, this is an informational message only; no corrective action is needed.
Messages FSCK Inconsistency and Error Messages Cause. A record with the indicated invalid inode number as a key was found in the current PXINODE file. The inode number is less than or equal to zero or is greater than 231. Effect. The record is omitted from the new PXINODE file. Recovery. If the REPAIR option was not specified or the REPAIR NONE option was specified in the Subsystem Control Facility (SCF) DIAGNOSE FILESET command, the catalog is still in an inconsistent state.
Messages FSCK Inconsistency and Error Messages Cause. The directory graph forms a loop that includes the indicated inode. The inode is linked through other inodes back to itself; this means that the directory referred to by the inode number is its own parent. Effect. The inode is unlinked and placed in the OSS /lost+found directory. Recovery.
Messages FSCK Inconsistency and Error Messages If the REPAIR ALL option was specified in the command, this is an informational message only; no corrective action is needed. 214 MINOR 214 - Catalog/File Label Mismatch, Inode=inode-number inode-number indicates the affected inode. Cause. For the indicated inode, either of the following is true: 1. The disk process’s copy of the number of links to the file is incorrect. 2.
Messages FSCK Inconsistency and Error Messages 216 MINOR 216 - ZYQ File Conflict - filename filename indicates the affected file. Cause. The inode corresponding to the indicated file does not refer to that file. Effect. A currently unused inode is allocated, the ZYQ file is renamed to correspond to the new inode, and the inode is placed in the OSS /lost+found directory. Recovery.
Messages FSCK Inconsistency and Error Messages 219 WARNING 219 - Root Fileset Cause. The FILESET ROOT option was not specified in the Subsystem Control Facility (SCF) DIAGNOSE FILESET command, but the fileset catalog is a root catalog. Effect. The FSCK utility does not check for the existence of those files that must be in a root fileset. Recovery. Informational message only; no corrective action is needed. 220 WARNING 220 - The Inode Table has Overflowed to Disk Cause.
Messages FSCK Inconsistency and Error Messages 222 WARNING 222 - Catalog will be converted from up-level format Cause. A catalog at a higher version than the current version is being repaired (for example, a catalog in D40 format is being repaired by a D30.02 release version update, or RVU, copy of the FSCK utility). Effect. A new catalog is produced that is suitable for mounting by an OSS name server from the same software RVU as the FSCK utility that issued the message. Recovery.
Messages FSCK Inconsistency and Error Messages Recovery. If the REPAIR ALL option was not specified in the Subsystem Control Facility (SCF) DIAGNOSE FILESET command, the catalog is still in an inconsistent state. Reissue the command with the REPAIR ALL option. If the REPAIR ALL option was specified in the command, this is an informational message only; no corrective action is needed. 302 ERROR 302 - Invalid or Corrupt PXCKSTAT File Cause.
Messages FSCK Inconsistency and Error Messages 304 WARNING 304 - filename Purged filename indicates the affected file. Cause. The indicated Guardian disk file is either an empty orphan ZYQ file or a file belonging to an incomplete catalog left behind when an earlier execution of FSCK was interrupted. Effect. The indicated file is purged and FSCK proceeds with its checking. Recovery. Informational message only; no corrective action is needed.
Messages FSCK Inconsistency and Error Messages 307 ERROR 307 - Can’t Create SQLCAT Process - PROCESS_CREATE_ Error error:detail error:detail is the error and error detail returned by the Guardian PROCESS_CREATE_ procedure. For further information, see the Guardian Procedure Errors and Messages Manual. Cause. The FSCK utility detected a file that contains SQL/MP data but failed to start the SQLCAT process needed to delete the file. The failure might be caused by a temporary shortage of system resources.
Messages • OSS Monitor Messages Perform the following steps: 1. Use the Subsystem Control Facility (SCF) START FILESET command to remount the fileset containing the affected file. 2. Use the File Utility Program (FUP) INFO, DETAIL command or the OSS pname utility to determine the pathname of the affected file. 3. Have a user with update authority for the SQL catalog delete the file using the OSS rm utility. 4. Use the SCF STOP FILESET command on the affected fileset. 5.
Messages Unnumbered Messages processname is the name of the failed process. Unless otherwise indicated in the specific message description, this name is always $ZPMON. interval is the elapsed time since the start of the failed process, in the form dd mmm yyyy hh:mm:ss.mil. code is the SPI error code of the failure. Values for this are indicated in the specific message descriptions. release is the major release version identifier for the failed process.
Messages Unnumbered Messages OSS Monitor failed in adding converted record to filename file -- Error: err filename is the Guardian filename of the affected OSS configuration file. err is the Guardian file-system error number that describes the failure. Cause. During automatic conversion of the OSS configuration files from a previous release version update (RVU), the OSS Monitor found an entry in the ZPCONFIG file that could not be written to the indicated configuration file. Effect.
Messages Unnumbered Messages filename is the Guardian filename of the affected OSS configuration file. err is the Guardian file-system error number that describes the failure. Cause. The OSS Monitor was unable to add a default record to the indicated OSS configuration file, during the initial startup of a database conversion from the D40 product. (For example, the entry for the root OSS name server could not be added to the ZOSSSERV file.) Effect.
Messages Unnumbered Messages err is the Guardian file-system error number that describes the failure. Cause. The OSS Monitor could not create the indicated OSS configuration file. This message might indicate that the automatic conversion of the configuration files from a previous release version update (RVU) has failed. Effect. The OSS Monitor process terminated abnormally with the SPI error code value 17. The OSS environment is not available. Recovery. Perform the following steps: 1.
Messages Unnumbered Messages OSS Monitor failed in opening filename file -- Error: err filename is the Guardian filename of the affected OSS configuration file. err is the Guardian file-system error number that describes the failure. Cause. The OSS Monitor could not open the indicated OSS configuration file. This message might indicate that the configuration file is corrupted. Effect. The OSS Monitor process terminated abnormally with the SPI error code value 17. The OSS environment is not available.
Messages Unnumbered Messages OSS Monitor failed in reading filename file -- Error: err filename is the Guardian filename of the affected OSS configuration file. err is the Guardian file-system error number that describes the failure. Cause. The OSS Monitor could not read the indicated OSS configuration file. This message might indicate that the configuration file is corrupted. Effect. The OSS Monitor process terminated abnormally with the SPI error code value 17. The OSS environment is not available.
Messages Unnumbered Messages OSS Monitor failed to get its process name Cause. The OSS Monitor could not determine its own process name. An internal error might have occurred. Effect. The OSS Monitor process terminated abnormally with the SPI error code value -7. The OSS environment is not available. Recovery. Restart the OSS Monitor with the process name $ZPMON. If the problem persists, check the Event Management Service (EMS) log for related messages.
Messages Numbered Messages 4. Restart the OSS Monitor to restart the database creation process: OSSMON /NAME $ZPMON, TERM $ZHOME, NOWAIT/ Process name of OSS Monitor must be $ZPMON Cause. The OSS Monitor was started with the name shown as processname in the SPI termination information for this message. Effect. The OSS Monitor process terminated abnormally with the SPI error code value -33. The OSS environment is not available. Recovery.
Messages Numbered Messages 1 OSS E00001 Internal error Cause. The OSS Monitor has detected a serious internal error during the execution of a Subsystem Control Facility (SCF) command. Effect. The OSS Monitor stops processing the command. Recovery. Check the Event Management Service (EMS) log for related messages.
Messages • Numbered Messages Examine the load on the processor that is running the primary OSS name server process, then do one or more of the following: ° ° ° Increase the execution priority of the OSS name server process. Start the OSS name server on a less-busy processor. Increase the cache size for the OSS name server and restart the server. 3 OSS E00003 Failed to start process -- error err, error-detail process is the process name of the OSS subsystem process that failed to start.
Messages Numbered Messages Cause. A Subsystem Control Facility (SCF) ADD or ALTER command contained a parameter value that was invalid. Effect. The command is not processed. Recovery. Reissue the command with a valid value. 6 OSS E00006 No attributes have been specified for this command. Cause. A Subsystem Control Facility (SCF) command contained a parameter without a required value or was missing a required parameter. Effect. The command is not processed. Recovery.
Messages Numbered Messages 9 OSS E00009 Failed to start fileset filesetname filesetname identifies the affected fileset. Cause. A Subsystem Control Facility (SCF) START FILESET command failed. Either another fileset is mounted on the mount point you want to use for the indicated fileset, or the fileset containing the mount point you want to use is not started. Effect.
Messages Numbered Messages 3. Enter an SCF DIAGNOSE FILESET command for the indicated fileset. 4.
Messages Numbered Messages Cause. A Subsystem Control Facility (SCF) START FILESET command was issued for a fileset whose mount point does not represent a valid OSS pathname. (The pathname does not exist or is not a directory.) Effect. The fileset is not started. Recovery. Perform the following steps: 1. Use the SCF INFO FILESET command to check the spelling of the mount point pathname. 2. Use the OSS shell ls command to verify that an OSS directory exists to serve the mount point. 3.
Messages Numbered Messages 17 OSS E00017 Unable to access configuration file filename -error err filename is the Guardian filename of the OSS configuration file that could not be opened, read, or written. err is the Guardian file-system error number returned by the input or output operation. Cause. The OSS Monitor could not open, read from, or write to a configuration file. Effect. No processing can be performed on the objects in the affected configuration file.
Messages Numbered Messages Effect. The command is not processed. Recovery. A serious problem exists. Perform the following actions: 1. If possible, check and correct the configuration of the fileset using the Subsystem Control Facility (SCF) INFO FILESET and ALTER FILESET commands. Reassigning an existing attribute to the fileset might fix the configuration record problem. 2. Restart the fileset. If the problem persists, check the Event Management Service (EMS) log for related messages.
Messages Numbered Messages 22 OSS E00022 Invalid value for attribute attribute-name attribute-name is the name of the affected attribute. Cause. A Subsystem Control Facility (SCF) ADD or ALTER command failed because a specified attribute has a value that is invalid. Effect. The command is not processed. Recovery. Reissue the command using a valid value for the attribute. 23 OSS E00023 Name Server server rejected the request server is the server name of the affected OSS name server. Cause.
Messages Numbered Messages 24 OSS E00024 Name Server server gave unexpected response to OSS Monitor: status. server is the server name of the affected OSS name server. status is the status value returned to the OSS Monitor. Cause. An OSS name server returned status information that is not recognized by the OSS Monitor. An OSS name server process might be malfunctioning. Effect. The command cannot be processed. Recovery.
Messages • • • • Numbered Messages A copy of the OSS fileset catalog files Description of the problem and accompanying symptoms Details from the message or messages generated Supporting documentation such as Event Management Service (EMS) logs, trace files, and a processor dump, if applicable If your local operating procedures require contacting the Global Customer Support Center (GCSC), supply your system number and the numbers and versions of all related products as well.
Messages Numbered Messages 28 OSS E00028 Failed in moving catalog files Cause. A Subsystem Control Facility (SCF) ALTER FILESET command was issued to change the catalog volume of a fileset, but the catalog files could not be moved to the specified volume.
Messages Numbered Messages b. Use the SCF INFO SUBSYS command to check the disk volume being used for the ZOSSVOL attribute. c. If the ZOSSVOL attribute does not specify the correct disk volume, use the SCF ALTER SUBSYS command to correct the problem. 3. Check the security permissions of the CVT and FUP utilities to ensure that the OSS Monitor can execute them. Correct them if necessary. Reissue the ALTER FILESET command to change the CATALOG value. If the problem persists: 1.
Messages Numbered Messages Recovery. If the affected server is... Take this action... OSS message-queue server Use the Subsystem Control Facility (SCF) STOP SERVER command to stop the server. OSS name server Use the SCF STOP FILESET command for all filesets managed by the server. OSS sockets local server 1. Stop all applications with open AF_UNIX sockets. 2. Use the SCF STOP SERVER command to stop the server. 1. Stop all applications with open sockets. 2.
Messages Numbered Messages 32 OSS E00032 Invalid combination of command options Cause. A Subsystem Control Facility (SCF) DIAGNOSE FILESET command was issued, but the command contained an invalid combination of options. This condition usually occurs when more than one of the options STATUS, REPAIR, UPGRADE, and DOWNGRADE are specified. Effect. The command is not processed. Recovery. Reissue the command with a single option or a valid combination of options. See DIAGNOSE FILESET Command on page 12-43.
Messages Numbered Messages 35 OSS E00035 Command reserved for SUPER.SUPER only Cause. A Subsystem Control Facility (SCF) command that requires the super ID was issued but you are not logged on with the super ID. Effect. The command is not processed. Recovery. Exit SCF, change your user ID to the super ID, and re-enter SCF. Reissue the command. If you are not permitted to use the super ID, contact your system administrator to have the command issued for you.
Messages Numbered Messages Effect. The command is not processed. Recovery. Reissue the command with a single specification for the attribute. 39 OSS E00039 Missing required attribute attribute attribute identifies the missing attribute. Cause. A Subsystem Control Facility (SCF) command was entered without a required attribute. Effect. The command is not processed. Recovery. Reissue the command with the required attribute specification. 44 OSS E00044 Reserved name for OSS transport agent server Cause.
Messages Numbered Messages 46 OSS E00046 Alter Name Server of active fileset is not allowed Cause. A Subsystem Control Facility (SCF) ALTER FILESET command attempted to change the OSS name server process for a fileset while the fileset is mounted. Effect. The command is not processed. Recovery. Stop the fileset and any filesets mounted on it, then reissue the command and restart all affected filesets. 47 OSS E00047 Mount point pathname must be an absolute pathname Cause.
Messages Numbered Messages 52 OSS E00052 Fileset fileset is not mounted. fileset identifies the affected fileset. Cause. A Subsystem Control Facility (SCF) CONTROL FILESET command was entered, but the fileset targeted by it is not mounted. Effect. The command is not processed. Recovery. Use a START FILESET command to start the fileset and reissue the CONTROL FILESET command. 53 OSS E00053 Name Server server is not running. server identifies the affected server process. Cause.
Messages Numbered Messages 55 OSS E00055 The MAXINODES value is lower than the number of currently inuse inodes inuseinodes for the fileset fileset. inuseinodes is the current number of inodes in use. fileset identifies the affected fileset. Cause. A Subsystem Control Facility (SCF) CONTROL FILESET command was entered, but the specified MAXINODES value is lower than the current number of inodes in use. Effect. The command is not processed. Recovery.
Messages Numbered Messages Effect. The command is not processed. Recovery. Use the ALTER SERVER command to change the BACKUPCPU attribute, and reissue the CONTROL SERVER command with the SYNC option. 58 OSS E00058 The migration of a primary or a backup Name Server server to a different processor failed. server identifies the affected server process. Cause. A Subsystem Control Facility (SCF) CONTROL SERVER command was entered, but the backup OSS name server process is not in a valid state. Effect.
Messages Numbered Messages 60 OSS W00060 The fileset is started, but MAXINODES value is changed to maxinodesvalue maxinodesvalue indicates the MAXINODES value being used. This value is also stored in the configuration database for the fileset. Cause. The specified MAXINODES value is less than 110 percent of the number of inuse inodes. This message can occur when an OSS name server tries to mount an existing fileset and the previous maximum value did not allow reasonable growth. Effect.
Messages OSSTTY Subsystem Messages OSSTTY Subsystem Messages The messages in this section are generated by the OSSTTY subsystem. OSSTTY generates the unnumbered messages described in Startup Messages on page A-58 when it is first started. It can log the EMS messages, described in the Operator Messages Manual, while running; the subsystem ID displayed by these messages is OSSTTY.
Messages Startup Messages OSSTTY: Both the attributes attr1 and attr2 are mutually exclusive attr1 indicates the first of two conflicting command flags or values. attr2 indicates the second of two conflicting command flags or values. Cause. The identified flags or values cannot be used in the same command. Effect. OSSTTY is not running. Recovery. Reenter the command without specifying one of the conflicting flags.
Messages Startup Messages Cause. Access to one of the redirection targets failed. Effect. OSSTTY is not running. Recovery. Check that: • • A specified process is running An existing EDIT output file is not corrupted, has the correct access permissions for use, and is not already opened by another process Reenter the command after correcting the problem. OSSTTY: Value out of range for attribute attr attr identifies the command flag that could not be processed. Cause.
Messages Startup Messages OSSTTY: ****Warning**** Unable to create the backup process, error error_number. Continuing... error_number indicates the Guardian file-system error returned by the underlying call to the Guardian PROCESS_LAUNCH_ procedure. For more information about this error, refer to the Guardian Procedure Errors and Messages Manual. Cause. The primary copy of the OSSTTY process could not start its backup copy. Effect.
Messages Startup Messages OSSTTY: ****Warning**** EMS Collector coll_name is not accessible, error error_number, Using $0 as the EMS collector. coll_name indicates the process name specified in the command. error_number indicates the Guardian file-system error returned by the underlying call to the Guardian FILE_OPEN_ procedure. For more information about this error, refer to the Guardian Procedure Errors and Messages Manual. Cause. The specified collector is not accessible.
B Manually Setting Up an OSS Environment Table B-1 summarizes the procedures for configuring and starting a new OSS environment by using manually entered operator commands instead of the OSSSETUP utility. Table B-2 on page B-9 summarizes the procedures for completing a new OSS environment preconfigured by HP or created by the OSSSETUP DEFAULTS command.
Manually Setting Up an OSS Environment Table B-1. Creating a Basic OSS Environment Without Using the OSSSETUP Utility (page 2 of 8) Task Subtask Prepare for configuration (continued) Verify system resources: • • • • • • Start related processes Example See Guardian User’s Guide 64 MB memory/ processor 500 MB disk space for root fileset Running software release version update (RVU) documente d in this guide FILEINFO $SYSTEM.SYS*.
Manually Setting Up an OSS Environment Table B-1. Creating a Basic OSS Environment Without Using the OSSSETUP Utility (page 3 of 8) Task Subtask Example See Provide required configuration files Create or modify a storage-pool file for each fileset FUP DUP ZXOSSMON.OSSPOOL, ZXOSSMON.ROOTPOOL Creating a Storage Pool on page 5-6 Ensure that preconfigured databases are accessible after a node number change FUP INFO $SYS*.*.
Manually Setting Up an OSS Environment Table B-1.
Manually Setting Up an OSS Environment Table B-1. Creating a Basic OSS Environment Without Using the OSSSETUP Utility (page 5 of 8) Task Subtask Example See Finish configuring the OSS environment Start an OSS shell OSH osh(1) reference page in the Open System Services Shell and Utilities Reference Manual Create the general profile file /bin/cd /etc Setting Up an /etc/profile File on page 9-2 /bin/cp profile.
Manually Setting Up an OSS Environment Table B-1. Creating a Basic OSS Environment Without Using the OSSSETUP Utility (page 6 of 8) Task Subtask Example See Finish configuring the OSS environment (continued) Configure and start network services (continued) /bin/cp smplinetd.conf inetd.conf Starting a Network Services Server on page 4-38 /bin/vi inetd.conf shell stream tcp nowait root /bin/rshd /bin/vi hosts.
Manually Setting Up an OSS Environment Table B-1.
Manually Setting Up an OSS Environment Table B-1. Creating a Basic OSS Environment Without Using the OSSSETUP Utility (page 8 of 8) Task Subtask Example See Configure OSS users Add users and specify an alias for each SAFECOM Managing Users and Groups on page 8-9 ADD USER USER.ONE, 001,001, PASSWORD Secure1 ALTER USER USER.ONE, GUARDIAN DEFAULT SECURITY NUNU ALTER USER USER.ONE, GUARDIAN DEFAULT VOLUME $DATA.
Manually Setting Up an OSS Environment Table B-2. Completing a Preconfigured Basic OSS Environment (page 1 of 4) Task Subtask Example See Prepare for configuration Log on as the super ID LOGON SUPER.SUPER Guardian User’s Guide Ensure that preconfigured databases are accessible after a node number change FUP INFO $SYS*.*.ZOSSFSET, DETAIL HP strongly discourages changes in node (system) numbers.
Manually Setting Up an OSS Environment Table B-2.
Manually Setting Up an OSS Environment Table B-2. Completing a Preconfigured Basic OSS Environment (page 3 of 4) Task Subtask Example See Finish configuring the OSS environment (continued) Create the general profile file /bin/cd /etc Setting Up an /etc/profile File on page 9-2 /bin/vi profile umask 022 #Only users have write permission on their files. set -o noclobber #Redirection can’t overwrite files. set -o trackall #Track all aliases. export MANPATH=/usr/share/man #Match PATH use.
Manually Setting Up an OSS Environment Table B-2. Completing a Preconfigured Basic OSS Environment (page 4 of 4) Task Subtask Example See Finish configuring the OSS environment (continued) Configure utilities for periodic tasks /bin/cd /usr/bin/cron Configuring the cron Process on page 2-35 /bin/cp at.deny.sample at.deny /bin/chmod 1775 * /bin/cd /var/adm/cron /bin/cp queuedefs.sample queuedefs /bin/cp .proto.sample .proto /bin/cp cron.deny.sample cron.
C OSS Management Utilities This appendix describes: • • The optional OSSTTY utility The OSS Easy Setup utilities provided in the T0585 product OSSTTY OSSTTY can run as a single-use process or as a server. OSSTTY provides an alternative to Telserv for OSS program terminal input or output using OSS standard files.
OSS Management Utilities Starting OSSTTY Command To run OSSTTY, enter this command: [RUN] OSSTTY [ [/ NAME process_name, [ IN stdin_redirection_target, [ OUT stdout_redirection_target, [ TERM stderr_redirection_target, [ NOWAIT [ -access { ALL | OWNER } [ -backupcpu processor [ -coll collector [ -help [ -idletimeout seconds [ -[no]prefixpid [ -[no]quiet [ -[no]server [ -[no]wrap [ -secure security_string ] ] ] ] ] / ] ] ] ] ] ] ] ] ] ] ] NAME $process_name specifies the Guardian process name for the OSS
OSS Management Utilities Starting OSSTTY If an EDIT file (file code 101) is specified but does not exist, the file is created with the security specified by the -secure option or the default security setting for the user. If you omit this option, the name used is the device name for the Telserv terminal from which the command is entered. TERM stderr_redirection_target specifies the Guardian process name for the process to which data for the OSS standard error file should be written.
OSS Management Utilities Starting OSSTTY -coll collector specifies the Event Management Service (EMS) collector process to receive the OSSTTY event messages. If you specify a collector process that cannot be used, OSSTTY issues a warning message. If you omit this option or if the specified collector process is not valid, OSSTTY uses $0 as the collector process. -help requests a summary of OSSTTY command options. This is the first message described in Startup Messages on page A-58.
OSS Management Utilities Starting OSSTTY -[no]quiet controls whether the warning and error messages described in Startup Messages on page A-58 are displayed. Suppressing message display can be useful when OSSTTY is used within a TACL macro or other script. When the -quiet option is specified, no messages are displayed. When the -noquiet option is specified, messages are displayed as necessary. If you do not specify either option, OSSTTY uses the -noquiet option.
OSS Management Utilities • • • Starting OSSTTY The OSS shell commands and utilities do not restrict the use of the standard error file to the display of error messages. As is true for many implementations of UNIX, the standard error file can be used when unbuffered output to the terminal is the desired behavior for a utility.
OSS Management Utilities Stopping OSSTTY Stopping OSSTTY To stop a copy of the OSSTTY process started through either the OSH command or from a TACL prompt, use the TACL STOP command. Use the STOP command carefully; when OSSTTY does not terminate normally, any EDIT files it has open might not be properly closed and data could be lost. To stop a copy of the OSSTTY process monitored by the persistence manager, use the SCF Kernel subsystem ABORT PROCESS $ZZKRN.#ZTTY command.
OSS Management Utilities Utility File Security Table C-1. The EasySetup Utilities (page 2 of 2) Component Description OSSTREE file and /tmp/oss.tree. ddmmmyyyy.system_name file An EDIT file and an ASCII text file, respectively, containing a list of all directories created in the OSS file system. These files are created by the OSSSETUP utility and contain the result of the following OSS shell command: find / -WNOE -WNOG -type d -print ddmmmyyyy.
OSS Management Utilities Diagnostic Messages You can stop any of the utilities by pressing the Break key. This action is acknowledged by the following message: Break or error terminated operation. task of the OSS subsystem on system-name: FAILED Utility-name terminated unexpectedly or encountered errors.
OSS Management Utilities Utility PARAMs OSS EasySetup EMS events are logged with the OSS subsystem ID and have the form: TANDEM.OSS.D30 00010 USER NOTICE (userID) : text where userID is your user ID, and text is the event message text. Diagnostic messages for nonfatal situations include a suggested response to avoid the problem.
OSS Management Utilities OSSSETUP Utility Examples 1. To issue progress messages as EMS events using a ZSPIDEF subvolume on $SYSTEM when initially configuring and starting OSS, enter: PARAM EASYSETUP^EMSVOL $SYSTEM OSSSETUP 2. To issue progress messages as EMS events using a ZSPIDEF subvolume on $SYSTEM on the remote node \NODE when initially configuring and starting OSS, enter: PARAM EASYSETUP^EMSVOL \NODE.
OSS Management Utilities OSSSETUP Utility OSSSETUP is intended to provide a quick way to initially configure an OSS environment. You can later build on the configuration OSSSETUP creates by manually configuring additional resources or altering your initial configuration. You must not use the OSSSETUP utility when an OSS configuration already exists on your node. For example, you cannot use OSSSETUP when: • • • Your system was preconfigured by HP, but OSSREMOV has not been run.
OSS Management Utilities OSSSETUP Utility Figure C-1.
OSS Management Utilities STARTOSS Utility Considerations • • • • • • • • • Beginning with the G06.17 release version update (RVU) and the G10 version of the OSS Monitor, the OSS Monitor no longer uses the $NULL process or requires the $ZSMP process for its own startup. However, not configuring or starting $NULL or $ZSMP could affect the functioning of other products or subsystems.
OSS Management Utilities • STARTOSS Utility Your OSS environment was not preconfigured by HP or configured by using OSSSETUP, but you have created an OSSINFIL file in the STARTOSS volume and subvolume and maintain your current OSS configuration in that file. Under other conditions, using the STARTOSS utility might not achieve your intention. Command To run STARTOSS, enter these commands: VOLUME $SYSTEM.
OSS Management Utilities • STOPOSS Utility Fileset mount points should be checked after all filesets are started to ensure that they have security permissions consistent with your site’s security policies. The OSS shell chmod command can be used to modify permissions as necessary. STOPOSS Utility The STOPOSS utility stops all filesets, regardless of whether they were configured in OSSINF or OSSINFIL. STOPOSS is intended for use before a system shutdown.
OSS Management Utilities • OSSREMOV Utility You can embed this command in the subsystem shutdown file as follows: == Stop the OSS subsystem, filesets, and servers: #Push #Defaults Volume $SYSTEM.ZOSSINS Run STOPOSS #Pop #Defaults The keyword Run is optional, but it is required if #PMSEARCHLIST does not include #DEFAULTS. OSSREMOV Utility The OSSREMOV utility deletes the configurations of all processes and filesets regardless of whether they are configured in the OSSINF file or OSSINFIL file.
OSS Management Utilities • OSSINF File The OSSREMOV command prompts you to determine whether it should remove the security manager server process $ZSMP and the $NULL process. These processes are often required by other products and usually should be allowed to continue running. OSSREMOV uses the Safeguard SAFECOM program to stop $ZSMP to ensure that $ZSMP does not restart itself. If your site has not licensed Safeguard, either do not use OSSREMOV or respond “no” to its prompt about stopping $ZSMP.
OSS Management Utilities OSSINFIL File Figure C-4. Example of an OSSINF File * File: OSS Monitor in-file ( Version 1.0 ) { <--MUST BE FIRST LINE OF FILE} * This is an in-file for the OSS T0585 utilities STARTOSS and OSSSETUP. The * OSS Monitor process manages all entries in this file. * * * * There is no need to refer to any name server or transport agent server. Name servers are automatically started when their corresponding filesets are started.
OSS Management Utilities • • • SERVER entries must contain the number sign (#) character. SERVER entries for #ZTAnn are not needed. SERVER entries for OSS name servers are not valid.
D Falling Back to a Previous Release Version Update You cannot fallback to a G-series RVU from an H-series RVU because of hardware considerations. The considerations in this appendix apply to OSS products only. There might be other considerations when falling back to a previous RVU. See the Release Version Update Compendium for the current RVU for fallback notes for products installed on your system and see the Interactive Upgrade Guide for detailed information about falling back to a previous RVU.
Falling Back to a Previous Release Version Update • Falling Back to G-series Release Version Update As Far Back as G06.
E Environment Limits This appendix summarizes the architectural and functional limits that apply to the Open System Services environment on NonStop servers. Limits for a specific release version update (RVU) are subject to increases as software product revisions (SPRs) occur. OSS application programs that conform to POSIX.1 conform to a set of limits defined by the POSIX standards; POSIX limits can be less than those imposed by the environment.
Environment Limits Table E-1. OSS Environment Limits (page 2 of 4) Description OSS file size Maximum Value For H06.06 and later RVUs: Approximately 1 terabyte (constrained by the space available on the disk volume) Comment To Check Amount Currently Used For RVUs H06.06 and later, the file size limit depends on the function used to open the file. For details, see OSS and Guardian Enscribe File Formats and File Size Limits on page E-5. For G-series RVUs and for H-series RVUs prior to H06.
Environment Limits Table E-1. OSS Environment Limits (page 3 of 4) Description Maximum Value Comment OSS file opens per application process, using the select() function to manage I/O 1,024 4,096 Prior to RVU G06.26 (prior to NPV T9055 G12), the limit was 1,024.
Environment Limits Table E-1. OSS Environment Limits (page 4 of 4) Description OSS filesets per node Maximum Value Functional limit: 1,000,000 Theoretical maximum: 33,553,952 Entries in an inode cache 500,000 Comment Functional limit is imposed by functional limit on directories (a fileset must have a directory as a mount point). To Check Amount Currently Used At a TACL prompt, use the SCF INFO FILESET *, DETAIL command. Theoretical maximum is determined by maximum number of device labels.
Environment Limits OSS and Guardian Enscribe File Formats and File Size Limits OSS and Guardian Enscribe File Formats and File Size Limits The size limit for a file depends on: • • • • The function used to open the file The RVU on which the file was created The type of API (OSS or Guardian) used to create the file The current size of the file File Size Limits For Files Created on H06.06 and Later RVUs Note. For H06.
Environment Limits File Size Limit Behavior for File Open Operations Table E-3. File Format and Limits Table for File Open Behavior Using This Function 32-Bit OSS APIs such as open( ) 64-Bit OSS APIs such as open64( ) Guardian APIs without 64-Bit elections Guardian APIs with 64-Bit elections OSS small file (< approx. 2 GB) created on H06.
Environment Limits File Size Limit Behavior for File Open Operations Table E-3. File Format and Limits Table for File Open Behavior Using This Function 32-Bit OSS APIs such as open( ) 64-Bit OSS APIs such as open64( ) Guardian APIs without 64-Bit elections Guardian APIs with 64-Bit elections Guardian Enscribe Format 1 file created using OSS APIs on H06.
Environment Limits File Size Limit Behavior for File Open Operations Table E-3. File Format and Limits Table for File Open Behavior Using This Function To Open This File... 32-Bit OSS APIs such as open( ) 64-Bit OSS APIs such as open64( ) Guardian APIs without 64-Bit elections Guardian APIs with 64-Bit elections Guardian Enscribe Format 1 file created using Guardian APIs on G06.
Glossary A7CINFO file. A Distributed Systems Management/Software Configuration Manager (DSM/SCM) configuration file distributed with OSS products that contains information about the contents of all pax archive files of an OSS distribution subvolume. absolute pathname. A pathname that begins with a slash (/) character and is resolved beginning with the root directory. Contrast with relative pathname. address space. The memory locations to which a process has access. API.
Glossary DSV the software product number and rrr is the base version identifier (such as G01) or software product revision (SPR) identifier (such as AAB.) DSV. See distribution subvolume (DSV). Epoch. The period beginning January 1, 1970, at 0 hours, 0 minutes, and 0 seconds Coordinated Universal Time (UTC). See also Coordinated Universal Time (UTC). FIFO. A type of special file that is always read and written in a first-in, first-out manner. file.
Glossary Guardian Guardian. An environment available for interactive or programmatic use with the HP NonStop operating system. Processes that run in the Guardian environment usually use the Guardian system procedure calls as their application program interface; interactive users of the Guardian environment usually use the HP Tandem Advanced Command Language (TACL) or another HP product’s command interpreter. Contrast with Open System Services (OSS). inode.
Glossary OSS OSS. See Open System Services (OSS). OSS environment. See Open System Services (OSS) environment. OSS Monitor. See Open System Services (OSS) Monitor. pathname. In the OSS environment, the string of characters that uniquely identifies a file within its file system. A pathname can be either relative or absolute. See also ISO/IEC IS 9945-1:1990 (ANSI/IEEE Std. 1003.1-1990 or POSIX.1), Clause 2.2.2.57. pathname component. See filename. pathname resolution.
Glossary storage pool storage pool. The set of disk volumes that contain files for a specific fileset or that can be used to contain files for a specific fileset. The storage pool includes the creation pool. storage-pool file. A file containing a list of disk volumes to be used when creating new files in a fileset. As these volumes are filled, more volumes can be added to the storagepool file. superblock.
Glossary user ID user ID. The unique identification of a user within a node. In the Guardian environment, the term “user ID” usually means the group number and member number pair; for example, the user ID of the super ID is usually described as (255, 255). In the OSS environment, the term “user ID” usually refers to the single number that is the scalar view of the user ID—a number called the UID. See also UID. UTC. See Coordinated Universal Time (UTC) working directory.
Index A A7CINFO file 6-4 ABENDED: processname (OSS Monitor message) A-27 acctcom utility (UNIX) 8-4 accton utility (UNIX) 8-4 aculog file (UNIX) 8-3 ADD ALIAS command (SAFECOM) 8-12, 8-20 ADD FILESET command (SCF) 12-7/12-15 ADD GROUP command (SAFECOM) 8-9 ADD PROCESS command (SCF) 2-9 ADD SERVER command (SCF) 12-16/12-19 ADD USER command (SAFECOM) 8-9, 8-20 Administrative groups 8-10, 8-12 adm/wtmp file (UNIX) 8-3 AF_INET sockets 1-15, 4-5, 4-34, 4-35, 4-36, 4-38 AF_INET6 sockets 1-15, 4-5, 4-34, 4-35, 4-3
Index B B Backup strategies 6-15/6-16 using Backup and Restore 2.0 6-18 using BRCOM 6-11 volume mode 6-23 BACKUPCPU attribute OSS Monitor 2-9 Bad free inode list (inconsistency checked by FSCK) 5-30 Bad parent list (inconsistency checked by FSCK) 5-31 basename.key file 4-28 basename.private file 4-28 batch command (OSS) 2-38 Berkeley Internet Name Domain (BIND) 9 4-6 bin user name (UNIX) 8-6 BIND 9.2.3 4-6 BIND 9.
Index D Configuration files 4-7/4-18 inetd process 4-24 storage-pool files 3-10/3-12, 4-17/4-18, 5-7 ZOSSFSET file See ZOSSFSET file ZOSSPARM file See ZOSSPARM file ZOSSSERV file See ZOSSSERV file ZPCONFIG file See ZPCONFIG file Configuration files, OSS ZPOS files D-1 console file (UNIX) 8-1 CONTROL FILESET command (SCF) 12-37/12-39 CONTROL SERVER command (SCF) 12-39/12-41 COPYOSS macro (TACL) 4-9, 6-4, 6-7, 6-8, 8-28 Corrupt Inode, Inode=inode-number (FSCK message) A-15 Corrupt PXLINK Record - Parent:par
Index E Directory graph 5-31 Directory names, characters, number of 3-1 Dirty Catalog using Fast Create; REPAIR ALL will be performed (FSCK message) A-11 discard service (inetd) 8-6 Disk Compression Program (DCOM) 9-9 Disk files, Guardian, referencing 3-8 Disk process 1-5, 1-6, 3-7 Disk Space Analysis Program/Disk Compression Program (DSAP/DCOM) 9-9 Disks defragmenting 9-10 sharing among storage pools 4-18 Distributed Systems Management/Software Configuration Manager (DSM/SCM) 6-4, 6-6, 6-7, 6-8, 6-9 DNS
Index F ERROR messages (FSCK) (continued) 303 - Catalog Inconsistent A-24 305 - Invalid { PXINODE | PXLOG | PXLINK } File A-25 306 - DP2 Cache Flush Write Error A-25 307 - Can’t Create SQLCAT Process PROCESS_CREATE_ Error A-26 308 - Unexpected SQLCAT Error error Purging File filename A-26 6 - Catalog Version not supported by this program A-9 98 - INTERNAL ERROR A-11 99 - HEAP OVERFLOW A-12 ERROR utility, Guardian 6-13 etc/rpc.
Index G Filesets (continued) customer-defined 4-8 deleting 5-34 diagnosing 5-24/5-34, 12-43/12-46 disk space 3-12 disk volume limit 5-8 examining state 5-13, 12-66 integrity checking See FSCK utility mounting 2-6, 5-5, 5-7, 12-64 naming 4-8 obtaining information about 5-13 reconfiguring 5-14 renaming 5-34 restarting 2-3, 2-6, 5-7 root 4-8, 5-32 size 3-10 starting 2-6, 5-7, 12-64 state, displaying 5-13, 12-66 stopping 2-3, 5-13, 5-14, 12-80 unmounting 2-3, 5-13, 5-14, 12-80 upgrading 5-40 volume allocation
Index H genxlt command (OSS) 9-6 get command options 9-11 getopts command (OSS) 9-11 getstats utility (UNIX) 8-4 gname command (OSS) 6-2 GNU 8-5 groff utility (UNIX) 9-4 Group list 8-10 Groups 8-10, 8-12 characteristics 8-12 groups file (UNIX) 8-1 gtacl command (OSS) 1-2 Guardian environment initial FTP access to 8-21 OSS access from 8-11 Guardian filenames See Filenames guest user name (UNIX) 8-6 H HEAP OVERFLOW (FSCK message) A-12 HELP command (SCF) 12-2 HELP OSS command (SCF) 12-2, 12-3 help user name
Index K Initial working directory (continued) use of 8-13 Inode cache 1-11 free list 5-30 in Guardian file identifier 3-3, 6-2 invalid number 5-32 lost number 5-31 maximum number 1-11 missing 5-30 number 1-6 orphan 5-31 parent numbers 5-31 install_obsolete directory (OSS) 8-3 Integrity checker See FSCK utility INTERNAL ERROR FSCK message A-11 Internal error (OSS Monitor message) A-36 Internal Error*** (CVT message) A-5 Internet domain name server (DNS) 4-6 Internet sockets 1-15 Interprocess communication
Index M LC_TIME environment variable 9-6 lib/aliases file (UNIX) 8-3 lightweight resolver utility 4-7 Links maximum number of 1-11 too many 5-30 links ilink number 1-6 Local sockets (AF_UNIX) 1-15 Locale behavior environment variable (LC_ALL) 9-5 locale command (OSS) 9-6 Localization creating new locale 9-6 environment variables 9-5 OSS Monitor 12-5 OSS shell 9-5/9-8 local/etc/http/logs/access_log file (UNIX) 8-3 logger command (OSS) 2-7, 8-5 logical network partitioning 1-16 Login displaying login name 8
Index N MINOR messages (FSCK) (continued) 215 - Invalid Timestamp, Inode=inodenumber A-20 216 - ZYQ File Conflict - filename A-21 301 - Invalid file in ZYQ Subvolume filename A-23 Missing inode (inconsistency checked by FSCK) 5-30 Missing Inode, Inode=inode-number (FSCK message) A-18 Missing Link - Parent:parent, Child:child, Name:link linkname (FSCK message) A-13 Missing or corrupt superblock (inconsistency checked by FSCK) 5-29 Missing required attribute (OSS Monitor message) A-52 Missing ZYQ file (inco
Index O NOE (-W flag option) 3-6 NOG (-W flag option) 3-6 nonblocking input and output 1-9 Nonzero mounted flag (inconsistency checked by FSCK) 5-29 NORMALIOMODE attribute 5-15, 5-20, 5-21 Not a Root Fileset (FSCK message) A-21 nroff utility (UNIX) 9-4 nsupdate utility 4-7, 4-27 ntpd process (UNIX) 8-4 null file (OSS) 8-3 NULL.
Index O OSS E00022 Invalid value for attribute attribute-name (OSS Monitor message) A-44 OSS E00023 Name Server server rejected the request (OSS Monitor message) A-44 OSS E00024 Name Server server gave unexpected response to OSS Monitor: status (OSS Monitor message) A-45 OSS E00025 Name Server server detected invalid data (OSS Monitor message) A-45 OSS E00026 Repair is needed for corrupted fileset filesetname (OSS Monitor message) A-46 OSS E00027 Root fileset is not started (OSS Monitor message) A-46 OSS
Index O OSS environment (continued) management activities 1-3 OSS Monitor interface 12-1 restarting 2-3, 2-6 stopping 2-3 OSS file caching 5-18/5-21 OSS file compression using pack 9-10 OSS file system automatic restart of 5-8/5-11 components 1-10/1-11 detailed description 3-1/3-12 differences from UNIX 1-5/1-8 restarting 2-6 stopping 2-3 OSS files 3-8 catalog portion of 3-7 data portion of 3-7 file code for 3-7 FUP INFO displays 6-3 Guardian access permissions for 3-7 large files 3-9 not backed up with G
Index O OSS Monitor failed in adding converted record to filename file -- Error: err (OSS Monitor message) A-29 OSS Monitor failed in adding default record to filename file -- Error: err (OSS Monitor message) A-29 OSS Monitor failed in creating filename file - Error: err (OSS Monitor message) A-30 OSS Monitor failed in opening filename file - Error: err (OSS Monitor message) A-32 OSS Monitor failed in reading filename file - Error: err (OSS Monitor message) A-33 OSS Monitor failed to get its process name
Index P OSS transport agent servers 4-4, 12-66, 12-82 OSS W00060 The fileset is started, but MAXINODES value is changed to maxinodesvalue (OSS Monitor message) A-57 OSS W00061 The fileset is started, but not all the volumes in the pool edit file are eligible for file creation (OSS Monitor message) A-57 OSSBUFFERED attribute value 5-17, 5-21 OSSBUFFEREDCP attribute value 5-16, 5-21 OSSCACHING C-14, C-18 OSSINF file C-19 OSSINFIL file 2-2, 4-29, 4-30, 4-31, 4-50, 5-7, C-19 OSSJOURN file C-7 OSSLIB file C-7
Index Q Persistent process OSS application 2-23/2-29 OSS Monitor 2-4, 2-7, 2-8, 2-9/2-10, C-11 OSSTTY C-1 security manager 2-9, C-11 $NULL C-11 Personalities 1-1 pgp utility (UNIX) 8-4 ph utility (UNIX) 8-4 PINSTALL command (TACL) 4-9, 6-4, 6-8, 8-28 PINSTALL utility 6-5, 6-8 pname command (OSS) 6-3 POOL attribute 5-15 portmap process 4-6, 4-42 POSIX^CONFIG^LOC PARAM 2-12 preserve file (OSS) 8-3 Previous release version update, falling back to D-1 Primary group 8-10 Primary user groups 8-12 PRIMARYCPU att
Index S Requests, handling, server process 4-5 resolv.
Index S SCF (continued) INFO PROCESS command 12-58/12-61 INFO SERVER command 12-52/12-57 INFO SUBSYS command 12-58/12-61 NAMES command 12-61 RENAME FILESET command 5-34 required 12-5 START FILESET command 2-6, 5-7, 12-64/12-65 START PROCESS command 2-10 STATUS FILESET command 5-13, 12-66/12-74 STATUS SERVER command 12-75/12-80 STOP command 2-4, 2-15 STOP FILESET command 5-13, 5-14, 12-80/12-81 VERSION MON command 12-82/12-85 VERSION PROCESS command 12-82/12-85 VERSION SUBSYS command 12-82/12-85 Security a
Index T SQL/MP programs restoring 6-24 SQL/MX compiler remote invocation 4-5 START FILESET command (SCF) 2-6, 5-7, 12-64/12-65 START PROCESS command (SCF for NonStop Kernel) 2-10 START SERVER command (SCF) 12-65/12-66 STARTMODE attribute OSS application 2-23 OSS Monitor 2-9 STARTOSS utility 2-2, 2-8, 4-29, 4-30, 4-31, 4-36, 4-37, 4-50, 5-7, C-14 STARTUPMSG attribute OSS application 2-23, 2-24, 2-25 OSS Monitor 2-9 STATUS command (TACL) 2-8 STATUS FILESET command (SCF) 5-13, 12-66/12-74 STATUS SERVER comma
Index U TCPIP_RESOLVER_NAME environment variable 4-25 TCPMON 1-19 telnet command (UNIX) 8-5 telnet user name (UNIX) 8-6 Telserv configuring access 7-2/7-5 configuring user login 7-2, 8-11 direct access to OSS 7-1, 7-3/7-5, 8-11 indirect access to OSS 7-1, 7-2, 8-11 logging in through an OSS program 7-4/7-5 through OSS shell 7-3/7-4 OSSTTY break key processing C-6 redirection of OSS standard files from C-1 TEMPPOOL 4-10, 4-18 TERM file 6-28 termcap file (OSS) 8-3 terminal helper process 1-9 tftpd utility (
Index V UNBUFFEREDCP attribute value 5-16 Unexpected Argument*** - token (CVT message) A-6 Unexpected SQLCAT Error error Purging File filename (FSCK message) A-26 UNIX cron utility 2-34 df command 2-15 fsck command 2-15 mkfs command 2-15 mount command 2-15 umount command 2-15 yellow pages 8-10 UNIX super group 8-7 UNIX super ID 8-7 Unknown keyword specified on the command line (OSS Monitor message) A-35 unlink() function 5-17 Unmounting a fileset 2-3, 5-13, 12-80 Upgrading a catalog 12-45 User definitions
Index X X xargs utility (UNIX) 8-5 Z ZEMSTACL C-9, C-10 ZFB* file 6-6, 6-8 ZINSPOOL 4-18 ZOLDFSET 4-11 zone-filename.
Index Special Characters $ZHOME 2-13, C-1 $ZLOG C-9 $ZMSGQ 1-15, 4-2, 4-19, C-16 $ZPLS 1-17, 1-19, 4-4, 4-21, 4-22, 4-23, C-16 $ZPMn 4-6 $ZPMON C-11, C-15, C-16 adding to NonStop Kernel subsystem 2-9 OSS Monitor process name 2-8, 2-13 verifying installation 2-8 $ZPNS 4-2, 4-8, 4-15, 4-49 $ZPPnn 1-15 $ZPTMn 1-19 $ZSAMn 1-17 $ZSMP 2-7, 2-9, C-11, C-15, C-16 $ZTAnn 1-16, 1-17, 4-4 $ZTCn 1-17 $ZTTnn 1-9 ***Command Error*** - token (CVT message) A-3 ***fsck needed -- subvolume.
Index Special Characters /tmp directory 4-8, 5-3 /tmp/oss.tree.ddmmmyyyy.system_name file C-8 /usr/etc/exportfs utility 8-5 /usr/etc/showmount utility 8-5 /usr/include directory 6-7 /usr/include/unistd.h 9-6 /usr/lib/cron/at.allow file 2-37 /usr/lib/cron/at.deny file 2-35, 2-37 /usr/local/man/man* directories (OSS) 9-4 /usr/share/man/cat* directories (OSS) 9-4 /usr/share/man/man* directories (OSS) 9-4 /usr/ucb directory 6-7 /var/adm/cron.
