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. 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. This guide supports J06.03 and all subsequent J-series RVUs, H06.08 and all subsequent H-series RVUs, and G06.
© Copyright 1995, 2014 Hewlett-Packard Development Company, L.P. Confidential computer software. Valid license from HP required for possession, use or copying. Consistent with FAR 12.211 and 12.212, Commercial Computer Software, Computer Software Documentation, and Technical Data for Commercial Items are licensed to the U.S. Government under vendor’s standard commercial license. Warranty The information contained herein is subject to change without notice.
Contents About This Document...................................................................................15 Supported Release Version Updates (RVUs)................................................................................15 Intended Audience..................................................................................................................15 Unsupported Utilities...............................................................................................................
Automatic Startup Service...................................................................................................47 Possible Ways to Stop the OSS File System............................................................................47 Manually Stopping the OSS File System and the OSS Environment...........................................47 Sending Warnings to Users............................................................................................
Small Files....................................................................................................................86 Large Files....................................................................................................................86 Fileset Size Considerations..................................................................................................87 OSS Configuration Files.....................................................................................................
Automatic Restart of OSS Name Servers After Processor Failure.............................................148 Potential Problems During Automatic Restart of Filesets..........................................................148 Auditing a Fileset..................................................................................................................149 Using the AUDITENABLED Attribute....................................................................................150 Audited SCF Operations.......
Removing Obsolete OSS Files and Directories...........................................................................181 Updating the whatis Database Files........................................................................................182 Backing Up and Restoring OSS Files.......................................................................................183 Considerations................................................................................................................
Configuring Expand Users................................................................................................213 Configuring FTP Users......................................................................................................213 Allowing Anonymous FTP Use............................................................................................214 Disallowing OSS Use by Specific Users...............................................................................
OSS Core Utilities User Commands.........................................................................................243 Executing OSS Core Utilities Commands.............................................................................244 Viewing OSS Core Utilities Reference Pages........................................................................244 11 Managing OSS Devices.........................................................................246 The Scope of OSS Device Management................
Examples...................................................................................................................296 FSCK Log File.............................................................................................................296 INFO FILESET Command...................................................................................................296 Example....................................................................................................................
B Manually Setting Up an OSS Environment..................................................386 Creating a Basic OSS Environment Without Using the OSSSETUP Utility......................................386 Ensure Target System is Ready for Software Updates.............................................................386 Update DSM/SCM If Necessary........................................................................................386 Install Current Guardian Product Files.....................................
E Summary of OSS Processes......................................................................409 F Environment Limits...................................................................................411 OSS and Guardian Enscribe File Formats and File Size Limits.....................................................418 Size Limits For Files Created on RVUs That Support Large OSS Files........................................418 File Size Limit Behavior for File Open Operations..............................
Figures 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 Guardian Filenames and OSS Files....................................................................................30 Pathname Resolution for Remote File Access Through the Guardian Expand Network................32 OSS File-System Components.............................................................................................34 Interprocess Communication Facilities..........................................................................
Examples 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 Sample Broadcast Message for Stopping the OSS File System................................................49 Sample Login Warning for Stopping the OSS File System......................................................50 Sample SCF LISTDEV Command Display.............................................................................58 TACL STATUS Display for an OSS Process.................................
About This Document This guide describes how to manage and operate the HP NonStop operating system Open System Services (OSS) environment. 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.
Unsupported Utilities HP is not responsible for the proper functioning of unsupported utilities or facilities and will not respond to product reports about them. Such utilities and facilities include those in the OSS /bin/unsupported directory. Use these utilities and facilities at your own risk. Acknowledgment The Portable Archive Interchange (pax) utility software was developed by Mark H. Colburn and is sponsored by the USENIX Association. ©1989 Mark H. Colburn. All rights reserved.
• • • Added information about the OSSSETUP utility automatically creating symbolic links for the TCP/IP product to: ◦ “Completing a Preconfigured Basic OSS Environment” (page 390) ◦ “Utility PARAMs” (page 401) ◦ “OSSSETUP Utility” (page 402) Added the following OSS Core Utilities information: ◦ Added information about differences between the OSS Utilities product (T8626) and the OSS Core Utilities product (T1202) in “About This Document” (page 15) and “Management Tools” (page 27) ◦ Added xinet
◦ “Starting an OSS Name Server” (page 127) ◦ “Determining the Current Configuration of a Server” (page 131) ◦ “Stopping a Specific OSS Name Server” (page 133) ◦ “Reconfiguring an OSS Name Server” (page 135) ◦ “Creating a Unique Fileset” (page 140) ◦ “Audited SCF Operations” (page 150) ◦ “Changing the Operating Parameters of a Fileset” (page 152) ◦ “Common and Unique Characteristics of OSS and UNIX Security” (page 216) ◦ New “Using an OSS Security Event-Exit Process (SEEP)” (page 231) sectio
• Revised information about logging OSS File System EMS events to both $0 and $ZLOG in “Event Management Service (EMS) Event Messages” (page 335). • Added process information for the resource agent servers to “Summary of OSS Process Information” (page 409).
Document Organization This guide contains information and procedures for managing and operating the OSS environment. It is divided into these chapters: Chapter 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. Chapter 2: “Operating the OSS Environment” Describes how to operate the OSS environment.
Appendix E: “Summary of OSS Processes” Provides a summary of names, descriptions, and start characteristics for common OSS processes. Appendix F: “Environment Limits” Provides guidance on current maximum values for various OSS features. Notation Conventions General Syntax Notation This list summarizes the notation conventions for syntax presentation in this manual. UPPERCASE LETTERS Uppercase letters indicate keywords and reserved words. Type these items exactly as shown.
K [ X | D ] address { } Braces A group of items enclosed in braces is a list from which you are required to choose one item. The items in the list can be arranged either vertically, with aligned braces on each side of the list, or horizontally, enclosed in a pair of braces and separated by vertical lines.
Bold Text Bold text in an example indicates user input typed at the terminal. For example: ENTER RUN CODE ?123 CODE RECEIVED: 123.00 The user must press the Return key after typing the input. Nonitalic Text Nonitalic letters, numbers, and punctuation indicate text that is displayed or returned exactly as shown. For example: Backup Up. Italic Text Italic text indicates variable items whose values are displayed or returned.
Related Information This subsection describes prerequisite and additional reading and training for this guide. Prerequisite reading and training are items that a reader of this guide must complete before using this guide. The following subsections briefly provide a context for each publication. Prerequisite Reading and Training NOTE: The HP NonStop OSS and Guardian documents are available from the NonStop Technical Library (NTL). The NTL is located at http://www.hp.com/go/nonstop-docs (external view).
Installation DSM/SCM User’s Guide NonStop Software Essentials Installation and Quick Start Guide Configuration • DNS Configuration and Management Manual • SCF Reference Manual for the Kernel Subsystem • SCF Reference Manual for the Storage Subsystem • TCP/IP Configuration and Management Manual • TCP/IP (Parallel Library) Configuration and Management Manual • TCP/IPv6 Configuration and Management Manual • Telserv Manual Periodic Task Management • Backup and Restore 2.
Publishing History Part Number Product Version Publication Date 527191-014 OSS Monitor T8622H03, T8622G12 February 2012 527191-015 OSS Monitor T8622H06, T8622G12 August 2012 527191-016 OSS Monitor T8622H06, T8622G12 February 2013 527191-017 OSS Monitor T8622H07, T8622G12 August 2013 527191-018 OSS Monitor T8622H08, T8622G12 February 2014 HP Encourages Your Comments HP encourages your comments concerning this document. We are committed to providing documentation that meets your needs.
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.
NOTE: Before using an OSS command, read the appropriate reference page either online or in the Open System Services Shell and Utilities Reference Manual to make sure that the command behaves in the way you expect. OSS commands conform to the XPG4 standards, but some OSS commands and utilities might have different options and behavior from the version of the UNIX operating system that you are familiar with.
Table 2 Management and Operations Tasks (continued) Task Subtask See Installing reference pages “OSS Management Utilities” (page 395), “Updating the whatis Database Files” (page 182), “OSSSETUP Utility” (page 402), or the Open System Services Installation Guide software “Manually Setting Up an OSS Environment” (page 386), “Installing New Product Files” (page 177), or the Open System Services Installation Guide reference pages and the whatis database “Updating the whatis Database Files” (page 182)
Table 2 Management and Operations Tasks (continued) Task Subtask See Using CVT “Managing and Repairing Fileset Catalog Files” (page 170) FSCK “Checking and Repairing Fileset Integrity” (page 159) OSS Monitor “Open System Services Monitor” (page 251) SCF “Using OSS Monitor Commands” (page 55), and the SCF Reference Manual for G-Series RVUs VPROC “Managing Problems” (page 249), and the Guardian User’s Guide OSS File System Concepts You use the OSS file system in the same way as you use a UNIX fi
OSS environment; these catalogs contain uniquely numbered data structures for each file and directory, called inode numbers and link numbers. Figure 1 shows that the OSS name servers resolve Guardian filenames and OSS pathnames to each other, then provide the information used by the system on behalf of the application program to communicate with the disk process, which provides access to the file on disk.
Figure 2 Pathname Resolution for Remote File Access Through the Guardian Expand Network The /E directory is described in more detail in “Understanding the OSS File System” (page 80). The /dev Directory The /dev directory is the device directory. It contains only the following files: tty The /dev directory itself is provided for compatibility with existing UNIX software. Do not put anything in this directory.
Components to Be Managed You need to manage the software described in the following subsections: • “Input/Output Utilities” (page 33) • “OSS Security” (page 33) • “OSS File-System Components” (page 33) • “Interprocess Communication Facilities” (page 35) Input/Output Utilities Input/output utilities include: • “OSSTTY Servers” (page 33) • “Terminal Helper Servers” (page 33) OSSTTY Servers Guardian administrative applications or OSS applications at your site might require you to make one or more c
Figure 3 OSS File-System Components An application running in the OSS environment makes calls for access to an OSS file through function calls in the OSS file-system library. The OSS file-system library is present in each processor. The library consists of functions in a shared run-time library (SRL) and in the system library. OSS file-system library functions are described in the Open System Services System Calls Reference Manual and in the Open System Services Library Calls Reference Manual.
The OSS file-system cache is used to buffer data transfers between the OSS disk files and an OSS application. Whether this cache is used for a specific file depends on: • Options used by an application to open the file • The setting of the OSSCACHING flag for the disk process of the disk volume that contains that file For more information on configuring this cache, see“Changing OSS File Caching for the Disks of a Fileset” (page 155).
• “Pipes and FIFOs” (page 39) • “OSS Sockets” (page 39) Figure 4 (page 37) shows these facilities with their approximate equivalents in the Guardian environment.
Figure 4 Interprocess Communication Facilities Components to Be Managed 37
Introducing Open System Services
OSS Shared Memory and Semaphores Semaphores allow one process to signal another about events such as the status of data in memory shared by the processes. OSS shared memory and semaphores are implemented using NonStop operating system features and do not require any installation or configuration actions before use. The Guardian environment has separate facilities for shared memory and semaphores that are not discussed in this guide.
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. • One transport-provider process must be running in the system. The transport-provider process provides configuration compatibility for open and close operations on sockets and for nonsensitive SCF commands.
your system. AF_INET sockets provide access through Internet Protocol (IP) version 4 addresses; AF_INET6 sockets provide access through IP version 6 addresses. • For Internet domain sockets using NonStop TCP/IP, the socket application communicates with the $ZTAnn process. The $ZTAnn process in turn provides Internet access through the transport-provider process of TCP/IP. • The AF_INET or AF_INET6 transport-provider process is named $ZTCn, $ZSAMn, or another name chosen by your site.
Figure 5 OSS AF_INET Sockets Servers for NonStop TCP/IP The application processes (Processes 1 through 4) communicate with each other through the transport-provider ($ZTCn) processes, with data being carried back and forth by the transport-agent processes. Transport-agent processes do not communicate with each other directly, but through the transport-provider.
3. 4. The message is forwarded to $ZTA01 (the transport-agent process on processor 1 of Node B, where Process 4 is running). The message is delivered to Process 4 when that process is ready to accept the message. • For NonStop Parallel Library TCP/IP, Internet domain socket application processes exchange data through their embedded library code; however, other processes not managed as part of the OSS environment must be running.
Figure 6 OSS AF_UNIX Sockets Servers (AF_UNIX Release 1) • Figure 7 (page 45)shows these processes for AF_UNIX Release 2 (J06.05 and later J-series RVUs and H06.16 and later H-series RVUs). In this case, the OSS name server ($ZPNS) for the root fileset provides the addressing information used by the OSS sockets local server processes for an AF_UNIX socket when that socket is created by Process 3 in a directory that is part of the root fileset.
Figure 7 OSS AF_UNIX Sockets Servers (AF_UNIX Release 2) For more information about AF_UNIX Release 1, AF_UNIX Release 2, and programming with sockets, see the Open System Services Programmer's Guide.
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 chapter describes what you do to operate the OSS environment that is unique to the OSS environment.
Automatic Startup Service The automatic startup service allows you to configure: • each fileset and its OSS name server • the OSS message-queue server • the OSS sockets local server (for systems running AF_UNIX Release 1 software only) so that they are automatically restarted whenever a system load occurs, when one of those server processes fails, or when a processor failure and reload occurs that affects one of those server processes.
2. Identify and stop all applications currently using the OSS environment: a. To identify all executing OSS processes, enter the OSS shell command: ps -e The output includes: b.
5. After the last fileset stops, stop the OSS Monitor process: • If the OSS Monitor is running as a standard process, enter the following at a TACL prompt: STOP $ZPMON • If the OSS Monitor is running as a persistent process, enter the following at an SCF prompt: ABORT PROCESS $ZZKRN.#ZPMON NOTE: The process device identifier #ZPMON is the convention used throughout this guide; your site might use another naming convention such as #OSMON or #OSSMN.
Example 2 Sample Login Warning for Stopping the OSS File System echo echo echo echo echo echo echo "*****************************************" "* Broadcast message at time *" "* WARNING! *" "* OSS environment *" "* stopping in five minutes. *" "* Please exit now. *" "*****************************************" For information about shell scripts, see the Open System Services User’s Guide.
EMS messages logged through the logger command or the syslog() function are described in the Operator Messages Manual. The OSS subsystem as described in this guide produces the informative and diagnostic messages discussed in “Messages” (page 335). Those messages are returned when you use either the SCF commands for the OSS Monitor or the CVT utility, as discussed under “Managing and Repairing Fileset Catalog Files” (page 170). You start the OSS subsystem by starting the OSS Monitor at least once.
AUTOSTART { AUTO | MANUAL } specifies whether the automatic startup service should start servers or filesets that are configured for automatic startup. If this parameter is specified on the command line, it overrides any previous specification of an AUTOSTART PARAM. This parameter also overrides the current value for the AUTOSTART attribute of the subsystem; see the SCF “ALTER SUBSYS, ALTER MON, and ALTER PROCESS Commands” (page 285)for more information. 1.
CPU FIRST, & AUTORESTART 5, & STARTMODE SYSTEM, & HOMETERM $ZHOME, & DEFAULTVOL $SYSTEM.SYSTEM 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 “Messages” (page 335), for explanation of any console messages received and possible corrective actions.
Table 3 Currently Used TACL PARAMs for the OSS Monitor (continued) PARAM Description value for the FSCKCPU attribute of the subsystem. If this PARAM is supplied after the first execution of the OSS Monitor, it replaces the current value for the FSCKCPU attribute of the subsystem. This value is kept in the ZOSSPARM file and controlled through the SCF ALTER SUBSYS command. See “ALTER SUBSYS, ALTER MON, and ALTER PROCESS Commands” (page 285) for more information about the values that can be specified.
Naming the OSS Monitor Process The OSS Monitor must be given the process name $ZPMON. An attempt to run the OSS Monitor terminates immediately if $ZPMON is already running or if the OSS Monitor is given a process name other than $ZPMON. If the wrong process name is specified, the OSS Monitor abends with completion code 3 (Abnormal, voluntary, but premature termination) and with the Subsystem Programmatic Interface (SPI) error-number token ZSPI^ERR^BAD^NAME.
• For example: ◦ The SCF command: INFO /OUT $S.#srv.info/ SERVER $ZPMON.* returns descriptions of all the servers in the ZOSSSERV file. ◦ The SCF command: STATUS /OUT $S.#file.stat/ FILESET $ZPMON.*user* and the SCF command: STATUS /OUT $S.#file.stat/ FILESET $ZPMON.*USER* both return the status of every fileset in the ZOSSFSET file that contains the string “USER” in its name. Creating Command Aliases: • You can use the SCF ALIAS command to create aliases for OSS Monitor commands.
Obtaining Information About the OSS Subsystem You can use SCF commands to obtain information as described in the following subsections: • “Checking Whether a Subsystem Process Is Running” (page 57) • “Listing the Objects Managed by the OSS Monitor” (page 58) • “Checking the Configuration of the OSS Monitor” (page 58) • “Checking the Version of the OSS Monitor” (page 59) Checking Whether a Subsystem Process Is Running There is no OSS Monitor command that provides the state of an OSS subsystem process
Example 3 Sample SCF LISTDEV Command Display LDev Name . . . 65 $ZZKRN 66 $ZZWAN 67 $ZZW05 68 $ZZW04 69 $ZZSTO 70 $ZZLAN 71 $ZZFOX 78 $ZSNET 79 $ZSMS 86 $ZQ09 . . . 438 $ZTNP0 . . . 580 $ZTC0 581 $ZPMON 587 $ZTC1 . . . PPID BPID Type 0,15 0,274 5,262 4,262 0,275 0,14 7,282 0,15 5,30 9,9 1,30 1,287 (66,0 (50,3 (50,0 (50,0 (65,0 (43,0 (27,0 (66,0 (52,0 (45,0 0,357 1,289 (46,0 ) 4,324 1,372 5,321 5,319 (48,0 ) 32000 170 \NODE1.$SYSTEM.SYS01.TCPIP (24,0 ) 4096 180 \NODE1.$SYSTEM.SYS01.
The information displayed is the configuration currently used by the OSS Monitor. The syntax of the SCF INFO SUBSYS command and an example appear under “INFO SUBSYS, INFO MON, and INFO PROCESS Commands” (page 307). Checking the Version of the OSS Monitor When the OSS Monitor is running, you can use the SCF VERSION command to determine the product-version information for the running copy of the OSS Monitor.
1. Notify all users that the OSS file system is shutting down. Follow the procedure described under “Sending Warnings to Users” (page 49). 2. Use the OSS shell ps command to identify all running OSS processes and then stop those processes with the OSS shell kill command. 3. Use the SCF INFO FILESET * command to locate the mount point of each fileset if you do not maintain a diagram of mount points. 4. Use the SCF STOP FILESET * command.
Monitoring OSS Resources For systems running J06.07 and later J-series RVUs and H06.18 and later H-series RVUs, Event Management Service (EMS) events can help you monitor OSS file-system resource usage. For each type of resource: • Limit warning events occur when resource usage exceeds 85% of the resource limit and are repeated hourly until resource usage drops below 65% of the resource limit. • Error events occur when a resource limit is reached or when an allocation error occurs.
Program file: \YOSFS1.$DATA02.ZYQ00000.Z00022L4 Attempted Allocation Amount: 1 You can translate the Guardian program file name to the OSS filename by using the Guardian FUP INFO command. For example: > fup info \node1.$DATA02.ZYQ00000.Z00022L4,detail $DATA02.ZYQ00000.
Monitoring OSS Processes From the Guardian Environment The TACL STATUS command in the Guardian environment reports information about processes running in the OSS environment: • A STATUS command for an OSS process returns information about the pathname of the executing program file and indicates that the process has an OSS process ID. For example, the TACL command: STATUS 4,354 would return a display similar to Example 4 when that process is the OSS shell.
Managing OSS Processes You can manage the availability of OSS processes in several ways: • Using the Kernel Subsystem, as described in “Making OSS Application Processes Persistent with the Kernel Subsystem” (page 64) • Using OSS tools that schedule periodic tasks, as described in “Using the cron Process” (page 73) • Using Guardian tools that schedule periodic tasks, as described in “Using the NetBatch Product” (page 75) You can manage workload-related behaviors of OSS processes best from the OSS envir
• ◦ STARTUPMSG – Specifies a string of up to 128 characters to pass to OSH ◦ USERID As an alternative to specifying the HOMETERM attribute, you can use OSS shell input/output redirection at the end of the STARTUPMSG value. For example: STARTUPMSG "...
See the osh(1) and run(1) reference pages either online or in the Open System Services Shell and Utilities Reference Manual for more information about the -name flags. • An OSS persistent process does not need to run in the same processor as the OSH process that starts it. The OSH command -cpu flag can be used in the STARTUPMSG attribute to start the OSS persistent process in a specific, currently available processor. For example: STARTUPMSG "-cpu n -p /bin/sh ...
tail -f log from an OSS shell prompt at a dedicated operator console. To make this process persistent, enter the following commands: 1. From a TACL prompt: WHO which produces output that includes the HOMETERM value you need to use, such as: Home terminal: $ZTN0A.#PT4KH30 ... 2. Then, at an SCF prompt for the NonStop Kernel subsystem: ADD PROCESS OSSAPP, NAME $OSH1, AUTORESTART 10, HOMETERM $ZTN0A.#PT4KH30, PRIMARYCPU 1, STARTMODE MANUAL, USERID OSS.APPS, PROGRAM $SYSTEM.SYSTEM.
which produces output that includes the HOMETERM value you need to use, such as: Home terminal: $ZTN0A.#PT4KH30 ... 2. Then, at an SCF prompt, enter: ADD PROCESS OSSAPP, NAME $OSH2, HOMETERM $ZTN0A.#PT4KH30, AUTORESTART 5, PRIMARYCPU 2, STARTMODE MANUAL, USERID OSS.APPS, PROGRAM $SYSTEM.SYSTEM.OSH ASSOCPROC $OSS2, STARTUPMSG "-ls -p /bin/sh startmyapp reload" 3.
1. From a TACL prompt: WHO which produces output that includes the HOMETERM value you need to use, such as: Home terminal: $ZTN0A.#PT4KH30 ... 2. Then, at an SCF prompt, enter: ADD PROCESS OSSAPP, NAME $OSH, HOMETERM $ZTN0A.#PT4KH30, AUTORESTART 5, CPU ALL, STARTMODE MANUAL, USERID OSS.APPS, PROGRAM $SYSTEM.SYSTEM.OSH ASSOCPROC $OSS, STARTUPMSG "-ls -p /bin/sh launchmyapp reload" 3.
Using the nice Command The nice command lets you start an OSS process at a modified priority. All users can lower the execution priority of a process, but only the super ID can increase the execution priority of a process. Priority values in a traditional UNIX system have a significance opposite that of the values used on a NonStop server.
3. Change the PATH environment variable in /etc/profile to something similar to the following: export PATH=/usr/local/script:/bin:/bin/unsupported: /usr/ucb:/usr/bin 4. Post a broadcast message to users, using the technique described in “Manually Stopping the OSS File System and the OSS Environment” (page 47). This procedure can be used to alter the scheduling priority of any process, including those released by HP as part of the OSS environment.
nn is the processor number of the processor you want the process to run in. object_file_path is the OSS pathname for the program object file to be executed as the process. "$@" causes any parameters entered for the process to be sent to the process. 1. Place the script file in a special directory you control, secured so that it can be executed by everyone but not altered by anyone. 2. Add your special directory to the PATH environment variable used to find program files.
The inetd command provides OSS sockets with specific services, sometimes by starting the corresponding server. See the inetd(8) reference page either online or in the Open System Services Shell and Utilities Reference Manual. More information about the inetd command can be found in “Starting a Network Services Server” (page 129). Scheduling Periodic Tasks You can schedule programs or multiple-task OSS shell scripts to run in the OSS environment at predetermined times and intervals.
File Description /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. You can create the working copies of these files by entering the following at an OSS shell prompt before you use the cron process for the first time: cp cp cp cp /usr/lib/cron/at.deny.
The at command and the cron process use these files to determine whether and what commands to run: File Description /var/spool/cron/atjobs/at Lists the commands to be run once. /usr/lib/cron/at.allow Lists the user IDs that are allowed to change the at file. /usr/lib/cron/at.deny Lists the user IDs that are denied access to the at file. You can use the at command if your login name appears in the /usr/lib/cron/at.allow file. If that file does not exist, the at command checks the /usr/lib/cron/at.
terminals cannot be opened unless OSSTTY is also used. OSSTTY can be used to redirect one or more OSS standard files to Guardian EDIT files or Guardian processes. See the osh(1) reference page either online or in the Open System Services Shell and Utilities Reference Manual for more information about starting OSSTTY with the OSH command and “OSSTTY” (page 395) for more information about starting OSSTTY as a server. • To the NetBatch product, the TACL OSH command is an executor program.
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. • The general form of the TACL OSH command for batch execution of an OSS shell script with output routed to OSS files is: OSH [-c] script_path <- >>script_out_file 2>>err_file -c implicitly runs the OSS shell to execute the script as a command. Note that -C (uppercase C) is not a valid command option.
script_out_file is the OSS pathname of the file to receive normal output. If the initial working directory is not the current working directory, a relative pathname is 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 normal output is saved in a file-code-180 file in that subvolume.
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.log and error output in the OSS file /script/processes.err. You would: 1. Create a valid job file for the NetBatch product, named \NODE.$SYSTEM.SYSTEM.
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.
Figure 8 OSS Files and Disk Volumes In contrast, Guardian filenames that are not administered through the NonStop Storage Management Foundation (SMF) are physical names: they mention physical storage devices. A fully qualified Guardian filename includes the name of a system node, a disk volume name, a subvolume name, and a name for the file itself. The relationship of Guardian files to disk volumes is illustrated in Figure 9 (page 82).
Figure 9 Guardian Files and Disk Volumes Each pathname for an OSS regular file has an underlying Guardian filename. The OSS name servers map such pathnames onto corresponding Guardian filenames. This mapping, known as name resolution, is necessary because OSS pathnames are different from Guardian filenames. Each OSS file therefore has two names: • An OSS pathname, such as /usr/henry/workfile • A Guardian filename, such as $VOL2.ZYQ00001.
Figure 10 Guardian Files in the OSS File System 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.
Remote OSS files can be found from the root (/) directory of the remote node. For example, the files in /usr/share/man on the remote node named NODE1 would appear to your system as /E/node1/usr/share/man. Files visible through /E on the remote node are not visible to your system. That is, if you are on NODE0, you cannot access files on NODE0 by looking in /E/node1/E/node0. Your local node also has an entry in its own /E directory.
UTILSGE variable accepts the same values as the -W command flag. To use UTILSGE, you either enter the appropriate one of the following at a shell prompt or add it to your .profile file: export UTILSGE=NOE export UTILSGE=NOG export UTILSGE=NOG:NOE NOTE: Do not enter NOE:NOG. This value for the variable UTILSGE is not valid. 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.
For information on accessing files in the /G directory, see the Open System Services User’s Guide. Relating OSS Files, Filesets, and Disk Volumes You perform OSS file-system configuration and administration tasks through the Guardian environment.
• Can be accessed using the 64-bit APIs such as creat64() and open64() only. Depending on the compilation environment, the 32-bit APIs can be automatically mapped to the 64-bit APIs. The utilities operators use have been enhanced to support large files. Only application programmers must be aware of which APIs to use for large files. • Cannot be restored to systems running RVUs that do not support OSS files larger than approximately 2 gigabytes.
listed in the storage-pool file can be viewed as the creation pool, a subset of the entire storage pool used by the fileset. Figure 11 (page 89) shows the difference between an OSS storage pool and the contents of the storage-pool file for the fileset DATA5; the creation pool is enclosed in a rectangle to indicate that it is the set of disk volumes identified in the storage-pool file.
Figure 11 Storage Pools and Disk Volumes While OSS filesets can span multiple physical disk volumes, individual files cannot. Thus if you allocate for a fileset a disk volume that is running out of space, you might not be able to extend existing files or write new large files on that volume, even if all other available volumes are nearly empty. After a fileset is created, you can check the individual disk volumes in the fileset’s storage-pool file to monitor the disk space that the fileset is using.
4 Managing Servers This chapter 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” (page 90) and “Configuration Files” (page 100) before attempting any operation on an OSS server.
Use of the OSSTTY utility as a server is application-dependent. The OSSTTY server is managed using TACL commands, as described in “OSSTTY” (page 395). Servers in other subsystems use OSS name servers for OSS pathname resolution.
The OSS Name Servers Open System Services supports multiple OSS name server processes. This feature can improve performance by allowing multiple processes to share the task of resolving OSS pathnames. You can run as many OSS name servers as you need on a system simultaneously. Each fileset is managed by only one OSS name server; however, one OSS name server can manage many filesets.
The OSS Sockets Local Servers The OSS sockets local servers vary and depend on whether the system is using AF_UNIX Release 2 software or AF_UNIX Release 1 software. AF_UNIX Release 2 Software Systems running J06.05 and later J-series RVUs and H06.16 and later H-series RVUs are using AF_UNIX Release 2 software. For systems using AF_UNIX Release 2 software, there is an OSS sockets local server 2 (LS2) server for each processor.
OSS File System Resources Overview There are four types of OSS File System (OSS FS) resources: • Memory pools • Bounded • Unbounded • Fileset Some memory pools are allocated from 32-bit addressable memory and some from 64-bit addressable memory. The OSS FS products obtain physical memory for these memory pools as needed. Pool limit errors happen when a memory pool has reached its configured limit.
Table 7 OSS File System Memory Pool Resources Resource Name in OSS EMS Events Measure Counter Name EMS Events Per Processor Limit PXS BUFS PXS-Buf-Bytes Limit error, allocation error, limit warning, and status normal events 72 MB buffer pool in 32-bit addressable memory, beginning with the J06.07 and H06.18 RVUs PXS64 BUFS PXS64-Buf-Bytes Limit error, allocation error, limit warning, and status normal events 128 MB buffer pool in 64-bit addressable memory, beginning with the J06.14 and H06.
Table 8 OSS File System Bounded Resources (continued) Resource Name in OSS EMS Events Measure Counter Name EMS Events Per Processor Limit PIPEFIFO OPENS Pipe-FIFO-Opens Limit error, allocation error, limit warning, and status normal events Maximum of 32,000 opens on unique pipes and/or FIFOs AF_INET SOCKET OPENS AF-INET-Socket-Opens Limit error, allocation error, limit warning, and status normal events Maximum of 32,000 opens on unique AF_INET sockets AF_UNIX SOCKET OPENS AF-UNIX-Socket-Opens L
To display resource usages and limits, use the following name and value combinations for param: RTYPE paramvalue The paramvalue specifies the OSS FS resource type. Keywords are used for paramvalue to designate a single resource or a group of related resources. For lists of single resource keywords see “Single OSS File System Resources for RTYPE paramname” (page 323). For lists of group resources keywords, see “Group OSS File System Resources for RTYPE paramname” (page 323).
Example 7 Configuring FILESET INODES in SCF > ASSUME $ZPMON > ALTER FILESET ROOT, MAXINODES 650000 > CONTROL FILESET ROOT, SYNC NOTE: The inodes (MAXINODES) limit cannot be set to less than the current usage. A warning limit EMS event is generated when the resource reaches the warning threshold. A limit error EMS event is generated when 100% is reached.
inetd is started from an OSS shell command line or script and listens for connections on certain Internet ports. When a connection request is received, inetd decides which service the request corresponds to and invokes a server program to service the request. After the program completes the request, inetd continues to listen. The inetd process allows one process to invoke several others, reducing load on the system.
The named process provides services comparable to the older Guardian-based T6021 DNS product but implements part of the Berkeley Internet Name Domain (BIND) 9 distribution from the Internet Software Consortium (ISC). Two versions of named are available: • Product T0685, a version based upon BIND 9.2.3, without security features • Product T0708, a secure version based upon BIND 9.
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 storage-pool 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.This file must be maintained by the system manager using the SCF ADD FILESET, ALTER FILESET, DELETE FILESET, or RENAME FILESET command.
This attribute increases the risk of losing file labels if the entire disk-process pair for the volume fails. This option is usually used only for a volume that will contain temporary files. When a default ZOSSFSET file is created by the OSS Monitor on a system that was not previously configured, it contains only an entry for a prototype root fileset with the device label 000000. The prototype root fileset does not contain a valid catalog disk volume name or a valid storage-pool file filename.
Attribute Fileset Fileset Fileset Single OSS name server #ZPNS option #ZPNS #ZPNS Multiple OSS name server #ZPNS option1 #ZPNH #ZPNS AUDITENABLED1 OFF OFF OFF BUFFERED1 LOG LOG CREATE DESIREDSTATE2 STOPPED STOPPED STOPPED FSCKCPU3 -1 (same processor as DP2 for fileset catalog) -1 (same processor as DP2 for fileset catalog) -1 (same processor as DP2 for fileset catalog) FTIOMODE4 UNBUFFEREDCP UNBUFFEREDCP UNBUFFEREDCP 30 30 30 500000 500000 500000 NORMALIOMODE OSSBUFFERED
Figure 13 (page 105) shows how the entries in the ZOSSFSET file correspond to disk files, disk volumes, and other attributes in the following list: Attribute Value in Figure 13 Fileset name ROOT Device label 000000 Storage-pool file ZOSSPOOL Catalog volume $ROOTCAT Mount point / OSS name server name #ZPNS AUDITENABLED OFF (default) BUFFERED NONE (default) READ-ONLY FALSE The files $DATA2.ZYQ00000.Z0000435 and $DATA1.ZYQ00000.
Figure 13 Relationship Among OSS Configuration Files, Processes, and Disk Volumes HP strongly discourages changes in node (system) numbers. However, in a few situations, such as during the first startup of a newly delivered system, you must change node numbers. When an Enscribe alternate-key database file such as ZOSSFSET is created, the current node number might be embedded in its ALTFILE attributes. If that happens, the attributes become invalid after the node number is changed.
be resolved, the database cannot be opened by the OSS Monitor at startup and the OSS file system cannot be restarted. To fix the situation, enter the following commands at a TACL prompt: VOLUME $SYSTEM.ZXOSSMON FUP ALTER ZOSSFSET, ALTFILE (0,ZOSSFS00) ALTER ZOSSFSET, ALTFILE (1,ZOSSFS01) EXIT After these commands are executed, FUP INFO, DETAIL output does not contain question marks because the Guardian file system is able to resolve the filename correctly. Subsequent attempts to open ZOSSFSET succeed.
Figure 14 OSS Configuration Files, Processes, and Disk Volumes Affected by Changing ZOSSPARM The ZOSSSERV File The ZOSSSERV file is an Enscribe file containing the OSS Monitor configuration database that stores information about the characteristics of each OSS name server. Entries in this file can be displayed with the SCF INFO SERVER command.This file must be maintained by the system manager using the SCF ADD SERVER, ALTER SERVER, and DELETE SERVER commands.
Attribute Default Values When ZOSSSERV Is Created by Initial Values When ZOSSSERV Was Delivered the OSS Monitor on a Preconfigured System INODECACHE 4096 4096 LINKCACHE 4096 4096 SERVER #ZPNS #ZPNS SQLTIMEOUT 60 60 TYPE NAME NAME Beginning with RVU G06.
For J06.03 and later J-series RVUs, H06.08 and later H-series RVUs, and G06.30 and later G-series RVUs, the other values are the same as the preceding table, but these default values are used for INODECACHE and LINKCACHE: Attribute Preconfigured or Single-Enclosure System Multiple-Enclosure System INODECACHE 64000 64000 LINKCACHE 64000 64000 For J06.15 and later J-series RVUs and H06.
Figure 15 OSS Configuration Files, Processes, and Disk Volumes Affected by Altering an OSS Name Server Entry in ZOSSSERV The Storage-Pool Files The storage-pool files define the disk volumes that each fileset is currently allowed to use when creating new files. The storage-pool files must be created or edited by the system manager using a Guardian text editor. They must be present and must contain valid data before the OSS Monitor is started.
8-character disk volume names to avoid problems with HP NonStop SQL/MP program objects that might reside in the OSS environment.) • One disk volume name can be specified on each line. • Up to 20 entries are allowed in the file. • The disk volumes specified: • • ◦ Cannot be optical disks. ◦ Cannot be administered through the NonStop Storage Management Foundation (SMF). ◦ Should not be in any other storage pool.
Configuration Database Files Used for the OSS Message-Queue Server To start the OSS message-queue server, the OSS Monitor requires a ZOSSSERV file in the subvolume $SYSTEM.ZXOSSMON. The OSS Monitor creates a default ZOSSSERV file with a #ZMSGQ entry that specifies the characteristics of the OSS message-queue server process, $ZMSGQ. To edit this file, use the SCF ALTER SERVER command. Beginning with RVU G06.
Attribute Single-Enclosure System Multiple-Enclosure System MSGMQB 65535 65535 MAXMQID 32 32 MAXMSG 32 * MAXMQID = 1024 32 * MAXMQID = 1024 MAXWAITTIME 0 0 MSGMSIZE 32000 32000 SERVER #ZMSGQ #ZMSGQ TYPE MSGQ MSGQ If your system is upgraded to a G06.17 or later G-series RVU, the OSS Monitor automatically upgrades an existing ZOSSSERV file. Beginning with the J06.17 and H06.
Beginning with the J06.05 and H06.16 RVUs, when the OSS Monitor creates a default ZOSSSERV file, it contains a #ZLSnn entry for every possible per-CPU instance of the OSS sockets local server 2 server process, $ZLSnn. Currently, 16 CPUs are supported (00–15), and the ZOSSSERV entries are #ZLS00–#ZLS15.
Beginning with RVU G06.15, when HP shipped a system with the OSS environment preconfigured or you used the OSSSETUP utility to configure your system and accepted all defaults, ZOSSSERV contained the following initial OSS sockets local server information: Attribute Single-Enclosure System Multiple-Enclosure System BACKUPCPU 0 2 CPU 1 1 SERVER #ZPLS #ZPLS TYPE LOCAL LOCAL Beginning with RVU G06.
Figure 16 OSS Configuration Files, Processes, and Disk Volumes Affected by Altering an OSS Sockets Local Server Entry in ZOSSSERV Configuration Database Files Used for the OSS Transport-Agent Servers When the OSS Monitor creates a default ZOSSSERV file, it contains a #ZTAnn entry for every possible per-CPU instance of the OSS transport-agent server process, $ZTAnn. Currently, 16 CPUs are supported (00–15), and the ZOSSSERV entries are #ZTA00–#ZTA15.
The OSS Monitor updates the OSS resource agent server entries in the ZOSSSERV file each time the OSS Monitor starts. Configuration Files for the Network Services Servers and Tools The following subsections briefly discuss the configuration files for the following programs: • “inetd” (page 117) • “rshd” (page 118) • “portmap” (page 118) • “RPCINFO” (page 119) • “BIND 9 Domain Name Server and Tools” (page 119) The remote execution server, rexecd, does not have a configuration file.
There are two versions of the resolver that you can use on a node. Each version has its own rules for the content of its resolver configuration file: • The BIND 4 version is described in the resolv.conf(4) reference page online and in the Open System Services Shell and Utilities Reference Manual • The BIND 9 version is described in the resolv.
RPCINFO The RPCINFO process uses the same Guardian files as the portmap process, plus an additional one, for a total of 3 files: $SYSTEM.ZTCPIP.HOSTS $SYSTEM.ZTCPIP.RESCONF $SYSTEM.ZRPC.RPC HP provides a default version of the HOSTS and RESCONF files. See the TCP/IP Configuration and Management Manual or the TCP/IPv6 Configuration and Management Manual for more information about those files. $SYSTEM.ZRPC.
You can copy and modify the sample files provided: File Purpose /etc/dns923/named.conf Sample of /etc/named.conf provided for the nonsecure version of named. /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.
Other servers, such as the network services servers, do not need to be added to an OSS configuration database. Such servers can be added to the OSS environment by starting them. See “Starting a Network Services Server” (page 129). Configuring a Server How and when you configure a server depends on the type of server.
2. Use the OSS Monitor SCF ALTER SERVER command to change the ZOSSSERV entry for the OSS message-queue server. a. To reassign the process to a new primary or backup processor, change the corresponding processor entry. If you do not explicitly assign a backup processor, the backup server process is started in the next higher-numbered processor after the processor that runs the primary server process. b.
If processors are added to your node or brought up after the OSS Monitor starts, the OSS Monitor configuration database will not contain entries for the OSS transport-agent servers in those processors. To force the OSS Monitor to update its database and add entries for those servers, you must stop and restart the OSS Monitor. Configuring the OSS Resource Agent Servers You cannot configure an OSS resource agent server. The OSS Monitor SCF ALTER SERVER command is not valid for OSS resource agent servers.
4. If one of the other files does not appear, create a symbolic link to its Guardian equivalent by entering one or more of these commands at the OSS shell prompt: ln -s /G/system/ztcpip/resconf resolv.conf ln -s /G/system/ztcpip/hosts hosts Beginning with the J06.16 and H06.27 RVUs, the OSS Core Utilities product (T1202) provides the extended Internet daemon, xinetd. This Open Source super-server daemon manages Internet-based connectivity and offers a more secure extension to inetd.
See the rshd(8) reference page either online or in the Open System Services Shell and Utilities Reference Manual for information about setting up the rshd process. See “Managing Security” (page 216), for more information about setting up aliases and initial working directories. rexecd To configure the rexecd process: 1.
ln ln ln ln -s -s -s -s /G/system/ztcpip/networks networks /G/system/ztcpip/protocol protocols /G/system/ztcpip/services services /G/system/ztcpip/hosts hosts To use the full NonStop TCP/IPv6 addressing capabilities instead of the /etc/hosts file, add the following NonStop operating system DEFINE to the TACLLOCL file that is executed whenever a TACL session is started: DEFINE =TCPIP^HOST^FILE, FILE $SYSTEM.ZTCPIP.
To create sockets in compatibility mode, choose $ZPLS. To create sockets in portability mode, choose $ZAFN2. NOTE: Neither $ZPLS nor $ZAFN2 is a process that runs on the system. The DEFINE determines the initial value of the AF_UNIX socket transport provider. If the DEFINE is not present, then sockets are created in compatibility mode (transport provider $ZPLS).
d. 3. fileset are started. See the “STATUS FILESET Command” (page 314) for a description of the output. Issue the OSS Monitor SCF START FILESET command for any unstarted filesets identified in the previous step. Issue the OSS Monitor SCF START FILESET command for the OSS name server’s fileset and for any unstarted filesets that are mounted on it. The OSS name server starts automatically when the first of its filesets starts. Beginning with the J06.15 and H06.
2. Enter an OSS Monitor SCF START SERVER command. For example: START SERVER $ZPMON.#ZTA15 Starting an OSS Resource Agent Server An OSS resource agent is started automatically when its processor starts. You do not need to start an OSS resource agent server for it to collect OSS File System resource information in a given processor. To restart a stopped OSS resource 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.
Internet-based connectivity and offers a more secure extension to inetd. For more details, see “OSS Core Utilities User Commands” (page 243). Obtaining Information About a Server You can obtain the following information: • Whether a server is running in the Guardian environment, as described in “Determining Whether a Server Is Running” (page 130).
Example 8 Sample SCF LISTDEV Command Display LDev Name PPID BPID Type RSize Pri . . . 142 $XMIOP 0,31 ( 6,4 ) 80 190 162 $Z003 0,48 (50,63) 3900 150 191 $ZLMG 1,36 0,34 (56,63) 132 149 192 $ZEXP 1,37 0,55 (63,30) 132 149 193 $ZNET 1,38 0,57 (50,63) 3900 149 215 $ZPMON 0,275 (24,0 ) 1024 149 222 $ZM00 0,65 (45,0 ) 132 201 . . . 228 $ZTC0 1,47 0,63 (48,0 ) 32000 180 More text? ([Y],N) y 229 $ZTNT 1,49 (46,0 ) 6144 155 \NODE1.$SYSTEM.ZTCPIP.TELSERV Program \NODE1.$SYSTEM.SYS00.XMIOP \NODE1.$SYSTEM.SYS00.
you can determine the following information for all servers administered through the OSS Monitor: • The type of server (OSS name server, OSS message-queue server, and so on) • The processor used by the primary server process • The processor used by the backup server process The following additional information is displayed when you specify the DETAIL option in the INFO SERVER command: • The user who created the server configuration, when it was created, the user who last modified the configuration, a
Stopping a Server How and when you stop a server depends on the type of server. • “Stopping a Specific OSS Name Server” (page 133) • “Stopping the OSS Message-Queue Server” (page 134) • “Stopping an OSS Sockets Local Server” (page 134) • “Stopping an OSS Transport Agent Server” (page 134) • “Stopping an OSS Resource Agent Server” (page 135) • “Stopping a Network Services Server” (page 135) Other servers used by OSS applications require separate procedures.
The fileset can be deleted as soon as it stops. • Reassign an existing fileset to the unstoppable OSS name server; use the OSS Monitor SCF ALTER FILESET command on the fileset to change the NAMESERVER attribute for the fileset. Stop that fileset; for example, if you decide to use the TEMP fileset to fix this problem, enter: STOP FILESET $ZPMON.TEMP Remember to reassign the fileset to its proper OSS name server and restart it again.
Identify and stop all OSS applications that use OSS AF_UNIX or OSS AF_INET sockets in the processor served by the OSS transport agent that you want to stop. An OSS transport-agent server will not shut down if it is servicing an open socket. Use an OSS Monitor SCF STOP SERVER command to stop the OSS transport-agent server. For example: STOP SERVER $ZPMON.#ZTA15 Stopping an OSS Resource Agent Server To stop an OSS resource agent server: 1. Make sure you are a member of the super group (255, nnn). 2.
DESIREDSTATE INODECACHE LINKCACHE MAXWAITTIME SQLTIMEOUT NOTE: Beginning with the J06.15 and H06.26 RVUs, you might also need to reconfigure OSS SEEP-related attributes for the OSS name server. For details of these attributes, see “ALTER SERVER Command” (page 278). Only a member of both the SOA group and Super group are allowed to set or alter the OSS SEEP-related attributes for the OSS name server. 3.
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. The OSS name server should stop as soon as the last of its filesets stops. 9. If the OSS name server you want to restart with changed attributes is not the OSS name server for the root fileset: a.
5. Locate and stop all OSS applications that use AF_UNIX sockets. When you stop the OSS sockets local server to restart it using the new configuration, the server notifies all applications using AF_UNIX sockets by closing their open sockets. Depending on how an application has been coded, this action might cause the application to fail. 6. Use the following OSS Monitor SCF command to stop the server: STOP SERVER $ZPMON.#ZPLS 7.
3. 4. 5. 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. Warn your users to make sure that all their files in affected filesets are closed and all OSS shell sessions using those filesets are terminated. You can use a method similar to the one described under “Manually Stopping the OSS File System and the OSS Environment” (page 47).
5 Managing Filesets This chapter describes how to perform operations on Open System Services (OSS) filesets.
1. Make sure that the mount-point directory exists; if it does not, create it and give it universal access permission. HP recommends against the use of symbolic links as mount-point names or within mount-point names; such use can cause problems when filesets are restarted.
3. Add a record for the fileset to the ZOSSFSET file by using the SCF ADD FILESET command. Select the appropriate settings for the fileset: • HP suggests that you select a consistent name for the fileset, directory mount point, and report file filename to make administration of the fileset easier. • If the fileset is frequently used, consider having it started or restarted automatically after a system load or processor failure.
The fileset becomes inaccessible until it is stopped, repaired using the SCF DIAGNOSE FILESET command (see “Checking and Repairing Fileset Integrity” (page 159)), and restarted. – Fileset recovery information is not recorded in the PXLOG catalog file (see “FSCK Log File” (page 160)). – After a double failure of an OSS name server, when the OSS Monitor attempts to remount a fast-create fileset, the remount request is rejected by the OSS name server.
• Keeps the fileset catalog on disk volume $CAT Figure 17 OSS Configuration Files, Processes, and Disk Volumes Involved in Adding a Fileset • Only checkpoints what would be PXLOG file entries • Allows up to 500,000 files (the default MAXINODES value) in the fileset • 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 u
• Is unrestricted by default • • Accepts the FTIOMODE and NORMALIOMODE default settings for file opens Start the fileset with the SCF START FILESET command. Figure 18 (page 145) shows how a fileset is mounted at a mount point when the fileset is started. The figure illustrates the effects of the following command: START FILESET $ZPMON.
To define the storage-pool file that is associated with a fileset record: 1. Use a Guardian text editor to create the storage-pool file. The storage-pool file must be in the same subvolume ($SYSTEM.ZXOSSMON) as the ZOSSFSET file. 2. 3. Enter the names of disk volumes that will contain the OSS files in this fileset, one disk volume name on each line. See “The Storage-Pool Files” (page 110) for the rules about specifying disk volumes in storage-pool files. Exit the editor.
Example 9 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. Multiple OSS FILESET objects may share a common POOL file.
If the OSS Monitor finds that a fileset was left in the STARTED state but its OSS name server process is not running, one of the following might have occurred: • A serious problem has occurred. • A fileset was not properly stopped before the OSS Monitor last stopped. • An OSS name server failed while the OSS Monitor was not running. That fileset might need repair.
Both the primary and backup OSS name server processors for a fileset can fail during the restart. If both the primary and backup OSS name server processors for a fileset fail, the OSS Monitor checks processor messages until one of the OSS name server processors is reloaded, then initiates the recovery sequence. You do not need to take any action. If a restart operation fails for a fileset, that fileset, which was in the STARTED state, is changed to the UNKNOWN state by the OSS Monitor.
OSS security auditing allows you to audit access to objects in the OSS filename space. Audit commands for OSS objects and operations are provided by Safeguard, and SAFEART allows you to search for audit records of operations on OSS files. Using the AUDITENABLED Attribute The OSS fileset AUDITENABLED attribute determines whether audit records are generated on objects within the fileset.
Beginning with the J06.15 and H06.
2. Do one of the following: • If the fileset is the root fileset, stop all filesets by entering the following OSS Monitor SCF command: STOP FILESET $ZPMON.* • If the fileset is not the root fileset, use the SCF STOP FILESET command. For example, to stop (unmount) the fileset USER1 and send informational messages to the file CMDLOG, enter the following SCF command: STOP /OUT CMDLOG/ FILESET $ZPMON.
• Catalog file volume (CATALOG attribute) • OSS file-system mount point (MNTPOINT attribute) • OSS name server identity (NAMESERVER attribute) • OSS SEEP protection (SEEPPROTECTED attribute) NOTE: OSS SEEP is supported beginning with the J06.15 and H06.26 RVUs. For details of OSS SEEP, see “Using an OSS Security Event-Exit Process (SEEP)” (page 231). To change any of these attributes, change the appropriate value in the ZOSSFSET file by using the SCF ALTER FILESET command.
Table 10 Effects of File I/O Fault-Tolerance Attribute Settings (continued) Setting Scenarios Results buffered data is written to diskApplication fails with possible loss of multiple buffered write requests if failure occurs before buffered data is written to disk DP2BUFFERED (the same behavior as Single DP2 processor failure System legacy OSS file caching turned off failure (or double DP2 processor [OSSCACHING OFF]) failure) Application processor failure Application fails with possible loss of multiple
Changing the OSS File System Mount Point Changing the mount point can affect the behavior of programs that use OSS files, particularly programs that use the symbolic-link feature. You should notify users well in advance of changing an existing mount point. Changing OSS File Caching for the Disks of a Fileset NOTE: HP recommends that you not modify OSS file caching at the disk level.
This command will stop all filesets in the correct order. This command begins with the last fileset mounted and stops the filesets in the reverse order in which they were last started. 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.
NOTE: HP recommends that the S_NONSTOP extension not be used in new applications. The S_NONSTOP extension is ignored on nodes running an RVU more recent than G06.26, H06.03, and all J-series RVUs; the standard O_SYNC feature used with the FTIOMODE attribute provides better data integrity and improved performance.
2. Do one of the following: • Use the SCF CONTROL FILESET command with the SYNC option to make the change take effect immediately. • Stop the fileset with the SCF STOP FILESET command, as described under “Stopping (Unmounting) a Fileset” (page 151), then restart the fileset using the SCF START FILESET command. The change takes effect when the fileset is restarted.
8. Restore the previously archived files to the mount point using the OSS shell pax command. For example: pax -rvf oss_files.pax 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 OSS name server process (both primary and backup) or a disk volume containing a fast-create fileset. Note that even with the fast-create option enabled, FSCK is needed only in cases of multiple failures and is run automatically by the OSS Monitor in most of these cases. NOTE: Before you use the DIAGNOSE FILESET command, make sure that the fileset you are about to check is stopped (unmounted). The OSS Monitor displays an error message if you attempt to diagnose a fileset that is not stopped.
Example 10 FSCK Log File Examples 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.
If the DETAIL option is also used, the mounted field contents are displayed in hexadecimal. Possible values are: Value Mounted state 0x0000 The fileset is unmounted. 0x0001 The fileset is being recovered. 0x0002 The fileset is mounted in BUFFERED LOG mode by an OSS name server that is not running the current code version. 0x0006 The fileset is mounted in BUFFERED CREATE mode by an OSS name server that is not running the current code version. 0x2004 The fileset is mounted in BUFFERED CREATE mode.
Table 11 Inconsistencies Checked by FSCK (continued) Inconsistency Type Explanation and only major inconsistencies are being corrected, FSCK does not create a new catalog but does: • Reset the mounted flag to zero. • Clear the bitmap of inconsistencies reported by an OSS name server. • Add any lost inode numbers to the appropriate free list. 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.
Table 11 Inconsistencies Checked by FSCK (continued) Inconsistency Type Explanation Orphan ZYQ file Minor There is a ZYQ file for which there is no inode.If the ZYQ file is empty, FSCK issues a warning message and purges the file. If the file is nonempty, FSCK obtains the current number of links for the file from the disk process.If the number of links is zero and the file is not currently open, FSCK purges the file. If the link count is nonzero, FSCK does the following: 1.
Table 11 Inconsistencies Checked by FSCK (continued) Inconsistency Type Explanation catalog, the ZYQ file is restored to its original name and the PXINODE file is updated to reflect the new CRVSN of the file. 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.
If the FSCK run fails, the fileset that FSCK is checking enters the UNKNOWN state rather than the STOPPED state. If this happens, check the FSCK log file (described under “FSCK Log File” (page 160)). If FSCK failed before writing anything to this file, the file is probably full. Either rename the file or copy it to another volume, then purge the original file. Deleting a Fileset If your site uses the OSSREMOV utility, that utility deletes all filesets.
Table 12 Configuration File Upgrades (continued) Old Configuration File (D40 Version of Old Configuration File (D46 Version of New Configuration File (G09 or Newer OSS Monitor) OSS Monitor) Version of OSS Monitor) ZPOSFS01 ZOSSFS01 ZPOSSERV ZOSSSERV ZPOSPARM ZOSSPARM Removing Older Configuration Files If you have upgraded from an earlier RVU, you might be able to save some disk space by deleting old database files. Whether you should delete files depends on whether you are: • “Upgrading From a G05.
5. Create a fileset named HOME, using a unique OSS name server, that has a DEVICELABEL of 000001. The use of multiple OSS name server processes can improve overall performance if the default name server, $ZPNS, is very busy. a. Start SCF. 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.
3. In SCF, stop the ROOT fileset by using the following commands: STOP FILESET $ZPMON.ROOT 4. Display and record all the configuration information for the ROOT fileset: INFO FILESET $ZPMON.ROOT, DETAIL 5. Delete the ROOT fileset and then redefine it using the recorded information: a. At a TACL prompt, go to the volume that is listed as CATALOG for the ROOT fileset: VOLUME $OSS.ZX000000 b. Delete all the files in the ROOT fileset by using the FSCK utility: FSCK PURGE 6.
2001-07-26 13:41:50 \NODE1.$ZPMON TANDEM.OSS.D40 STATE changed from Stopped to Started because of Request. 2001-07-26 13:41:50 \NODE1.$ZPMON TANDEM.OSS.D40 STATE changed from Started to Stopped because of Unmount by OSS Monitor. 4. -00003 $A , Operator -00003 $A , Automatic When you are unsure of the effect of an action on a fileset, use the STATUS FILESET command DETAIL option to obtain the last internally reported error information for it. Look up that error in “Numbered Messages” (page 361).
created on systems running J06.03 and later J-series RVUs, H06.08 and later H-series RVUs, and G06.29 and later G-series RVUs are created as Version 3 catalog filesets. To upgrade a fileset catalog file: 1. Determine whether the fileset is mounted by using the SCF STATUS FILESET command. 2. If the fileset is mounted, stop the fileset by using the procedure described under “Stopping (Unmounting) a Fileset” (page 151). 3.
$vol is the name of the disk volume where the ZOSS subvolume is located. 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.
CVT - T8621G09 - (01FEB01) OSS Catalog Volume Tool Copyright Compaq Computer Corporation 1994, 1995, 1996, 2001 CVT { RENAME [ TO ] } { PURGE SERIAL [IN ] } where and are: { CURRENT } [ IN ] {SERIAL [ IN ] } CVT Examples Here are some examples of CVT commands: • To purge the files PXIN0004, PXLI0004, and PXLO0004 on subvolume $VOL.ZX000003, enter the following command at a TACL prompt: RUN $SYSTEM.ZOSS.
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” (page 177). • Remove obsolete files, as described in “Removing Obsolete OSS Files and Directories” (page 181) and “Erasing and Removing OSS Files” (page 195).
Interpreting Guardian Filenames for OSS Files OSS files have a distinctive form of Guardian filename in the Guardian environment. In addition to the usual node name and volume name information, the Guardian filename for an OSS file has the following form: • Subvolume names begin with ZYQ. These subvolume names correspond to OSS filesets. The digits that follow ZYQ are the device identifier for the fileset within the ZOSSFSET database file used by the node.
NOTE: In filename, you must put another backslash character before the backslash (\) and dollar sign ($) characters or else the shell interprets these characters with their special shell meanings rather than as plain characters. Example 12 shows examples of using pname. When the file is on another NonStop node, the pathname of the file relative to the / directory on that node is displayed after the prefix of /E/ and the node name. Example 12 OSS pname Command Examples $ pname \\NODE1.\$VOL.ZYQ00000.
Example 13 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 $VOL.ZYQ00000.Z0000DV4 26 Jul 1994, 13:24 OSS PATH: /usr/henrysp/stuff OWNER 254,254 SECURITY: -rw-rw-rwCREATION DATE: 2 Jun 1994, 10:38 ACCESS TIME: 19 Jul 1994, 19:52 EOF 142 5> fup info \NODE1.$DATA.ZYQ00001.Z0000DV2, detail \NODE1.$DATA.ZYQ00001.
them when necessary. If you do not select Manage OSS Files, NonStop Software Essentials or DSM/SCM just installs the files needed to set up the OSS environment into the TSV. • The TACL COPYOSS macro, which automatically runs the Guardian PINSTALL utility on all the archive files in the subvolume. • The PINSTALL utility, which you can run manually for individual archive files. The use of either COPYOSS or PINSTALL also requires you to use the OSS Pcleanup utility to remove obsolete files.
COPYOSS is used when: • A system is initially set up manually • The OSSSETUP utility invokes it • You do not use NonStop Software Essentials or DSM/SCM to install and maintain your OSS product files but a major upgrade requires you to load multiple new or revised pax archives into the OSS file system For example: • To load all the OSS product files using the COPYOSS macro, enter these commands from a TACL prompt: VOLUME $tsvvol.
chmod a=rwxt /tmp /usr/tmp /var/tmp /var/preserve /usr/tandem/sqlmx/USERMODULES • Under rare conditions, the PINSTALL command used by COPYOSS can return an error message that indicates a disk file could not be found; this message can be ignored when the named file is an empty directory in the corresponding pax 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.
• Files installed by PINSTALL are not always secured in conformance with the best practices at your site. You should always resecure files installed by PINSTALL.
NonStop Software Essentials and DSM/SCM automatically deletes directories left empty after obsolete files are removed but the Pcleanup utility does not. To remove unneeded directories after using Pcleanup, use the OSS shell rmdir command CAUTION: Invisible OSS files with names containing .dsmscm are used by NonStop Software Essentials or DSM/SCM for OSS file/configuration management and should never be deleted.
Backing Up and Restoring OSS Files This subsection describes the tasks for backing up and restoring the OSS environment and user files on your system. OSS files can be backed up with the OSS pax command or the BRCOM interface for Backup and Restore 2.0. An entire OSS fileset can be backed up by using its mount point as the current directory when running either utility.
NOTE: You cannot restore an OSS file that is larger than approximately 2 gigabytes to a system running an RVU that does not include support for OSS large files. Guardian Files and the pax Utility You should not use the pax utility to back up or restore files in /G. Such files should be backed up and restored using TMF or a version of Backup/Restore. See “Backing Up User Files” (page 186). Unless otherwise noted, the pax utility makes no distinction between Guardian and OSS files.
where n is a Guardian file-system error number. A single archive can span more than one reel. The pax utility issues this message during reel switching: Mount next tape to continue? Backing Up the OSS Environment Using a Version of Backup/Restore To back up the OSS environment: 1. Record all configuration information for the OSS environment.
4. Backup and Restore 2.0 preserves OSS file access information. However, if you are using the Backup and Restore utilities (T9074), you need to record all user ID 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 USER *.
NOTE: Do not begin find or pax operations at the root (/) directory. find and pax perform recursive operations within directories unless you use the UTILSGE=NOG:NOE environment variable or the -W NOG and -W NOE flags. The / directory contains /G and /E. If you begin an operation that recursively processes directories at /, you can unintentionally process an entire Guardian file system and the operation will take a very long time to finish.
The blocksize argument can have values no greater than 32,256 for disk archives and 28,672 for tape archives. The blocksize argument can be specified as a series of digits (0 through 9) followed by a flag letter, “b” or “k”: • If “b” is used, the blocksize value is multiplied by 512. • If “k” is used, the blocksize value is multiplied by 1024. For example, “10b” translates to a blocksize value of 5120 bytes (10 * 512). The default blocksize value for cpio archive format is 10b (5120 bytes).
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 2.0 You can specify part or all of multiple directory hierarchies and selectively restore files from within them using Backup and Restore 2.0.
',^//*usr//*,,' translates into an ed replacement string as follows: ' The opening and closing single quotation marks (') enclose the substitution expression. , The first comma (,) indicates the end of the omitted line address argument. The second and third commas indicate the beginnings of the two omitted arguments. ^ The circumflex (^) means search from the beginning of the line. / The first slash (/) means search for the first occurrence of a slash.
2. Use the following form of the TACL DSAP command to determine the size of the fileset: DSAP oss_volume_name, BYSUBVOL You need to do this for all volumes that contain files with the subvolume name for the fileset. To determine the list of volumes with subvolumes, you can use the command: FUP SUBVOLS $*.ZYQnnnnn For the HOME fileset example, the commands would be: FUP SUBVOLS $*.
Backing Up OSS Files to Other Expand Nodes You can perform a remote backup of OSS files indirectly by using the pax utility and writing the backup archive to a directory within the Guardian file system (a directory within/G), where the archive file is a Guardian file. For more information on creating a backup archive in a /G directory, see “Creating a pax Backup of OSS Files in the Guardian File System” (page 190).
4. 5. 6. 7. 8. 9. 10. 11. 12. 13. Stop all disk volumes used for OSS catalogs and OSS data files on the source system. Use the BACKUP VOLUME mode option to backup the OSS catalog disk volume from the source system. Use the BACKUP VOLUME mode option to backup the OSS data file disk volumes from the source system. Use the RESTORE VOLUME mode option to restore the OSS catalog disk volume to the target system. Use the RESTORE VOLUME mode option to restore the OSS data file disk volumes to the target system.
1. 2. 3. Restore the security database files using the RESTORE MAP NAMES and OPEN options as new files in a temporary location (subvolume). Use the File Utility Program (FUP) to rename the old security database files and to move the restored copies to the correct subvolume. Stop and restart Safeguard and all Expand lines to close the old security database files and open the restored ones.
pax -r -v -f archive-name -W clobber match-pattern -r reads an archive file from the standard input file. -v writes archive member pathnames to the standard error file. -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.
In a UNIX environment, these three files are usually associated with the user’s terminal: standard input is read from the terminal keyboard, standard output is sent to the terminal display, and standard error is an output logging mechanism that is usually 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.
Figure 20 (page 197) illustrates the behavior when OSSTTY is run as a server named $ZTTY before starting an application program explicitly designed to use it, such as mysample3 in the following command set: RUN OSSTTY / NAME $ZTTY, IN $VHS,OUT $DATA.OSSLOG.EDITFIL, TERM $VHS, NOWAIT / -server OSH -p "/usr/mysample3" For information about using OSSTTY through the OSH command, see the osh(1) reference page either online or in the Open System Services Shell and Utilities Reference Manual.
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. Use the SCF ALTER FILESET command MAXINODES option to change that value. The change takes affect as soon as the fileset is started.
7 Managing Terminal Access This chapter 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” (page 199) • “Configuring Telserv Access” (page 199) • “Configuring FTP Access” (page 203) How Users Gain Access to the OSS Environment A user gains access to the OSS environment through a server process.
Example 15 Telserv Login Using Default Telserv Services Connecting.. .Connected to 123.456.789.111. Escape character is '^]'. WELCOME TO node1.subdom1.corporate.net [PORT $ZTCF0 #23 WINDOW $ZTNF0.#PTMNBVY] TELSERV - T9553D40 - (29JUN2001) - (IPMADI) Available Services: TACL EXIT Enter Choice> TACL TACL 1> logon guest.user1 Password: Welcome to \NODE1. The time is 8:37:07 am 08/21/2002.
Example 16 Telserv Login Using an OSS Shell Direct Login Service Connecting... Connected to node1.subdom1.corporate.net. Escape character is '^]'. WELCOME TO node1.subdom1.corporate.net [PORT $ZTCF0 #23 WINDOW $ZTNF0.#PTMNBVY] TELSERV - T9553D40 - (29JUN2001) - (IPMADI) Available Services: OSS TACL EXIT Enter Choice> OSS login: guest.user2 Password: /G/SYSTEM/SYSTEM: In this example, the OSS service allows the user to log in directly to the OSS environment.
Example 17 Telserv Login Using a Site-Written Direct Login Service Connecting... Connected to node1.subdom1.corporate.net. Escape character is '^]'. WELCOME TO node1.subdom1.corporate.net [PORT $ZTCF0 #23 WINDOW $ZTNF0.#PTMNBVY] TELSERV - T9553D40 - (29JUN2001) - (IPMADI) Available Services: OSH3 TACL EXIT Enter Choice> OSH3 login: guest.any Password: Hello: Considerations: • Specify the SCF ADD SERVICE command with ACCESS set to SYSTEM.
Configuring FTP Access Configuration of FTP server access is fully described in the TCP/IP Configuration and Management Manual and the TCP/IPv6 Configuration and Management Manual. Whether an anonymous user of FTP initially enters the Guardian environment or the OSS environment depends on which anonymous login user name is used and whether an OSS initial directory is configured for that user name. “Managing Security” (page 216), shows how to set up an OSS user for anonymous FTP access.
8 Managing Users and Groups This chapter provides the following user and user-group management concepts that are relevant to the Open System Services (OSS) environment: • “How Users Gain Access to the OSS Environment” (page 204) • “User and User-Group Attributes” (page 205) • “Assigning an Initial Working Directory” (page 206) • “Assigning an Initial Program” (page 209) Beginning with the J06.16 and H06.
The FTP server provides access in the following ways: • Directly to the Guardian environment, when either of the following is true: ◦ No initial working directory is configured for the user of the FTP client. ◦ An initial working directory with a name that begins with the characters /G/ is configured for the user of the FTP client. The FTP user must use an FTP client quote OSS command to gain access to the OSS environment.
• User groups can exist independently of user definitions. Therefore: ◦ A user group can be created before any users are added to it. ◦ 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.
Creating an Initial Working Directory in the OSS Environment NOTE: HP strongly recommends that you create a separate fileset for initial working directories. To create an initial working directory in the OSS environment, use either the mkdir command from within the OSS environment or the OSH command from within the Guardian environment.
Example 18 TACL Macro to Configure an OSS User ?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.
You can also assign an initial working directory 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 before using the OSH command: PARAM HOME pathname where pathname is the pathname of an existing OSS directory to be used as the initial working directory.
alias-name is an existing alias for an existing user definition. /code-dir/program is the initial program, expressed as an OSS pathname. For example, the following SAFECOM command assigns the initial program /user/deml.o to the alias donl of the user definition with the user name scribes.DONALD: ALTER ALIAS donl, INITIAL-PROGRAM /user/deml.o The assignment process can be partially automated by creating a TACL macro file and then editing that file appropriately for each new user.
• “userdel Command” (page 212) – deletes an existing user or an alias and optionally removes related files • “newusers Command” (page 212) – reads a file that contains user or alias names, and their attributes, to create or update users or aliases To view full details of the above OSS commands, you can use any of the following methods: • Use the -h option for the command.
Example 22 usermod Command – Change Home Directory for an Existing User The following usermod command changes the home directory for the sdev user. usermod –d /home/users/newdev sdev Example 23 usermod Command – Modify Expiration for a User's Alias The following usermod command modifies the expiration date to July 31, 2014 for the sdevUK alias. usermod –A sdevUK –e 2014–07–31 userdel Command The userdel command deletes an existing user and optionally removes related files.
Licensing the OSS Monitor to the Super Group The OSS Monitor contains privileged procedures but is not a licensed program. If anyone other than the super ID (255,255 in the Guardian environment, 65535 in the OSS environment) attempts to start it, the TACL error message Unlicensed privileged program is issued. HP strongly suggests that you do not license the OSS Monitor. In most cases, the only user of the OSS Monitor should be the system administrator (or other authorized person).
Allowing Anonymous FTP Use To configure FTP for anonymous access to the OSS file system, use the Safeguard product. For example, enter SAFECOM commands similar to: ADD USER NULL.FTP, 0,254, OWNER 255,255, PASSWORD guest ALTER USER NULL.FTP, GUARDIAN DEFAULT SECURITY OOOO ALTER USER NULL.FTP, GUARDIAN DEFAULT VOLUME $guest.ftp ADD ALIAS anonymous, 0,254, PASSWORD guest ALTER ALIAS anonymous, GUARDIAN DEFAULT SECURITY OOOO ALTER ALIAS anonymous, GUARDIAN DEFAULT VOLUME $guest.
• • The initial working directory for an anonymous FTP user should be set up to have appropriate security in the OSS environment for the user as a type “other” user of the directory. ◦ A read-only anonymous user would have the OSS file security for the directory set to “drwxrwxr-x”. ◦ A write-only anonymous user would have the OSS file security for the directory set to “drwxrwx-wx”.
9 Managing Security This chapter provides information about: • “Common and Unique Characteristics of OSS and UNIX Security” (page 216) • “Differences Between OSS and UNIX User and User-Group Configuration” (page 222) • “OSS Security Auditing” (page 223) • “Protecting Your System” (page 225) Common and Unique Characteristics of OSS and UNIX Security Basic file security is the same for the OSS environment as on a UNIX system.
Beginning with the J06.15 and H06.26 RVUs, an OSS Security Event-Exit Process (SEEP) is supported and provides additional file-access authorization. For details of OSS SEEP, see “Using an OSS Security Event-Exit Process (SEEP)” (page 231). Certain differences might require you to code the security-management portions of a shell script in a manner specific to the OSS environment.
tftpdaccess.cf ◦ In any of various directories: .plan .project sendmail.cf ◦ The /home/quotas file, used to establish user disk space quotas Some of these files and directories also provide mechanisms that intruders can use to compromise UNIX system security or integrity.
tty ◦ In /etc: hosts hosts.equiv inetd.conf install_obsolete magic named.boot printcap and printcap.sample profile profile.sample protocols resolv.conf services termcap ◦ In /var: /adm/cron/.proto /adm/cron/cron.allow /adm/cron/cron.deny /adm/cron/queuedefs< preserve spool/cron spool/pcnfs tmp ◦ In the user’s home directory: .netrc .rhosts Take normal security precautions with these files and directories.
finger (a TACL FINGER command exists for the Guardian environment) ftpd fsck fsirand getstats identd in.
Many of these user and group names can provide mechanisms that intruders can use to compromise UNIX system security or integrity. 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.
Figure 21 Major Components and Interfaces for OSS Security Management See the Security Management Guide for an overview of both Guardian and OSS security. See the Safeguard Audit Service Manual for a description of the SAFEART program and the information logged for audited OSS files. Beginning with the J06.15 and H06.26 RVUs, the OSS name server consults the OSS SEEP for operations on files in OSS SEEP-protected Version 3 catalog filesets, if the OSS SEEP is running.
• Perform auditing for OSS files and the OSS environment’s filesets • Control access to the /E directory using the REMOTEPASSWORD attribute • Use the OSS EasySetup product This guide illustrates the use of the Safeguard product in the Guardian environment. You create a user definition with the SAFECOM ADD USER command, and you create a user group definition with the SAFECOM ADD GROUP command. A NonStop operating system user can have alternate user names, called aliases.
Audit Records for OSS Objects Audited events are recorded in the Safeguard audit files (collectively referred to as the audit trail).
The name logged in an operation depends on the type of object being audited. Formats are: Object Format OSS fileset $ZPMON.Znnnnn:yyyymmddhhmmss, where nnnnn is the fileset device number and yyyymmddhhmmss is the local civil time (LCT) when the fileset was created. Example: $ZPMON.Z00000:19980119152451 OSS regular file (disk file) $vol.ZYQnnnnn.
OSS Shell Commands Useful for Security Administration This subsection covers the following topics: • “Creating a Logon Session With the su Command” (page 226) • “Displaying Your User Login Name” (page 226) • “Changing Your User Group” (page 226) NOTE: • Restricted-access filesets deny the super ID special access privileges. • The getfilepriv and setfilepriv commands can be used to list and to set file privileges for executable files, user libraries, and ordinary DLLs.
Use of suid Scripts Some scripts, known as suid scripts, enable users to perform some activities that require appropriate privileges; that is, the scripts could be used by an intruder to assume the identity of the super ID (255,255 in the Guardian environment, 65535 in the OSS environment). As a result, it is a good practice to make sure that there are no such script files on your system. As installed, the OSS environment does not contain suid script files.
used by the undocumented RPC interfaces underlying such products as the Network File System (NFS) for Open System Services. • To protect sensitive files from access by the super ID, place those files in a restricted-access fileset. For more information about restricted-access filesets, see “Using Restricted-Access Filesets and File Privileges” (page 228). • Use the shred command to permanently erase data and optionally remove sensitive files.
File Privilege File Attribute On systems running J06.11 or later J-series RVUs or H06.22 or later H-series RVUs, files have an additional file privilege attribute that specifies which special privileges, if any, a file has when accessing files in a restricted-access fileset.
PRIVSOARFOPEN File Privilege The PRIVSOARFOPEN file privilege allows a process to directly access any file in a restricted-access fileset on the local system, but only if that executable file has been started by a locally-authenticated member of the Safeguard SECURITY-OSS-ADMINISTRATOR (SOA) group.
Identifying Attempts to Break Security Checking the file system for changes in the ownership and permissions of important files and directories can reveal the presence of an intruder. You can monitor permissions by entering the following OSS shell command periodically: ls -alt pathname > file pathname is the OSS pathname for the mount-point directory of a fileset you want to monitor (such as /bin and /etc). file is the OSS filename of a file to receive the output.
The following fileset attribute is supported for OSS SEEP management: • SEEPPROTECTED { ON | OFF } — Specifies whether the fileset has OSS SEEP protection set or not. The default value is OFF. OSS SEEP consultation for access authorization happens only if the SEEPENABLED attribute of the OSS name server is set to ON. The following OSS name server attributes are supported for OSS SEEP management: • SEEPENABLED { ON | OFF } — Specifies whether the OSS SEEP needs to be running or not.
Starting an OSS SEEP The OSS name server starts and manages the OSS SEEP. Each time the OSS name server starts the OSS SEEP, it first attempts to kill any process with the same process name. An EMS message notes this action. If the OSS name server determines that the OSS SEEP was not started, it does not send messages to the OSS SEEP until it successfully kills the process and restarts it. Be careful to avoid name collisions.
The backup OSS name server does a close of the OSS SEEP when the primary OSS name server closes the OSS SEEP, which results in the graceful stop of the OSS SEEP. OSS SEEP-Related EMS Events The OSS name server generates EMS events to $0 and $ZLOG under the following conditions: • The OSS SEEP is successfully started. The EMS event reports the OSS SEEP state change from STOPPED to RUNNING. • The OSS SEEP is stopped or an OSS SEEP death occurs.
10 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 chapter 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.
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 .profile File Each user’s environment can be set up by means of a .profile file. This file is executed automatically every time a user logs in. Although a .profile file is not essential, it can make use of the OSS environment easier. A default .
Example 27 Sample /etc/profile File /etc/profile.
When the MANPATH variable is not defined, the man command searches directories in the following order: /usr/share/man/man* /usr/local/man/man* /usr/share/man/cat* /usr/local/man/cat* nroff Formatting Tool Beginning with the J06.16 and H06.27 RVUs, the nroff formatting tool is available with the OSS Core Utilities product.
Table 14 Localization Environment Variables Variable Meaning and default LANG The locale of your shell, which consists of a language, territory, and code set. The The The The The default default default default default locale is the C locale. language is English. code set is ASCII. territory is US. code set is ASCII. LC_ALL The behavior for all aspects of the locale, unless overridden by another variable. LC_COLLATE The collating sequence. The default value is C/POSIX.
Table 15 Locale Names and Filenames (continued) Language of locale Filename for locale definition Finnish fi_FI.ISO8859-1 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.SJIS Japanese-English, Japan en_JP.
Monitoring the OSS Environment With the Shell Potential problems that you might want to monitor using the shell include: • Slow performance • Overuse of resources The examples given in this subsection are not exhaustive. You might also want to remove files from directories that expand automatically, as discussed in “Controlling the Growth of Directories” (page 241). Slow Performance Slow performance might be the result of many processes left running that are no longer being used.
Here are some examples of using the find command to locate and delete files: • Delete all files in /var/tmp whose names begin with TMP that have not been accessed in thirty days: find /var/tmp -name 'TMP*' -atime -30 | xargs rm • Delete all files that are named a.out or whose names end in .o that have not been accessed in sixty days: find / -W NOG -W NOE \( -name a.out -o -name '*.
Executing Remote Shell Commands The rsh command executes a specified shell command remotely. It executes commands at the host system where the commands are to be run. The rsh command sends standard input from the local host to the remote command and receives standard output and standard error file data from the remote command. NOTE: If the remote host is a NonStop server, you must specify the -l flag and provide your password in clear text form.
The OSS Core Utilities comprises the following Open Source software: • • • On NonStop systems running J06.14 or later J-series RVUs or H06.25 or later H-series RVUs, the following Open Source software is provided: ◦ GNU Core Utilities (Coreutils) – see http://www.gnu.org/software/coreutils/. ◦ GNU compression and archival tools: – bzip2 – see http://bzip.org/. – gzip – see http://www.gnu.org/software/gzip/. – tar – see http://www.gnu.org/software/tar/.
coreutils/share/man directory must precede /usr/share/man in the MANPATH environment variable, as in the following example: export MANPATH=/usr/coreutils/share/man:/usr/share/man For more information about viewing OSS reference pages, see “Controlling Reference Page Searches and Display” (page 237).
11 Managing OSS Devices This chapter 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.
Specifying a Default Printer You specify a default printer: • By specifying the Guardian spooler-location name for a system default printer in the /etc/printcap file. Specifying a system default printer allows you to control which printer is used for high-volume print jobs. • By specifying the Guardian spooler-location name for a default printer for a specific user in one of the following ways: ◦ Specify the Guardian spooler-location name for the chosen printer in the user’s printcap file.
Example 28 Sample /etc/printcap File /etc/printcap.sample default ps titan land build \COMM.$S.#DEFAULT \COMM.$S.#POST7 \COMM.$S.#TITAN7 \COMM.$S.#LAND7 \COMM.$S.#BUILD You assign aliases other than that of the default printer so users can refer to printers by names that are more convenient than the Guardian spooler-location names. For example, you might want to use the alias titan in place of the Guardian spooler-location name \COMM.$S.#TITAN7.
12 Managing Problems Most operational problems are easily resolved by following the recovery recommendations listed in “Messages” (page 335). However, some of the messages indicate that a problem should be reported to HP. This chapter 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.
/G/SYSTEM/ZOSS/PINSTALL Binder timestamp: Version procedure: Target CPU: AXCEL timestamp: 17MAR95 15:38:44 T8626D30_26MAY95_OSSUTL_AAB UNSPECIFIED 17MAR95 15:39:02 For help with interpreting the information in VPROC displays, see the vproc(1) reference page either online or in the Open System Services Shell and Utilities Reference Manual. To use the vproc command, you must first know the location of the file for which you want information.
13 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 “Managing Filesets” (page 140), “Managing Servers” (page 90), and “Operating the OSS Environment” (page 46).
• The possible causes of each error • The recommended action to recover from each error You invoke the help facility from an SCF prompt by entering HELP OSS. This command produces the display shown in Example 29. Example 29 SCF HELP OSS Display The Open Systems Services (OSS) subsystem is HP’s open computing interface implementation. It provides an open interface to the NonStop operating system for supporting portable applications. OSS conforms to the POSIX.1 and POSIX.
Example 30 SCF HELP Command OSS Menu +------------------------------OSS Menu-------------------------------------+ | | | Commands: | | ADD ALTER CONTROL DELETE DIAGNOSE INFO NAMES RENAME | | START STATUS STOP VERSION | | | | Objects: | | FILESET NULL MON PROCESS SERVER SUBSYS | | Attributes: | | | | | | Error Numbers: | | [ E | W ] [ - ] | | | +------------------------Next Menu Selection--------------------------------+ | | | Options: | | QUIT MAIN RETURN | | | +-----------------------------------
Example 32 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: . . . More text? ([Y],N) If you answer “Y” to the prompt for more text, you get more text or, if there is no more relevant text, you are returned to the previous screen (in this case, the SCF HELP command OSS FILESET menu).
processor halt, the OSS Monitor restarts the OSS name server when the processor becomes available again. The OSS Monitor remembers the state of all filesets (DIAGNOSING, STARTED, STOPPED, or UNKNOWN) if it stops and is restarted, even after a system load. Software Requirements SCF is required for the OSS environment, and the version of SCF must be compatible with that of the OSS PM.
“DELETE SERVER Command” (page 292) “DIAGNOSE FILESET Command” (page 293) “INFO FILESET Command” (page 296) “INFO SERVER Command” (page 301) “INFO SUBSYS, INFO MON, and INFO PROCESS Commands” (page 307) “NAMES Command” (page 309) “RENAME FILESET Command” (page 312) “START FILESET Command” (page 312) “START SERVER Command” (page 313) “STATUS FILESET Command” (page 314) “STATUS SERVER Command” (page 322) “STOP FILESET Command” (page 331) “STOP SERVER Command” (page 332) “VERSION SUBSYS, VERSION MON, and VERSIO
[ , SEEPPROTECTED { ON | OFF } ] 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 specifies the name of the OSS fileset to be added. filesetname can contain up to 32 alphabetic or numeric characters. The first character must be a letter. Only names not currently defined in the ZOSSFSET file are valid values. Fileset names are not case-sensitive.
For restricted-access filesets: • The default value for this attribute is DISABLED. • You can alter this attribute only if you are logged on to the local system as a user ID that is a member of the Safeguard SECURITY-PRV-ADMINISTRATOR (SPA) security group and a member of the super group (255, nnn), but is not the super ID (255, 255) or a member of the Safeguard SECURITY-OSS-ADMINISTRATOR (SOA) group. BUFFERED { CREATE | LOG | NONE } specifies the amount of catalog write buffering used for the fileset.
If the processor indicated by a nonnegative processor value is not available, then the processor specified for the SUBSYS FSCKCPU option is used. If that processor is also unavailable, the processor in which the OSS Monitor is running is used. If an FSCKCPU option is omitted from the FILESET configuration, then the processor specified for the SUBSYS FSCKCPU option is used when a fileset recovery is automatically started.
on the NonStop server using the actual standard OSS permissions or OSS ACL permissions (if present) on the object. The attribute values for NFSPERMMAP are: Value Description RESTRICTIVE The other and owning group fields of the permissions bits returned to NFS V2 clients are modified such that only access that would be granted to everyone in the ACL, excluding the owner, is granted in the permissions bits.
See the Open System Services NFS Management and Operations Guide for more information about NFS. 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: Value Description UNBUFFEREDCP Use unbuffered input/output with checkpointing. This behavior provides maximum fault tolerance but with reduced performance.
SEEPPROTECTED { ON | OFF } specifies whether the fileset has OSS SEEP protection set or not. ON Specifies that the fileset is OSS SEEP-protected and OSS SEEP needs to be consulted for file operations under this fileset that require access authorization. Only a member of both the SOA group and Super group can change the setting of this attribute. OFF Specifies that the fileset is not OSS SEEP-protected. The default value is OFF.
DP2BUFFERED OSSBUFFERED • 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.
Examples • To add the fileset USER1 with the storage-pool file ZOSSPOOL using the volume catalog on $DATA1 and the mount point /user1, enter this command: ADD /OUT CMDLOG/ FILESET $ZPMON.USER1, CATALOG $DATA1, & POOL ZOSSPOOL, MNTPOINT "/user1", & BUFFERED LOG, NAMESERVER #ZPNS This command: • ◦ Sends informational messages to the file CMDLOG. ◦ Specifies the fileset name as USER1 and the mount point as /user1.
• ◦ Specifies that the catalog files for the fileset USER3 are stored on the disk volume $DATA2. ◦ Assigns the fileset USER3 to the OSS name server process $ZPNS by default. ◦ Uses the BUFFERED LOG specification by default. ◦ Specifies that the inode cache for the fileset USER3 should be flushed to disk after approximately 10 seconds to make its recovery faster by decreasing the time when the cache is considered dirty.
[ , SEEPPRI [ priority ] ] [ , SEEPPARAMTEXT [ startup-param-text ] ] 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 added. server_processname has the following form: [ $ZPMON.]servername servername specifies the name of the server to add.
BACKUPCPU backup_processor specifies the processor number of the processor where the backup server process will run. backup_processor is either an integer in the range 0 through 15 or equal to -1. The value -1 specifies that the OSS name server will not run as a process pair. If the BACKUPCPU parameter is omitted, the processor for the backup server process is unchanged.
nseconds can have a value in the range 1 through 900. There is no option to wait indefinitely. The default value is 5 seconds, and a null entry for nseconds resets this attribute to the default value. SEEPPROGFILENAME prog-filename specifies the name of a Guardian local file containing the Guardian program to be run when the OSS SEEP needs to be started. This attribute must be defined before the SEEPENABLED attribute can be set to ON.
message-queue server, and OSS transport-agent servers are started by the START SERVER command. • You must be a member of the super group (255, nnn) to use this command. • The ADD SERVER command is intended to be used for OSS name servers of filesets other than the root fileset. • When Open System Services is first installed, the OSS Monitor adds the default OSS name server for the root fileset the first time the OSS Monitor is run.
The syntax of the ALTER FILESET command is: ALTER [ /OUT filename1/ ] FILESET [$ZPMON.
OFF Security-sensitive operations against this fileset are not logged in the Safeguard security audit trail. For filesets that have a RESTRICTEDACCESS attribute of DISABLED, only the super ID (255,255) can change the value of the AUDITENABLED attribute.
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. If the DESIREDSTATE option is omitted, the previous value for the fileset is unchanged. FSCKCPU processor specifies the processor number of the processor that should run the FSCK program when a fileset recovery is automatically initiated, where processor is in the range 0 through 15 or is -1.
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.
Value Description Setting NFSPERMMAP to this value guarantees that users who have read permission in the OSS ACL for an object on the NonStop system will be able to read the object on NFS V2 clients. However, it also allows users on NFS V2 clients who do not have read permission in the OSS ACL for an object on the NonStop server to be able to read data from the object when the data is cached on NFS V2 clients.
POOL filename2 specifies the Guardian filename of the storage-pool file that contains the volume list to use for file creation for the fileset. The specified file must exist and must reside on the same Guardian volume and subvolume as the ZOSSFSET and ZOSSSERV files. Storage-pool filenames are not case-sensitive.
Value Description (255, nnn), but is not the super ID (255, 255) or a member of the Safeguard SECURITY-OSS-ADMINISTRATOR (SOA) group. See also “Considerations” (page 276). SEEPPROTECTED { ON | OFF } specifies whether the fileset has OSS SEEP protection set or not. ON Specifies that the fileset is OSS SEEP-protected and OSS SEEP needs to be consulted for file operations under this fileset that require access authorization.
SECURITY-PRV-ADMINISTRATOR (SPA) security group and a member of the super group (255, nnn), but is not the super ID (255, 255) or a member of the Safeguard SECURITY-OSS-ADMINISTRATOR (SOA) group. For more information about restricted-access filesets, see “Using Restricted-Access Filesets and File Privileges” (page 228). • You can use the ALTER FILESET command on a fileset that is not in the STOPPED state. However, the changes do not take effect until the fileset is stopped and restarted.
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. • OSS SEEP considerations (J06.15/H06.26 and subsequent RVUs): ◦ Only a member of both the SOA group and Super group are allowed to set or alter the SEEPPROTECTED attribute.
This command is not valid for: • OSS transport-agent servers • OSS local server 2 servers (#ZLSnn) • OSS resource agent servers The syntax of the ALTER SERVER command is: ALTER [ /OUT filename/ ] SERVER server_processname { [ , AUTORESTART ntimes ] [ , BACKUPCPU backup_processor ] [ , BACKUPCPUOK { TRUE | FALSE } ] [ , CPU primary_processor ] [ , DESIREDSTATE { STARTED | STOPPED } ] [ , INODECACHE size1 ] [ , LINKCACHE size2 ] [ , MAXMQID maxmqid ] [ , MAXMSG maxmsg ] [ , MAXWAITTIME seconds2
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: AUTORESTART ntimes specifies the persistence count of the server, where ntimes is an integer value in the range 0 through 10. The persistence count is the number of times the OSS Monitor will automatically restart the server during a 10-minute period.
DESIREDSTATE { STARTED | STOPPED } specifies the desired end state of an OSS sockets local server or the OSS message-queue server when the SUBSYS AUTOSTART AUTO feature is used. STARTED Attempts are made to start the server when the OSS environment is restarted. STOPPED No attempt is made to start the server during a restart of the OSS environment. If the DESIREDSTATE option is omitted, the previous value for the server is unchanged.
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.
The default value for this attribute is 5 seconds, and a null entry for nseconds resets this attribute to the default value. SEEPPROGFILENAME prog-filename specifies the name of a Guardian local file containing the Guardian program to be run when the OSS SEEP needs to be started. This attribute must be defined before the SEEPENABLED attribute can be set to ON. If the SEEPENABLED attribute is set to ON and an attempt is made to set this field to null, the command is rejected.
• 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 “PARAMs Used by the OSS Monitor” (page 53)). If the TACL PARAM BACKUPCPU is not defined or is invalid, the OSS message-queue server automatically picks an available processor with the next higher or lower processor number than the primary processor.
process and starts the OSS SEEP with that name. An EMS event is generated to notify the stopping of the running process.
If object-type is omitted, process must be specified unless ASSUME PROCESS $ZPMON was previously specified. process specifies the manager process of the subsystem whose parameter values you want to alter. process has the following form: process_name $ZPMON is the only supported value for process_name. process_name must be specified unless ASSUME PROCESS $ZPMON or ASSUME $ZPMON was previously specified. You can specify the following options in any order.
The initial value (set when the configuration database is created) is blank. The value of NULL or blank indicates that the output file should be a file created in the same volume and subvolume as the OSS Monitor OSSMON object file. The created file is a file-code-180 file (a C language text file) with a file identifier that consists of the characters ZX0 followed by the device label of the fileset processed by the command that launched the FSCK utility: $SYSTEM.SYSnn.
CONTROL FILESET Command The CONTROL FILESET command can do either or both of the following: • Reset the highwater mark for inode use by the fileset to the number of inodes currently in use • Update the fileset attributes used by an OSS name server for a mounted fileset from the current values in the ZOSSFSET configuration file The syntax of the CONTROL FILESET command is: CONTROL [ /OUT filename1/ ] FILESET filesetname { [ , RESET MAXINODEUSED ] [ , SYNC ] } OUT filename1 specifies the name of a Guardi
Considerations • The CONTROL FILESET command can be used only by super-group users (255,nnn). • When you use the SYNC option and the fileset has the READONLY TRUE attribute, the BUFFERED attribute value in the ZOSSFSET file is ignored and the OSS name server uses the BUFFERED NONE attribute. • When you use the SYNC option and the fileset has the BUFFERED CREATE attribute, the POOL attribute is ignored and the catalog disk volume is the only disk volume used for the storage pool.
OUT filename1 specifies the name of a Guardian output file for informational messages. You can either read this file with a text editor or display it with the FUP COPY command. SERVER server_processname specifies the server being altered. server_processname has the following form: [$ZPMON.]servername servername specifies the name of a running OSS name server to be altered. Only names currently defined in the ZOSSSERV file are valid values. The first character of the name must be a pound sign (#).
from OFF to ON, the OSS SEEP is started if there are any OSS SEEP-protected filesets mounted. ◦ The new values of the SEEPPROGFILENAME, SEEPPROCESSNAME, SEEPSWAPVOL, SEEPCPU, SEEPPRI, and SEEPPARAMTEXT attributes will be used when the OSS SEEP gets started the next time. ◦ If the CPU specified through the SEEPCPU attribute is not available at startup, the OSS SEEP is created using the backup CPU specified in the first word of the startup-param-text.
CONTROL SERVER $ZPMON.#ZPNS, RESET MAXINODECACHEUSED DELETE FILESET Command The DELETE FILESET command removes an OSS fileset from the configuration administered through the OSS Monitor. This command affects the contents of the ZOSSFSET file. The syntax of the DELETE FILESET command is: DELETE [ /OUT filename/ ] FILESET [$ZPMON.]filesetname OUT filename specifies the name of a Guardian output file for informational messages.
Considerations • You must be in the super group (255, nnn) to use this command. • You can delete only OSS name servers that are not running. • You cannot delete an OSS name server that is associated with a fileset. You must first alter or delete that fileset so that it is not associated with the OSS name server you want to delete. • You cannot delete the OSS name server for the root fileset. • You cannot delete an OSS resource agent server.
If you do not specify a processor by either means, the default processor is the processor on which the copy of DP2 used for the fileset catalog is running. DETAIL produces a detailed report and places it in the FSCK log file. If the DETAIL option is omitted, only a summary report is produced. OPTION { START | STOP } starts or stops a diagnosis by the FSCK utility. Value Description START Starts a new diagnosis and sets the state of the fileset to DIAGNOSING.
When the REPAIR option is specified, you cannot specify the STATUS, UPGRADE, or DOWNGRADE option. UPGRADE reformats the catalog for the fileset from an earlier version to a current version. For example, you can • Upgrade a D30-version catalog (Version 1 fileset) to a D40-version catalog (Version 2 fileset) to add support for symbolic links. • Upgrade a Version 2 catalog fileset to a Version 3 catalog fileset to add support for OSS ACLs.
can be remounted with the SCF START FILESET command unless serious inconsistencies remain unrepaired. • You can use the SCF STATUS FILESET command to determine whether the integrity check is complete. • If FSCK fails, the fileset is placed in the UNKNOWN state. • The FSCK utility checks for the inconsistencies listed in Table 11 (page 162). • The DIAGNOSE FILESET command can be used only by super-group users (255,nnn).
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 Table 5 (page 55) for the definition of UNIX wildcard characters.) 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: DETAIL produces a detailed report.
*MntPoint-FilesetName.. *MntPoint-NameServer... *MntPoint.............. *AuditEnabled.......... *DesiredState.......... *FsckCPU............... *Report................ *MaxDirtyInodeTime..... *MaxInodes............. *FTIOmode ............. *NormalIOmode ......... *RestrictedAccess...... *SEEPProtected.........
NFSPermMap PermMap-value (J06.09 and later J-series RVUs or H06.20 and later H-series RVUs only) Specifies how the permissions for an OSS object protected by optional access control list (ACL) entries are mapped to the permissions (rwxrwxrwx) used by Network File System Version 2 (NFS V2) clients on open, read, write, and directory search operations. Write permissions are always enforced on the NonStop server using the actual standard OSS permissions or OSS ACL permissions (if present) on the object.
AlterBy user-name2 is the user name of the user who last modified the configuration or state of the fileset. AlterTime time2 is the timestamp for the last time that the fileset configuration was changed, in the form dd mmm yyyy hh:mm:ss.mil. MntPoint-FilesetName filesetname is the name of the fileset on which this fileset is mounted. MntPoint-NameServer servername2 is the server name of the OSS name server for the fileset identified by filesetname.
SEEPProtected status indicates whether the fileset has OSS SEEP protection set or not. Value Description OFF Indicates that the fileset is not OSS SEEP-protected. ON Indicates that the fileset is OSS SEEP-protected and OSS SEEP needs to be consulted for all operations under this fileset that require access authorization.
servername specifies the name of the server whose information is to be displayed. Only names currently defined in the ZOSSSERV file are valid values. The first character of the name must be a pound sign (#). servername can contain wildcard characters. (See Table 5 (page 55) for the definition of UNIX wildcard characters.) The $ZPMON prefix can be omitted if you have previously specified $ZPMON in an SCF ASSUME command. You can specify the following options in any order: DETAIL produces a detailed report.
processor2 is the processor number of the processor where the backup server process runs. The returned value is from the database; it might not reflect the actual processor of the running process because the active process might have been switched or recreated its backup process in the next available processor, as implemented in the OSS message-queue server.
because the active process might have been switched or re-created its backup process in the next available processor. For the OSS message-queue server, the value -1 specifies the backup process will be started on the next available CPU in the node. For all other OSS server types, the value -1 indicates the server will not run as a process pair. CreateBy user-name1 is the user name of the user or process that created the initial configuration of the server.
MSGMSIZE msgmsize is the maximum number of bytes allowed for a message. Valid values are in the range 1 through: • 32000 – Prior to the J06.17/H06.28 RVUs. • 2000000 (~2 MB) – Beginning with the J06.17/H06.28 RVUs. This parameter is displayed only for the OSS message-queue server. BackupCPUOK choice specifies whether the server can be started in its backup processor during a restart when its primary processor is unavailable. Valid values are TRUE or FALSE.
OSS Detailed Info SERVER \NODE1.$ZPMON.#ZPNS Type.................... *CPU..................... *BackupCPU............... CreateBy................ CreateTime.............. AlterBy................. AlterTime............... *Params.................. *BackupCPUOK............. *MaxWaitTime............. *DesiredState............ *AutoRestart............. *PersistenceCount........ NAME 1 0 $ZPMON 12 Dec 2001 13:28:28.331 SUPER.SUPER 03 Mar 2002 01:29:14.
NOTE: This example is valid for systems beginning with the J06.14 and H06.25 RVUs only. Example Report for the OSS Message-Queue Server, #ZMSGQ To obtain a configuration report of the OSS message-queue server, #ZMSGQ, and send the report to the CMDLOG file, enter the following SCF command: INFO /OUT CMDLOG/ SERVER $ZPMON.#ZMSGQ, DETAIL A display such as the following is written to CMDLOG: OSS Detailed Info SERVER \NODE1.$ZPMON.#ZMSGQ Type................... *CPU.................... *BackupCPU............
DETAIL produces a detailed report. If the DETAIL option is omitted, only a summary report is produced.
FsckCPU processor is the processor number of the processor in which the FSCK program runs when the SCF DIAGNOSE FILESET or DELETE FILESET command is used and either no processor is configured for the fileset or the processor configured for the fileset is unavailable. Report filespec is the Guardian filename or spooler location of the default FSCK log file. ZOSSVol 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.
OUT filename specifies the name of a Guardian output file to receive the list of objects. You can either read this file with a text editor or display it with the FUP COPY command. object-type specifies the types of objects whose names are to be listed. object-type is one of the following: MON | PROCESS | SUBSYS | FILESET | SERVER MON, PROCESS, and SUBSYS object types all specify the OSS Monitor.
servername ... node is the Expand name of the node on which the subsystem is running. filesetname is the name of an OSS fileset. servername is the server name of a configured OSS server. 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.
#ZRA10 #ZTA04 #ZTA14 #ZRA11 #ZTA05 #ZTA15 #ZRA12 #ZTA06 #ZRA13 #ZTA07 #ZRA14 #ZTA08 #ZRA15 #ZTA09 #ZTA00 #ZTA10 #ZTA01 #ZTA11 #ZTA02 #ZTA12 #ZTA03 #ZTA13 RENAME FILESET Command The RENAME FILESET command changes an existing fileset name. The syntax of the RENAME FILESET command is: RENAME [ /OUT filename/ ] FILESET [$ZPMON.]filesetname, [$ZPMON.]newname filename specifies the name of a Guardian output file for informational messages.
Considerations • The START FILESET command can be used only by super-group users (255,nnn). • The mount point for the specified fileset must be in a fileset that is already started. • If the OSS name server for the fileset is not already running, it is started when the fileset is started. • The MAXINODES attribute is ignored when an OSS name server tries to mount a fileset.
NOTE: On systems running AF_UNIX Release 2 software, the OSS sockets local server 2 servers are started when the processors are loaded. • The transport-agent servers are started when the processors are loaded. • The resource agent servers are started by the OSS Monitor when the system is coldloaded or when the processors are loaded. • The START SERVER command can be used only by super-group users (255,nnn).
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 Description DIAGNOSING The fileset is being diagnosed by the FSCK utility. STARTED The fileset is started (mounted). STOPPED The fileset is ready to be started (mounted) or diagnosed. UNKNOWN The fileset is in an unknown state.
*NFSTimeout............................. time4 *NFSPool................................ number1 *NFSPermMap..............................PermMap-value *MaxInodes.............................. number2 InuseInodes............................ number3 MaximumCountInodesUsed................. number4 TimestampMaximumCountInodesUsed........ time5 TimestampMaxiximumCountInodesReset..... time6 *FTIOmode .............................. ftiomode *NormalIOmode........................... normaliomode NumVols.............
AlterAfterSyncOrMount status1 indicates whether the configuration of the fileset has changed after its most recent start or synchronization with ZOSSFSET: Value Description FALSE The fileset configuration has not changed. TRUE The fileset configuration has changed. This field is displayed only if the fileset is in the STARTED state. LastError error is the error number of the last OSS Monitor error that occurred during an operation on the fileset.
ReadOnly status3 indicates whether the fileset is mounted for read-only access: Value Description FALSE The fileset has read and write access. TRUE The fileset is mounted for read-only access. This field is displayed only if the fileset is in the STARTED state. NFSTimeout time4 is the number of seconds that the OSS name server retains the results of nonretryable Network File System (NFS) operations for the fileset. This field is displayed only if the fileset is in the STARTED state.
Value Description permission in the OSS ACL for an object on the NonStop server to be able to read data from the object when the data is cached on NFS V2 clients. UNMODIFIED The other and user fields of the permissions bits returned to NFS V2 clients are unmodified. The group field of the permissions bits returned to NFS V2 clients are the permissions of the class entry of the ACL.
Volumes: volume_list1 lists the names of the disk volumes in the fileset’s storage pool. If the fileset is in the STARTED state, the complete volume list for the storage pool is displayed; otherwise only the creation pool volume list is displayed. The following fields are displayed only when the fileset is in the STARTED state: VolumesElligibleForFileCreation: volume_list2 lists the names of the disk volumes in the fileset’s storage pool that have space available for new file creation.
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.
$OSS5 4238 2735.44 64 4 2735.40 $OSS6 (unavailable -- device error 201 ) 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.USER1 A display such as the following is written to CMDLOG: OSS Status FILESET USER1 FilesetName USER1 State STARTED LastError 0 ErrorDetail 0 STATUS SERVER Command The STATUS SERVER command displays status information about a server administered by the OSS Monitor.
You can specify any of the following single resource keywords for paramvalue when using RTYPE.
Table 17 Group OSS File System Resources for RTYPE paramname (continued) Group Resources Keyword for RTYPE paramname OSS File System Resource Disk cache buffers Pipe/FIFO data buffers 32-bit PXS buffers 64-Bit global priv addressable buffers AF_INET_SOCK AF_INET socket opens AF_INET socket request-wait operations AF_INET socket data buffers AF_UNIX_SOCK AF_UNIX socket opens AF_UNIX file-descriptors passing operations AF_UNIX socket data buffers BUFFERS AF_INET socket data buffers AF_UNIX socket data
any paramvalue, an error message is displayed listing the possible values for the missing paramvalue. If resource information is not available for any resource (due to an error returned by the resource agent server), then blank values are displayed for current and peak usage information for that resource. 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.
InodeCacheInUse........ MaxInodeCacheUsed...... TimeMaxInodeCacheUsed.. TimeMaxInodeCacheReset. *LinkCache.............. LinkCacheInUse......... MaxLinkCacheUsed....... TimeMaxLinkCacheUsed... TimeMaxLinkCacheReset.. *SEEPEnabled............ SEEPState.............. SEEPLastError.......... *SEEPParams............. *SEEPParams............. *SEEPParams............. *SEEPParams............. *SEEPParams............. *SEEPParams............. *SEEPParams.............
StartTime time1 is the timestamp for the time when the server was last started, in the form dd mmm yyyy hh:mm:ss.mil. LastError error1 is the error number of the last OSS Monitor error that occurred during an operation on the server. See “Numbered Messages” (page 361) for an explanation of a specific error number. ErrorDetail error-detail is the error number of the Guardian file-system error (if any) reported with the error value.
TimeMaxInodeCacheReset 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. LinkCache size4 indicates the number of links that the OSS name server is allowed to cache. LinkCacheInUse size5 indicates current link cache usage by the OSS name server, displayed as the number of entries in use and as a percentage of size4.
SEEPPARAMTEXT startup-param-text indicates the startup parameter text in the startup message sent to the OSS SEEP. Considerations • The DETAIL display for an unstarted OSS name server differs from that of a started OSS name server. If an OSS name server is not in the STARTED state, its inode cache information, link cache configuration, and OSS name server attributes are not displayed. • Beginning with the J06.14 and H06.25 RVUs, the param field is valid only for OSS resource agent servers.
NOTE: • This example is valid for systems running AF_UNIX Release 2 software only. To obtain summary information about all configured servers and send informational messages to the file CMDLOG, enter the following command: STATUS /OUT CMDLOG/ SERVER * • Beginning with the J06.14 and H06.25 RVUs, to obtain detailed information about the ALL_OPENS resource for the $ZPMON.#ZRA01 resource agent server, enter the following command: > STATUS SERVER $ZPMON.
ResourceLimit is the maximum limit of the resource being displayed. If the resource is a buffer then the value is in Megabytes, otherwise, the value is an absolute integer value. CurrentUsage is the current usage value (either in Megabytes or in absolute value) of the resource and also the current usage percentage with respect to the resource limit of the resource.
STOP /OUT CMDLOG/ FILESET $ZPMON.* STOP SERVER Command The STOP SERVER command stops an OSS server. The syntax of the STOP SERVER command is: STOP [ /OUT filename/ ] SERVER server_processname OUT filename specifies the name of a Guardian output file for informational messages. You can either read this file with a text editor or display it with the FUP COPY command. SERVER server_processname specifies the server being stopped. server_processname has the following form: [$ZPMON.
The syntax of the VERSION SUBSYS, VERSION MON, and VERSION PROCESS command is: VERSION [ /OUT filename/ ] [ object-type ] [ processname ] [ , DETAIL ] OUT filename directs any output generated for this command to the specified file. You can either read this file with a text editor or display it with the FUP COPY command. object-type specifies the types of objects whose version information is to be returned.
subsystem-name1 is the name of the SCF subsystem associated with the object. version-information-banner provides the product-version information for the software named on the same display line. See “Managing Problems” (page 249), for more information about product-version information and the related Guardian VPROC utility.
A Messages Status, warning, and error messages are sent to the console by: • The Event Management Service (EMS). These messages are describe under “Event Management Service (EMS) Event Messages” (page 335). • The Open System Services (OSS) EasySetup utilities. These messages are described under “OSS EasySetup Utility Messages” (page 335). • The Guardian Catalog Volume Tool (CVT) utility. These messages are described under “CVT Messages” (page 336).
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.ZSPIDEF.ZEMSTACL or $SYSTEM.ZSPIDEF.ZOSSTACL. Cause The utility cannot find the indicated event definition file, a Subsystem Programmatic Interface (SPI) file containing the Event Management Service (EMS) event definitions that the utility needs to issue event messages.
CVT Error Messages CVT issues the following error messages, which are listed in alphabetic order. After issuing an error message, CVT terminates. ***Command Error*** - token token identifies the invalid parameter. Cause The value token appears where CVT expects something else. Effect CVT terminates. Recovery Check the command and correct it as necessary. FILE_OPEN_ Error error on file filename error is a Guardian file-system error value returned by the underlying Guardian FILE_OPEN_ procedure call.
Cause The Guardian FILE_RENAME_ procedure call could not be completed. Effect CVT did not rename the indicated file. Recovery Check the command and correct it as necessary. ***fsck needed -- subvolume.PXCKSTAT exists*** subvolume identifies the affected subvolume. 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.
Cause The value token followed the keyword SERIAL in the command line but is not a valid FSN value. 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” (page 165) for more information about FSN values. ***Invalid Subvolume Name*** - token token identifies the invalid subvolume name.
Effect None. Recovery Informational message only; no corrective action is needed. *** NO INCONSISTENCIES DETECTED -- EXISTING CATALOG RETAINED *** Cause FSCK detected no inconsistencies and retained the current catalog files. Effect None. A new catalog is not created. Recovery Informational message only; no corrective action is needed.
3 { WARNING | ERROR } 3 - operation Error error-number (description) on filename operation identifies the uncompleted operation. error-number provides the Guardian file-system error number of the error that occurred. description describes the error that occurred. filename identifies the affected file. 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 error-number and description.
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.
16 WARNING 16 - Dirty Catalog using Fast Create; REPAIR ALL will be performed Cause In a Subsystem Control Facility (SCF) DIAGNOSE FILESET command, the REPAIR option was specified for a catalog that used the fast-create fileset option and had not been unmounted cleanly. Effect The FSCK utility proceeds as if the REPAIR ALL option were selected, because catalogs that use the fast-create option can be repaired only by using the REPAIR ALL option.
• 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. 100 SERIOUS 100 - Corrupt PXLINK Record - Parent:parent, Child:child, Name:linkname parent indicates the inode number of the parent end of the link.
102 SERIOUS 102 - Missing Link - 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 The indicated PXLINK record was missing in the old PXLINK file. 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. 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.
Recovery If the REPAIR option was not specified or the REPAIR NONE, REPAIR SERIOUS, or REPAIR OPEN 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 ALL option was specified in the command, this is an informational message only; no corrective action is needed.
If the REPAIR ALL option was specified in the command, this is an informational message only; no corrective action is needed. 209 { MINOR | SERIOUS } 209 - Invalid Inode Number, Inode=inode-number inode-number indicates the affected inode. For the MINOR message: Cause A record with the indicated invalid inode number as a key was found in the current PXINODE file. The inode number falls within the range reserved for future special inodes.
inode-number indicates the affected inode. Cause This inode appears in the PXINODE file but has no links in the PXLINK file. Effect The inode is added to the OSS /lost+found directory. 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.
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. 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.
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.
Recovery Informational message only; no corrective action is needed. 221 WARNING 221 - File Omitted from New Catalog, Inode = inode-number inode-number indicates the affected inode. Cause During a downgrade operation, the indicated inode cannot be included in the new catalog; the new catalog version does not support files of this inode’s type. The FSCK utility issues this message for each symbolic link found in the catalog if you are downgrading the catalog to a D30 version.
301 MINOR 301 - Invalid file in ZYQ Subvolume - filename. filename indicates the affected file. Cause The indicated file exists in a subvolume associated with this catalog, but it is not a valid OSS file. Effect A currently unused inode is allocated, the invalid file is renamed to correspond to this inode, and the inode is placed in the OSS /lost+found directory.
Recovery Informational message only; no corrective action is needed. 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.
Effect The file cannot be deleted from the OSS fileset, and fileset repair or deletion cannot be completed. Recovery Use the Subsystem Control Facility (SCF) DIAGNOSE FILESET command to run FSCK in a different processor. 308 ERROR 308 - Unexpected SQLCAT Error error Purging File filename error is the Guardian file-system error returned by the underlying Guardian procedure call. For further information, see the Guardian Procedure Errors and Messages Manual. filename indicates the affected file.
ABENDED: processname CPU time: interval 3: Premature process termination with fatal errors or diagnostics Termination info: code Subsystem: TANDEM.143.release text-message 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.
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. See the Guardian Procedure Errors and Messages Manual for a description of the file-system error err and possible corrective actions. Correct the problem. 2. Remove the invalid database files by entering the following at a TACL prompt: VOLUME $SYSTEM.ZXOSSMON PURGE ZOSSFSET ZOSSFS00 ZOSSFS01 ZOSSSERV ZOSSPARM 3.
If the problem persists, check the Event Management Service (EMS) log for related messages.
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. See the Guardian Procedure Errors and Messages Manual for a description of the file-system error and possible corrective actions. Correct the problem. 2. Remove the invalid database files by entering the following at a TACL prompt: VOLUME $SYSTEM.ZXOSSMON PURGE ZOSSFSET ZOSSFS00 ZOSSFS01 ZOSSSERV ZOSSPARM 3.
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. 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.
Effect The OSS Monitor process terminated abnormally with the SPI error code value -33. The OSS environment is not available. Recovery Restart the OSS Monitor with the process name $ZPMON: OSSMON /NAME $ZPMON, TERM $ZHOME, NOWAIT/ This version of OSSMON only runs on Himalaya S Series servers Cause The version of the OSS Monitor is G09 or later and the system is an HP NonStop K-series system (release versions D4x). Effect The OSS Monitor process issues the error message on its home terminal and terminates.
2 OSS E00002 Server did not respond server-name server-name is the server name of the OSS name server affected by the error. Cause During the execution of a Subsystem Control Facility (SCF) command, the OSS Monitor did not receive a response from an OSS name server process within the time specified by the IOTIMEOUT attribute of the OSS subsystem. This might indicate any of the following conditions: • The IOTIMEOUT attribute value is too small.
Recovery Perform one or more of the following actions: • Use the SCF INFO SERVER command to check for invalid information in the configuration of the failing server. • If this system is running AF_UNIX Release 2 software, modify startup scripts to remove entries that start the OSS local server process, $ZPLS. • Verify that the processors used by the server are running. • Check the server program file security permissions.
Recovery Perform one or more of the following actions: • Try to determine and correct the cause of the I/O retries. If the failing entity is an OSS name server, do one of the following: ◦ If the OSS name server is still running, stop all filesets managed by that OSS name server. Remount those filesets. ◦ If the OSS name server is not running, stop and restart the OSS Monitor so that it executes the automatic fileset restart sequence.
2. 3. 4. Enter another SCF STOP FILESET command for the indicated fileset. Enter an SCF DIAGNOSE FILESET command for the indicated fileset.
3. Use the SCF ALTER FILESET command to change the configuration of the fileset so that it uses a valid OSS directory as the mount point, then reissue the START FILESET command. For information about the INFO FILESET, ALTER FILESET, and START FILESET commands, see “Managing Filesets” (page 140). 16 OSS E00016 Unable to access catalog volume volname volname is the name of the affected disk volume.
3. Restart the OSS Monitor. 18 OSS E00018 Configuration contains invalid data Cause A fileset cannot be started because the configuration record for it has become corrupted. 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.
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 A Subsystem Control Facility (SCF) START FILESET command failed. Effect The command cannot be processed.
25 OSS E00025 Name Server server detected invalid data. server is the server name of the affected OSS name server. Cause An OSS name server received invalid data from the OSS Monitor. The OSS Monitor might be malfunctioning. Effect The command cannot be processed. Recovery Reissue the command.
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. During successful processing of this command, the CVT utility renames the current catalog files to a temporary subvolume with a name that begins with ZZ on the original catalog disk volume; the renamed files have a Guardian file identifier that ends with 9999.
Cause A Subsystem Control Facility (SCF) START FILESET command was issued, but the fileset containing the mount point for the specified fileset is not yet started. Effect The command is not processed. Recovery Start the fileset that contains the mount point. Reissue the original START FILESET command. 30 OSS E00030 Failed to stop server servername servername is the server name of the affected OSS server. Cause The indicated OSS server could not be stopped.
Recovery Reissue the command with a valid option value. See “DIAGNOSE FILESET Command” (page 293). 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.
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. 36 OSS E00036 Alter mount point of Root fileset is not allowed Cause A Subsystem Control Facility (SCF) ALTER FILESET command was issued for the root fileset, and the command contained a MNTPOINT value other than /.
Recovery Reissue the command with the required attribute specification. 44 OSS E00044 Reserved name for OSS transport agent server Cause An ADD SERVER command was attempted to create a server configuration record with the OSS transport-agent server reserved name. Effect The OSS Monitor stops processing the command. Recovery Reissue the command with a different server name.
Cause The mount point pathname specified in a Subsystem Control Facility (SCF) ADD FILESET or ALTER FILESET command begins with /E or /G. Effect The command is not processed. Recovery Select a mount point in the local node’s OSS file system and reissue the command. 49 OSS E00049 OSSMON/Database Record version incompatibility Cause A record in the OSS Monitor database has been updated by a newer version of the OSS monitor than is currently running.
Cause A Subsystem Control Facility (SCF) CONTROL FILESET command was entered, but the fileset targeted by it is corrupted. Effect The command is not processed. Recovery Use the SCF STOP FILESET command to stop the fileset. Recover the fileset using the SCF DIAGNOSE FILESET command with the REPAIR option. Restart the fileset with the START FILESET command, then reissue the CONTROL FILESET command if necessary.
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 The command is not processed.
Cause The specified MAXINODES value is less than 110 percent of the number of in-use 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 This is an informational message only. Recovery No immediate action is required. The disk capacity of the fileset should be checked to determine if its MAXINODES value should be increased further.
65 OSS E00065 RESTRICTEDACCESS attribute set or altered only by a member of both SPA and SUPER groups (but not super ID or SOA group) Cause The user that attempted to alter the RESTRICTEDACCESS attribute of a fileset did not have the appropriate privilege to execute the command. Effect The RESTRICTEDACCESS attribute is not changed.
Cause The user that attempted to set or alter the SEEPPROTECTED attribute of a fileset did not have the appropriate privilege to execute the command. Effect The SEEPPROTECTED attribute value is not changed. Recovery Log on to the local system as a user ID that is a member of both the Safeguard SECURITY-OSS-ADMINISTRATOR (SOA) security group and the Super user (255,n) group and reissue the command.
74 Attempt to set SEEPENABLED to ON with no program filename configured Cause An attempt was made to set the SEEPENABLED attribute value to ON without first configuring the value of the SEEPPROGFILENAME attribute. Effect The command is not processed. Recovery Configure the SEEPPROGFILENAME attribute and then reissue the command.
{$process-name | EDIT-filename} {$process-name} [options] [-backupcpu ] [-access < ALL | OWNER >] [-secure ] [-[no]server] [-[no]wrap] [-[no]prefixpid] [-coll ] [-idletimeout ] [-[no]quiet] [-help] { A | G | O } -- local security { N | C | U } -- remote security { - } -- "super" security Cause The user entered the OSSTTY command with the -help flag specified.
value indicates the value that caused the diagnosed condition. stdxx_redirection_target indicates the redirection target specification (IN, OUT, or TERM) that contains an invalid object type. Cause The Guardian file-system object specified in the command is not supported for use as the indicated redirection target file. For example, this message occurs when an EDIT file is specified for the OSS standard input file (IN). Effect OSSTTY is not running.
error_number indicates the Guardian file-system error returned by the underlying call to the Guardian FILE_CREATE_ procedure. For more information about this error, refer to the Guardian Procedure Errors and Messages Manual. Cause OSSTTY could not create the specified EDIT file to receive redirected OSS standard output file data. Effect OSSTTY stops.
Effect OSSTTY is not running. Recovery See “Starting OSSTTY” (page 395) for a description of the valid values and flags for the OSSTTY command. Reenter the command with a corrected flag or value. 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.
B Manually Setting Up an OSS Environment “Creating a Basic OSS Environment Without Using the OSSSETUP Utility” summarizes the procedures for configuring and starting a new OSS environment by using manually entered operator commands instead of the OSSSETUP utility. “Completing a Preconfigured Basic OSS Environment” (page 390) summarizes the procedures for completing a new OSS environment preconfigured by HP or created by the OSSSETUP DEFAULTS command.
Subtask Example Reference software that supports Security Event Exit Process Verify that the system has TCP/IP configured, has Telserv running, and that you have a VT100 or Xterm emulator or terminal Start the security manager process TCP/IPv6 Configuration and Management Manual or TCP/IP Configuration and Management Manual, and Telserv Manual >OSMP / NAME $ZSMP, PRI 190, “Starting the OSS Monitor” (page 51) NOWAIT, CPU 0 / 1 Provide Required Configuration Files Subtask Example Reference Create or m
Complete the OSS Configuration Subtask Example Reference 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 /bin/cp profile.sample profile File” (page 236) /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.
Subtask Example Reference man, and whatis commands Start OSS Servers Subtask Example Reference Start the OSS message-queue server SCF START SERVER $ZPMON.#ZMSGQ “Starting the OSS Message-Queue Server” (page 128) Start the OSS sockets START SERVER $ZPMON.#ZPLS, CPU 1, BACKUPCPU “Starting the OSS Sockets local server $ZPLS (This Local Server” (page 128) 0 EXIT task is for systems running AF_UNIX Release 1 software only.
Completing a Preconfigured Basic OSS Environment Prepare for Configuration Subtask Example Reference Log on as the super ID >LOGON SUPER.SUPER Guardian User’s Guide Ensure that preconfigured databases are accessible after a node number change HP strongly discourages changes in node (system) numbers. However, in a few situations, such as during the first startup of a newly delivered system, you must change node numbers.
Subtask Example Reference Configure utilities for periodic tasks /bin/cd /usr/bin/cron /bin/cp at.deny.sample at.deny /bin/chmod 1775 * /bin/cd /var/adm/cron /bin/cp queuedefs.sample queuedefs /bin/cp .proto.sample .proto /bin/cp cron.deny.sample cron.
Subtask Example Reference Add the user home directories to the OSS file system OSH /bin/mkdir /home/user1 /bin/chmod 0744 /home/user1 “Creating a Unique Fileset” (page 140) Configure the default printer alias by specifying a Guardian spooler process /bin/cd /bin/cp /bin/vi default “Specifying a Default Printer” (page 247) /etc printcap.sample printcap printcap /NODE1.$S.
C Falling Back to a Previous Release Version Update or Upgrading to a New Release Version Update NOTE: HP recommends that you use NonStop Software Essentials or DSM/SCM to manage your Open System Services files. The default for the Manage OSS Files option in NonStop Software Essentials and in the DSM/SCM planner interface is selected for J-series and H-series Release Version Updates (RVUs).
OSS Monitor Migration and Fallback Considerations Consider the following key points about the OSS Monitor database to eliminate most migration and fallback issues in the OSS environment: • Records in the database are associated with a version of the OSS Monitor. • After a migration to a newer RVU, records in the database are upgraded and new records are added when a new version of the OSS Monitor first opens the database.
D 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.
appears as $process_name.#stdin to a process using the Guardian file system or /G/process_name/#stdin to a process using the OSS file system. If you specify this option, the process cannot be a collector process or a spooler. If you omit this option, the name used is the device name for the Telserv terminal from which the command is entered.
-help requests a summary of OSSTTY command options. This is the first message described in “Startup Messages” (page 381). -idletimeout seconds specifies the time in seconds that OSSTTY waits before terminating after its last opener closes it. The time waiting allows another OSS process to open it if necessary for additional processing. Valid values are in the range 0 through 32767. A value of 0 indicates that OSSTTY should terminate immediately after its last opener closes it.
-secure security_string specifies the access permissions for the OUT file in the Guardian file system when the OUT file is an EDIT file and must be created by the command. Valid values are any four-character combination of the read, write, execute, and purge permissions allowed for Guardian file-system objects (N, U, O, A, G, C, or - for the super ID). If you omit this option when an EDIT file must be created, OSSTTY uses the default file-system permissions for the user ID of your terminal session.
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.
default_response is the action taken or the value used if the user presses Return. After you start a utility, you cannot back up within its dialog and choose alternate responses to prompts. Only the OSSSETUP utility provides help as a TACL command line option. 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.
Utility PARAMs Beginning with the T0585AAB version, the EasySetup utilities support the following TACL PARAM declarations: PARAM Name PARAM value Used by EASYSETUP^EMSVOL A valid disk volume name that identifies the disk on which the Event Management Service (EMS) subvolume (ZSPIDEF) resides. This subvolume must contain the event definition files ZOSSTACL and ZEMSTACL.
Quotation marks are optional if only one value is specified for a PARAM. If you specify a value that is not recognized as a keyword, a warning message appears and the value is ignored. To have the intended effect, PARAMs must be declared before a utility is started.
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. • OSSSETUP has previously been run, but OSSREMOV has not been run. • You have manually configured your OSS environment. Running OSSSETUP For information about running OSSSETUP, see the Open System Services Installation Guide.
Example 34 Example of Servers, Subsystem Processes, and Other Information Display Servers, Subsystem Processes and other Information Process Subsystem/Server Processors (-1 = N/A) $ZPMON $ZSMP $NULL $ZLSxx $ZMSGQ $ZPNS $ZPNH $ZTAxx *OSS Monitor *Security Manager *NULL Monitor *Local Servers Message Queue Server ROOT's Name Server HOME's Name Server *Transport Agents System load processor 0 1 0 1 One configured for each processor 2 3 0 1 -1 -1 One configured for each processor Processes with asterisks
for its own startup. However, not configuring or starting $NULL or $ZSMP could affect the functioning of other products or subsystems. • If $NULL or $ZSMP is already running and not configured as a generic object, and you have used the EASYSETUP^STARTUP param to specify $NULL or $ZSMP, you must stop it before you can use OSSSETUP to configure it as a generic object. To stop $NULL, use the TACL STOP command. To stop $ZSMP, use the SAFECOM STOP command.
The STARTOSS utility can be used when any of the following is true: • Your OSS environment was preconfigured by HP or configured by using OSSSETUP and has not been modified. • Your OSS environment was preconfigured by HP or configured by using OSSSETUP, and all subsequent modifications have been entered into the OSSINFIL file in the STARTOSS volume and subvolume.
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, stopping $NULL or $ZSMP could affect the functioning of other products or subsystems.
• The OSSREMOV utility does not back up any files from the OSS file system before removing filesets. • The OSSREMOV command removes all OSS files and all fileset catalogs. As a result, it also removes any directories or files kept in the OSS file system by products such as NonStop SQL/MX or iTP WebServer. • The OSSREMOV command prompts you to determine whether it should remove the security manager server process $ZSMP and the $NULL process.
E Summary of OSS Processes This appendix summarizes the system processes related to the Open System Services environment. Table 19 Summary of OSS Process Information Process Name File Name in $SYSTEM.SYSnn Description Started By Number of Processes $ZLSnn OSSLS2 local server 2 Processor load 1 per processor Transport-provider process for OSS AF_UNIX Release 2 sockets (AF_INET, AF_INET6, and AF_UNIX Release 1 transport-provider sockets use different names).
Table 19 Summary of OSS Process Information (continued) Process Name File Name in $SYSTEM.SYSnn Description Started By Number of Processes $ZPPnn OSSPS Pipe server Processor load 1 per processor Supports the transfer of data between OSS processes that use pipes or FIFOs between processors. If this process terminates abnormally it halts the processor in which it runs. For more information, see “Pipes and FIFOs” (page 39).
F 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.
Table 20 OSS Environment Limits (continued) Description Maximum Value Comment OSS processes per node • Approximately 160,000 processes (active and zombie), beginning with the J06.14 and H06.25 RVUs Beginning with the J06.14 and H06.25 RVUs, EMS events occur for: limit warning, limit error, and status normal.1 • Approximately 63,000 for J-series RVUs prior to J06.14 and H06.11–H06.24 RVUs • Approximately 29,000 for H-series RVUs prior to H06.11 and G06.
Table 20 OSS Environment Limits (continued) Description Maximum Value OSS file opens per processor2 • 128,000 beginning with • Beginning with the the J06.14 and H06.25 J06.07 and H06.18 RVUs RVUs, EMS events occur for: allocation error, limit • 64,000 for warning, limit error, and J06.07–J06.13 and status normal.3 H06.18–H06.24 RVUs • Approximately 12,000 for J06.06 and earlier J-series RVUs, H06.17 and earlier H-series RVUs, and G06.
Table 20 OSS Environment Limits (continued) Description Maximum Value • 16,000 for J06.07–J06.13 and H06.18–H06.24 RVUs OSS pipes/FIFOs per processor2 Comment for: allocation error, limit File System Resources” warning, limit error, and (page 96) for information status normal.1 about viewing usages. • 4,095 for J06.06 and earlier J-series RVUs, H06.17 and earlier H-series RVUs, and G06.24 and later G-series RVUs • Bounded resource3 • 32,000 beginning with the J06.14 and H06.
Table 20 OSS Environment Limits (continued) Description Maximum Value H06.18–H06.24 RVUs OSS NFS opens per processor2 • 32,000 beginning with the J06.14 and H06.25 RVUs • 16,000 for J06.07–J06.13 and H06.18–H06.24 RVUs Comment To Check Amount Currently Used • Constrained by the ALL OPENS limit or PXS and PXS64 memory • Maximums are for the number of unique disk files that can be accessed via NFS operations per processor. • Unbounded resource4 Beginning with the J06.14 and H06.
Table 20 OSS Environment Limits (continued) Description Maximum Value size per processor J06.07 and H06.18 RVUs Comment J06.07 and H06.18 RVUs, EMS events occur for: allocation error, limit warning, limit error, and status normal.1 To Check Amount Currently Used and H06.25 RVUs, see “Using SCF to View OSS File System Resources” (page 96) for information about viewing usages. • AF_INET sockets utilize AF_INET socket data buffers only when the process is configured on a Legacy TCP/IP address.
Table 20 OSS Environment Limits (continued) To Check Amount Currently Used Description Maximum Value Comment OSS filesets per node Functional limit: 1,000,000 Theoretical maximum: 33,553,952 Functional limit is imposed At a TACL prompt, use the by functional limit on SCF INFO FILESET *, directories (a fileset must DETAIL command. have a directory as a mount point). Theoretical maximum is determined by maximum number of device labels.
1 2 3 4 5 For details of messages for these EMS events, see the OSS Messages chapter in the Operator Messages Manual. The total number of file opens per processor includes opens of all types: disk opens, terminal opens, pipe/FIFO opens, AF_INET socket opens, AF_UNIX socket opens, NFS opens, directory opens, and directory stream opens. Each of these types of opens is limited both by the limit for the type of open and by the total number of file opens per processor.
Table 22 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 Guardian APIs with 64-Bit elections 64-Bit elections OSS small file (< 2GB-8KB approx. 2 GB) created (2,147,475,456 on J06.03 and later bytes) J-series RVUs, H06.06 and later H-series RVUs, or G06.
Table 22 File Format and Limits Table for File Open Behavior (continued) Using This Function To Open This File... 32-Bit OSS APIs such as open() 64-Bit OSS APIs such as open64() using OSS APIs on bytes), Enscribe J06.03 and later Format 2 J-series RVUs, H06.06 and later H-series RVUs, or G06.29 and later G-series RVUs (and file is less than approx.
Glossary A7CINFO file A configuration file, used in the In the NonStop software management environment, that is distributed with OSS products and 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. access control list (ACL) A structure attached to a software object that defines access permissions for multiple users and groups.
See also Coordinated Universal Time (UTC). Event Management System The product on NonStop systems that manages the events created by other products executing on NonStop systems. FIFO A type of file that is always read and written in a first-in, first-out manner. file An object to which data can be written or from which data can be read. A file has attributes such as access permissions and a file type.
limit error A resource error that occurs when a resource’s current usage is at (or so near) its limit that an allocation cannot be attempted. A limit error generates an EMS event. A resource allocation cannot be attempted when the resource’s current usage plus the requested resource amount is greater than the resource’s defined limit. link A directory entry for a file. Measure A system measurement tool. See the Measure Reference Manual for details.
Open System Services (OSS) Socket Library The OSS Layer 3 code for OSS sockets that runs as system library. Open System Services (OSS) Transport Agent A per-processor OSS helper process that acts as a messaging agent between socket requests made on a processor and the corresponding transport provider. orphan file A file with no corresponding inode in the PXINODE file. orphan inode An inode that appears in the PXINODE file but has no links in the PXLINK file. OSS See Open System Services (OSS).
save file A file created through the Inspect or Debug product. A save file contains enough information about a running process at a given time to restart the process at the same point in its execution. A save file contains an image of the process, data for the process, and the status of the process at the time the save file was created. A save file can be created through an Inspect SAVE command at any time.
UID A nonnegative integer that uniquely identifies a user within a node. The UID is a scalar number consisting of the group number multiplied by 256 and added to the member number; for example, the UID of the super ID is (255*256) + 255, or 65535. The UID is used in the OSS environment for functions normally associated with a UNIX user ID. See also user ID. Unbounded open resource An OSS file-system resource that is used to control the number of OSS opens or operations on OSS opens.
Index Symbols #### security permissions, 85 #ZLSnn, 113 #ZMSGQ, 112 #ZPLS, 113, 114 #ZPMON, 49 #ZPNH, 103, 108 #ZPNS, 101, 102, 103, 107, 108 #ZRAnn, 116 #ZTAnn, 116 $0, 396 $DSMSCM, 111 $NULL, 402, 406, 407 $OSS disk volume, 102, 111 $SYSTEM, 111 $SYSTEM.STARTUP.STARTUP, 406 $SYSTEM.SYSnn.ZX0devicelabel, 287 $SYSTEM.SYSTEM OSS Monitor subvolume, 251 $SYSTEM.SYSTEM.UNISTDH, 239 $SYSTEM.ZRPC.RPC, 119 $SYSTEM.ZTCPIP.HOSTS, 117, 118, 119 $SYSTEM.ZTCPIP.IPNODES, 117 $SYSTEM.ZTCPIP.NETWORKS, 117 $SYSTEM.ZTCPIP.
/home/quotas file (UNIX), 218 /tmp directory, 101, 143 /tmp/oss.tree.ddmmmyyyy.system_name file, 399 /usr/etc/exportfs utility, 220 /usr/etc/showmount utility, 220 /usr/include directory, 179 /usr/include/unistd.h, 239 /usr/lib/cron/at.allow file, 75 /usr/lib/cron/at.deny file, 73, 75 /usr/ucb directory, 179 /var/adm/cron.deny file, 219 /var/adm/cron/.proto file, 74, 219 /var/adm/cron/cron.allow file, 74 /var/adm/cron/cron.
basename.private file, 120 batch command (OSS), 75 Berkeley Internet Name Domain (BIND) 9, 100 bin user name (UNIX), 220 BIND 9.2.3, 100 BIND 9.
Device subtype required for /E use, 83 DEVICELABEL attribute OSS fileset, 258 df command, 220 dfs/dfstab file (UNIX), 217 DIAGNOSE FILESET command (SCF), 159, 293 diff command (OSS), 84 change detection, 231 dircmp command (OSS), 195 Directories comparing, 195 controlling the growth of, 241 Directory graph, 163 Directory names, characters, number of, 80 Dirty Catalog using Fast Create; REPAIR ALL will be performed (FSCK message), 343 Disk Compression Program (DCOM), 242 Disk files, Guardian, referencing, 86
Failed to stop fileset filesetname (OSS Monitor message), 364 Failed to stop server servername (OSS Monitor message), 371 Fast-create fileset option, 101, 142 fd file (UNIX), 217 File code 100, 85 File code 444, 85 File compression using compress, 242 File compression using pack, 242 File hierarchy, copying, 187 File Omitted from New Catalog, Inode = inode-number (FSCK message), 352 File transfer protocol (FTP) see FTP File-sharing groups, 205, 223 File-system cache, 35 FILE_OPEN_ Error error on file filena
FTPUSERS file, 204 indirect access to OSS, 205 initial access through Guardian, 213 user access through, 199 ftp user name (UNIX), 220 ftpd process (UNIX), 220 FTPUSERS file, 204 ftpusers file (UNIX), 217 Full disk volume, managing, 158 FUP INFO display, OSS file, 176 G games user name (UNIX), 220 gated process (UNIX), 219 GB18030 locale support, 239 gencat command (OSS), 239 genxlt command (OSS), 239 get command options, 243 getfilepriv command, 229 getopts command (OSS), 243 getstats utility (UNIX), 220
Invalid inode number (inconsistency checked by FSCK), 164 Invalid Inode Number, Inode=inode-number (FSCK message), 348 Invalid mount point in fileset filesetname (OSS Monitor message), 365 Invalid or Corrupt PXCKSTAT File (FSCK message), 353 Invalid parameter parameter_name (OSS Monitor message), 372 Invalid Parent List, Inode= inode-number (FSCK message), 345 Invalid Serial Number*** - token (CVT message), 339 Invalid Subvolume Name*** - token (CVT message), 339 Invalid Timestamp, Inode=inode-number (FSCK
OSS server, 267, 282 md5 utility (UNIX), 219 MEDIACOM utility, 184 mem file (UNIX), 217 MEMPAGES attribute OSS application, 64 merge_whatis command, 181 merge_whatis command (OSS), 182 Message queues, 39 MINOR messages (FSCK) 205 - Missing ZYQ File, Inode=inode-number, 346 207 - OSS Name Server Failed while Fileset was Mounted, 347 208 - There are nnn Inode Numbers Unaccounted For, 347 209 - Invalid Inode Number, Inode=inode-number, 348 213 - Orphan ZYQ File - filename, 349 214 - Catalog/File Label Mismatch
nsupdate utility, 100, 119 ntpd process (UNIX), 220 null file (OSS), 218 NULL.
OSS E00055 The MAXINODES value is lower than the number of currently inuse inodes inuseinodes for the fileset fileset. (OSS Monitor message), 376 OSS E00056 The primary Name Server server failed to migrate to the processor specified by the CPU attribute. (OSS Monitor message), 376 OSS E00057 The backup Name Server server failed to migrate to the processor specified by the BACKUPCPU attribute.
D46 version, 167 database file security, 227 default subvolume, 251 description of, 251 device subtype, 255 device type, 255 entering SCF commands, 55 error handling, 55 fault tolerance, 255 G09 or newer version, 167 licensed, 213 managing OSS environment, 251 messages, 361 obtaining information about, 57, 130 online help, 252 PARAMs used, 53 process name $ZPMON, 51, 54, 55 SCF commands see SCF specifying a home terminal for, 54 starting as a normal process, 51 starting as a persistent process, 52 starting
STATUS SERVER considerations, 329 stopping, 134, 233 ZOSSSERV file attributes, 109 OSS SEEP process attribute can be set or altered only by a member of both SOA and Super group (OSS Monitor message), 380 OSS Server Process (OSP), 35, 98 OSS servers adding, 120 configuring, 121 reconfiguring, 135 removing, 138 starting, 127 stopping, 133 OSS shell bash, 235 commands, recursive, 84 script, batch execution of, 77 OSS Sockets Local Server ZOSSSERV file, 113 OSS sockets local server default process for AF_UNIX R
and DSM/SCM, 180 and NonStop Software Essentials, 180 PINSTALL utility, 102, 177, 178, 227 pname command (OSS), 175 POOL attribute, 152 OSS fileset, 257, 275 portmap process, 99, 132 POSIX^CONFIG^LOC PARAM, 54 preserve file (OSS), 219 Primary group, 223 Primary user groups, 205 PRIMARYCPU attribute OSS application, 64 OSS Monitor, 52 print_gb18030 utility (OSS), 239 printcap file, 247 printcap file (OSS), 219, 247 printcap.
functions, 99 using with the portmap process, 132 rsh command (OSS), 220, 243 rshd process, 132 configuring, 124 functions, 99 removing, 139 starting, 99 runcat command (OSS), 239 RVU falling back Manage OSS Files option, 393 upgrading Manage OSS Files option, 393 S S_NONSTOP file open flag, 157 Safe threshold, OSS resource monitoring, 61 SAFECOM commands ADD ALIAS, 205, 213 ADD GROUP, 223 ADD USER, 213, 223 ALTER ALIAS, 207, 209, 213 ALTER USER, 207, 209, 213 Safeguard, 33 using for file audit reduction,
sendmail/aliases file (UNIX), 217 SERIOUS messages (FSCK) 100 - Corrupt PXLINK Record, 344 101 - Duplicate Link ID , 344 102 - Missing Link, 345 200 - Invalid Parent List, Inode= inode-number, 345 201 - Broken Free List, Inode=inode-number , 345 202 - Corrupt Inode, Inode=inode-number, 346 204 - Too Many Parents, Inode=inode-number, 346 206 - Parent Inode Not Directory, Inode=inode-number, 347 209 - Invalid Inode Number, Inode=inode-number, 348 210 - Missing Inode, Inode=inode-number, 349 211 - Loop in Dire
TCP/IP symbolic links, 388, 390, 401 TCP6MON, 43 TCPIP^PROCESS^NAME, 124, 126 TCPIP^RESOLVER^NAME, 118 TCPIP_RESOLVER_NAME environment variable, 118 TCPMON, 43 telnet command (UNIX), 220 telnet user name (UNIX), 220 Telserv configuring access, 199 configuring user login, 199, 204 direct access to OSS, 199, 200, 204 indirect access to OSS, 199, 204 logging in through an OSS program, 201 logging in through OSS shell, 200 OSSTTY break key processing, 398 redirection of OSS standard files from, 395 TEMPPOOL, 10
visitor user name (UNIX), 220 Volume mode backup, 192 vproc command (OSS), 249 VPROC utility, 249 W wall command (OSS), 220 WARNING - Event definitions file (filename) not loaded because file could not be found (OSS EasySetup message), 336 WARNING - Variable (variable_name) needed but does not exist (OSS EasySetup message), 336 WARNING messages (FSCK) 13 - Can’t UPGRADE/DOWNGRADE catalog with CORRUPT/MISSING Super Block, 342 14 - Catalog Already Upgraded, 342 15 - Catalog Already Downgraded, 342 16 - Dirty
