NonStop S-Series Operations Guide (G06.24+)

Starting and Stopping the System
HP NonStop S-Series Operations Guide522459-007
16-17
Getting a Corrupt System Configuration File
Analyzed
When you consider performing a tape load from a SIT:
You must contact the GCSC for guidance in restoring your system disk from a SIT.
You can configure the Distributed Systems Management/Software Configuration
Manager (DSM/SCM) to create a SIT whenever a significant software update is
performed, or you can request one each time a new SYSnn is created. For more
information, see the DSM/SCM User’s Guide.
Keep a copy of the contents of your system disk on backup tape. Unlike an
alternate system disk, the system image tape that DSM/SCM can create does not
contain all the files that were on the $SYSTEM disk, including almost important
system configuration information.
Do not perform a tape load to bring up a new software release under normal
conditions. The normal procedure for bringing up a new software release is to use
the DSM/SCM to create a new configuration revision (including the system image)
and then to load the system from the new SYSnn subvolume. For more
information, see the Gnn.nn Software Installation and Upgrade Guide.
Getting a Corrupt System Configuration File Analyzed
If the current system configuration file is corrupted, you can send it to your service
provider for an analysis. Follow these steps:
1. Return to a saved, stable configuration file by following the procedure outlined in
Procedure to Use the System Startup Dialog Box on page 16-7.
2. Once the system is up and stable, copy to a backup tape the corrupt CONFSAVE
file. For example:
> BACKUP $TAPE, $SYSTEM.ZSYSCONF.CONFSAVE, LISTALL
Do this before you load the system a second time. Another system load operation
will overwrite the CONFSAVE file you want analyzed.
3. Submit the tape to your service provider for analysis, along with a copy of any SCF
command file or SCF log file of the commands that were part of the process that
created the corrupt configuration.
Starting Other System Components
Once the system is started, you must start individual devices, processes, applications,
and communications lines. Procedures for starting objects differ depending on how the
objects are configured to start. HP recommends that you bring your system up in
stages, verifying each stage, to facilitate recovery if any step fails.
Some processes are started automatically by the $ZPM persistence monitor. These
processes include the operating system, SLSA subsystem, storage subsystem, and
WAN subsystem.