BackPak H4.00 Troubleshooting and Messages Manual

Troubleshooting
15 BackPak H4.00 Troubleshooting & Messages
Common and Special Cases
Tape subsystem reset procedure
This procedure is applicable in extreme cases, when the tape subsystem
infrastructure (ZSERVER and DSM/TC) seems impacted by severe operational
difficulties, when $ZSVR volume recognition does not execute correctly and DSM/TC
does not reply properly to commands.
This procedure is not a replacement to the HP documentation for managing HP
NonStop components. This is just a help to reset the whole NonStop tape system
including the BackPak components.
The “Reset procedure” below resets the tape subsystem of a single NonStop node,
but some components can be shared with remote nodes not showing the same level
of operational problem.
Some of the tape system components on the other nodes might have to be reset just
as side effect of this procedure. See “Procedure complements for other nodes
right after the “Reset procedure”.
Reset procedure
1. Log in with SUPER.SUPER
2. Stop the BackPak EMS Extractor(s)
ex: STOP $BBEXT, or SCF ABORT PROCESS $ZZKRN.#BBEXT
3. Exit any MEDIACOM session, issue a Session logout in all BackPak GUI instances.
4. Stop any tape application, except TMF audit dumps
5. Wait any TMF audit dump ends or aborts
6. Stop the Guardian ZSERVER process
STOP $ZSVR
7. Stop processes related to DSM/TC
STATUS *, PROG $SYSTEM.SYS*.MEDIA*
8. If the previous command shows only MEDIACOM, MEDIASRV and MEDIADBM
programs,
stop them by: STATUS *, PROG $SYSTEM.SYS*.MEDIA*, STOP
else
stop individually all occurrences of MEDIACOM, MEDIASRV and MEDIADBM
9. Stop all tape devices
10. SCF (to get the SCF prompt)
RESET TAPE $*, FORCE (reply Y to confirmation questions)
11. Wait for all tape IOP are stopped, this can easily take more than 20 minutes
SCF STATUS TAPE $*