Savepnpc problem, post script doesnt runs

Question

We have one Networker 7.3.2 environment.
Both server (sunebs) and client (pluto) runs Solaris 9.
The customer uses gUprd_offline group to backup three clients:
- pluto : cluster node1
- syrius : cluster node2
- sapserv : cluster moving address (backup command is savepnpc)
The cluster runs Oracle and SAP.

The gUprd_offline .res file look like:
type: savepnpc;
precmd: "/root/pre";
pstcmd: "/root/post";
timeout: "5:50am";

pre command:
#!/bin/ksh
/usr/cluster/bin/scswitch -n -j sap-ci_sapserv
/usr/cluster/bin/scswitch -n -j orasrv_sapserv
root@pluto # cat /root/post

post command:
#!/bin/ksh
/usr/cluster/bin/scswitch -e -j orasrv_sapserv
/usr/cluster/bin/scswitch -e -j sap-ci_sapserv

If we start the gUprd_offline group, the pre command runs succesfully, but the
post command fails, doesnt run.
The savepnpc.log look like:

05/13/07 22:05:58 pstclntsave: Savegroup is not running.
05/13/07 22:05:58 pstclntsave: All savesets on the worklist are done.
05/13/07 22:05:58 pstclntsave: Cannot stat the temp file created by preclntsave.
05/13/07 22:05:58 pstclntsave: Exited.

No tmp remains in the /nsr/tmp directory.

How can we solve it to run the post script?

pgerke · Answer

Not a lot of help here, but...You might check the /nsr/logs/messages file to see if it can provide any output from the post backup commands, but since the message says the server can't stat the processes, there probably isn't anything there.It sounds like you need to open a case with support.Good Luck,Paul

dk3 · Answer

On I found an hint that the behavior may result from different NetWorker versions on client and server. You've written that your environment is 7.3.2 - is this true for client and server (even patch level?).

The message "Savegroup is not running." seems to indicate that postclntsave is not able to read /nsr/tmp/ .lck that is typically created by preclntsave. Are you using any special user to perform your backups? There might be a write permission issue in this case. Are there any messages in /nsr/logs/savepnpc.log on clients and/or server?

Another question: What platform is your backup client? There's another forum thread describing similar problems with 64bit platform - without solution. Maybe contacting support would help.

ble1 · Answer

savepnpc will store data in local /nsr/tmp. If you have client installed as HA then /nsr/tmp at the time savepnpc is executed will be global one. Once pre script is executed global one is replaced by local one and data from /nsr/tmp is lost. That could be another possible problem which would explain what you see.

It is strange for me to see customers are willing to protect their db via HA solution, but for backup they still use cold backup approach instead of online module

paulo3 · Answer

Hi!

Both client and server on the same Networker platform.
(7.3.2 jumbo)

Both client and server are running on Solaris 9.s

savepnpc will store data in local /nsr/tmp.
If you have client installed as HA then /nsr/tmp at the time savepnpc is executed will be
global one. Once pre script is executed global one is replaced by local one and data from
/nsr/tmp is lost. That could be another possible problem which would explain what you >see.

It could be a right solution, but i dont understand it clearly.
Client is installed with the normal pkgadd utility.
The group contains the moving ip cluster ip address.

How can i correctly install the client?
I should ssh to the moving ip?

ble1 · Answer

For start, log to node which is active one (it is important that it runs) and see where /nsr points to. If it is link which points to cluster managed data then this could mean my theory was good.

NetWorker

Savepnpc problem, post script doesnt runs

Was this post helpful?