TDP schedules failing after instance name change

[mmbdcis]

Hi,


I have a problem with a certain TDP backup for an Oracle11 DB running on a Unix-AIX server.
The DB and TDP worked fine since they were first configured, until the Oracle instance name had to be changed.

After the instance name has been changed, the scheduled backups won't run anymore.
I have changed the instance name in the TDP/RMAN scripts, as well as in the configuration files. The schedule on the node has been restarted.
The dsmerror file only states that the schedule had failed with an error code of 1, the dsmsched file gives more or less the same output.

The TDP/RMAN scripts are known to work OK, since they manually complete successfully.

My best guess is that I might have forgotten to update the instance name in any of the configuration files, so I'd like to verify which files exactly are to be changed in the scenario.
Also, if you have any other ideas that could help...

Thanks

 

[nopalh]

pls check schedule scripts

 

[mmbdcis]

pls check schedule scripts

The scripts are OK - the instance name has been changed, and they work fine when run manually.

This is as far as I know for now:
The TSM scheduler is running, and it does start to run the script configured under "description". The dsmsched.log records that the scheduled task (the script) has started to run, and then no other output is added, also the next scheduled backups on this node fail to start - most ptobably meaning that the script is "stuck", making the scheduler think that it is still running.

I have tried to run the script manually with the same users/priviliges the TDP scheduler uses, and it works just fine. I'm not sure what other differences between running the script manually and running it with the TDP scheduler could affect its way of work.

 

[tsmdumber]

did you check the permission for the script file?

usually rman is run with the oracle id which has sufficient previlages to the databases.

may be you should add a line to" su to the oracle id " to the script and try t run it again!

 

[mmbdcis]

did you check the permission for the script file?

usually rman is run with the oracle id which has sufficient previlages to the databases.

may be you should add a line to" su to the oracle id " to the script and try t run it again!

Yes, I have checked the permission as well.

The RMAN scripts indeed usually run with the oracle user. I have tried running the scripts manually with the oracle user, as well as running them with a "su" command from the root user (which is how the TDP scheduler runs them). It worked in both cases manually, but still nothing when running the scheduler.

Is it possible that the TDP installation gets some properties from the DB when being installed, and therefore has issues with instances that have been changed/installed after the TDP? My best guess for now is to reinstall the TDP client, but I'd rather find a solution than this workaround.

 

[mmbdcis]

Hi,


Sorry to bump this old thread, but the problem hasn't been resolved yet, and I now need once again to deal with this server.
Last thing I tried was making the scheduler run a very simple script (only an "echo"), at which it has also failed with a return code 1, meaning the problem is surely not in the scripts but rather in the TSM/TDP configuration somewhere.

Are there any other tests I could run?

 

[opermty]

try reset daemon dsmcad

regards

 

[vaibhao.wagh]

Hi mmbdcis,

Could you please copy & paste error message from "tdpoerror.log" and oracle logs.

 

[opermty]

please, review "tdpoconf password", if you change instance name you have to relink the same with tdpo to make up the backup

regards