Stopping resheduler on failure of bp_start 1

[AnotherAlan]

Hi All,

I'm a bit green when it comes to Netbackup, have had a dig around but can't find an answer.
I have a bpstart script with timeout 7200. However, this calls an RMAN backup which is taking close to two hours.
Without changing the timeout I wanted to know if I can stop this backup, and only this backup, from resubmitting if it should fail originally.
Unfortunately at the moment the second parse is rewriting the RMAN backup.
I have asked the DBA's to remove the RMAN portion from the bp_start script and run it from cron. However, in the process of being a good (as in nice - not as in great) Sys Admin I wanted to now if there was anything I can do at the netbackup level.

Thanks all, sorry it's a long one.
Alan

 

[MaineDCAdmin]

If this is the only client in the Policy why not have bpstart notify make the policy inactive . I do not think bpend_notify works with DB Agents, so that option is out. Perhaps you could use the info from bpstart_notify and bpplinfo to set the Policy to inactive. Once bpstart finishes, the backup commences, so you will need to find a means to terminate the current backup. Hope this gets you started at the very least. The NetBackup CLI is quite powerful, and I am just learning to use it with Korn Shell

 

[AnotherAlan]

Hi MaineDCAdmin, and thanks for the reply.
That is a good idea and something I will look at.
What I was considering is adding a flag to the bp_start script to check if a succesful rman has already been completed.

Another question if I may: At the moment NB makes two passes or attempts at each backup. Is there a way I can restrict it to just one. i.e If it fails first time then do not attempt to requeue?

With regards the CLI, it does have some great commands, trouble I find is that the syntax of them can be rather long winded.
Anyway, a star for your trouble.

Thanks again,
Alan

 

[MaineDCAdmin]

The number of retries is set as a parameter of the Master Server, so you are stuck with what is set in the Master. If you are NOT using a DB Agent, you could have bpend_notify, if the backup is successful clear the file you "touched"; the next time the backup ran bpstart_notify could check for the file and abort if it is found. We did something like this so we do not get paged on the first backup failure, an at noon every day the touch files are deleted. There are probably more elegeant ways to test the file creation date of the touch file, but it works....

 

[AnotherAlan]

Thanks again Maine,

I managed to get my DBA's to run the rman part from cron instead. But your ideas are good ones and I am sure they will be used some time soon.

Regards