Duplication to Tape Failing

[Seth_E]

Hi all,

We have a duplication (basic disk) schedule that runs daily at 4am which copies data from a DSU to Tape. It's set in the following ways.

Multiple Copies

1 copy goes to Tape (onsite)

1 copy goes to Tape (offsite)

The company I work for will not invest in Vault. Therefore, this is the method I use.

for the most part, this work very well. Today, I ran out of scratch volumes in the offsite pool so, duplication succeeded only to onsite. My question is, is there a simple way (once i load scratch volumes) to ensure that I can get the data to the offsite pool? Will running a manual relocation work?  I'm assuming it may not since the data has already moved to onsite pool sucesfully.

Thanks all for your help!

[Nicolai · Accepted solution]

Since the local duplication succeeded this indicate to me that the inline setting "For each image if this copy fails" is set to continue and therefor you will properly need to run manual duplication to the images not copied to the offsite pool.

To ensure the data get to both pool you can set "fail all copies" instead of "continue"

Can you please verify the setting

This tech note is not really related but mention the lnline copy setting I mention.

http://www.symantec.com/docs/TECH153222

[BTLOMS]

what Nicolai said

or

what is the output of this:

nbstlutil stlilist -image_incomplete -U

[Seth_E]

I do not get an output when I run tis command.

[Seth_E]

Great technote! Now, i have to determine if I want the entire job to fail, or continue as is. The problem with failing the whole job, is that data will not copy to onsite group, thus, the disk will fill up and in the end, backups will fail as a result of not enough scratch volumes in offsite pool.

Lets assume that I want to keep the setting so that the job continues on error, is there a simple way to rerun the duplication so that the images get copied to offsite pool?

[Amol_Nair]

The "nbstlutil" command is used for SLP.. You do not see any output since you are using Basic Disk and not SLP... The easier way would be to find a list of all the backupids that you want to manually duplicate and use the bpduplicate command with the bidfile switch...

[Seth_E]

That's what I was thinking, but, hoping to find an easier way. Thanks very much everyone.,

[Nicolai]

Amol is right - no SLP in use here.

This is a difficult one 

I think you can do a bpimagelist for the job. Failed images are those that  that does not have 2 copies (field 21 in bpimagelist)  AND the primary copy is copy 2 (field 28).

Primary copy being 2 indicate duplicated images (DSSU) and only one copy indicate inline copy failed for those images.

http://www.symantec.com/docs/TECH5584

Is this clear speech ?

[Amol_Nair]

If all these duplications were done recently and the jobs are still available in the activity monitor then maybe you could just create a filter in the activity monitor with DSSU and get all the jobids of the successfully completed onsite jobs... Once you have the jobids, the ffilelogs folder in "netbackup\db\jobs" will give you the backupids to create your BID file... Another thing thay came to my mind simcw you want something easy... Set the maximum duplication copies to 3, then select all the images on tapes in your onsite volume pool and then run the duplications.. If the 3rd copy i.e. copy in offsite volume pool exists the duplication will fail with status 191. The only problem would be to cleanup the failed jobs.. ;)