cancel
Showing results for 
Search instead for 
Did you mean: 

Number of concurrent restores per client 7.5.0.4

nairan
Level 2

Hi, Any parameters available in netbackup to control the number of concurrent restores per client.

it seems it is 1 by default.

Thanks,

AN

1 ACCEPTED SOLUTION

Accepted Solutions

sclind
Moderator
Moderator
   VIP   

What do you have for "maximum jobs per client" for the master or media  server?  If '1' that might be the problem.

View solution in original post

9 REPLIES 9

sclind
Moderator
Moderator
   VIP   

What do you have for "maximum jobs per client" for the master or media  server?  If '1' that might be the problem.

Marianne
Moderator
Moderator
Partner    VIP    Accredited Certified
What is Max Jobs per Client set to on the Master?
Not sure if this setting applies to backups and restores.

There may be other reasons.
Are restore jobs queued?
If so, what is the reason listed in job details?

Sorry, I am not sure how it became an "Accepted solution"

Max Jobs / client is kept at 25.

Thanks you.

it is kept at 25.

Both restore job were initiated almost some seconds difference.

Both jobs were just hanging saying "begin reading" for 10 min or so. I just killed one of them to complete the second job.

The restores were from a DataDomain.

 

sclind
Moderator
Moderator
   VIP   

I thought that you meant the second job was queued for execution.  Both jobs actually starting is another situation.

Marianne
Moderator
Moderator
Partner    VIP    Accredited Certified

Since both jobs were active and already at "begin reading" stage, it means that the restore process was handed off to bptm on the media server.

Activity Monitor will show the bptm PID next to each "begin reading" entry.

You will need to look at bptm log on the media server to see what each bptm process is doing.
Level 0 log will probably not have sufficient info. 
I would personally try level 3 log (Veritas Support will ask for level 5).

One more thing:

I see the following in the Data Domain Best Practice White Paper:

...  “limit I/O streams per volume,” is set within a given disk pool....

This setting on the DD will limit max amount of read and write streams.
If this max limit has been reached, then maybe bpdm log on the media server will display appropriate message/entry received from DD, but this is a wild guess.

I think that DD logs may be more appropriate if this is the case.

 

Marianne, Thanks !

Just verified, we have not enabled that option at all.

When i started the restore nothing much was going on the master and on the media servers. Hardly 20 jobs in progress. Our log setting was 0 so nothing much there. Will increse it next time and check.

Can we have multiple restore jobs running for different clients in parallel which were backed up on to same tape storage?

sclind
Moderator
Moderator
   VIP   

Yes.

They must be submitted at roughly the same time though.