cancel
Turn on suggestions
Auto-suggest helps you quickly narrow down your search results by suggesting possible matches as you type.
Showing results for 
Search instead for 
Did you mean: 
Veritas.com
Support

Jobs vs Policies vs Clients

jplally
Level 3

‎03-10-2008 08:11 AM

Hi,
 
We're just getting started with Netbackup (6.5.1) coming from Legato. One big issue we're dealing with is what Netbackup calls a job. It seems that Netbackup creates a 'job' for every filesystem or mount point that exists on my Unix (AIX) servers. Some of theses servers have close to 100 file systems mounted and I would much prefer to look at the status of the overall client instead of 100 separate jobs (one for each filesystem). Any ideas? Thanks.
0 Kudos
6 REPLIES 6

Stumpr2
Level 6

‎03-10-2008 08:29 AM

It's working exactly as you instructed it to in the backup policy. Read the chapter called "Policies" in the admin guide volume 1 There are many ways to set up the policy and it is better for you to read the chapter than to get a one-time quickie answer.
 
If you have specific questions concerning the policy then use the "help" button for a quick explaination. If you have questions concering what you read then please post them here.
 
0 Kudos

jplally
Level 3

‎03-10-2008 09:42 AM

Thanks for your quick response. I'm re-reading the policy section of the admin manual but other than limiting the number of jobs per policy or not allowing multiple streams is there a way to still get the performance of multiple streams and drives yet still treat it as one 'job'?
0 Kudos

Stumpr2
Level 6

‎03-10-2008 09:54 AM


jplally wrote:
Thanks for your quick response. I'm re-reading the policy section of the admin manual but other than limiting the number of jobs per policy or not allowing multiple streams is there a way to still get the performance of multiple streams and drives yet still treat it as one 'job'?

no. each stream is an entity and individual image.
0 Kudos

TomerG
Level 6
Partner Employee Accredited Certified

‎03-10-2008 06:22 PM

Yes you can limit it by using streams...
 
i.e. in stead of your backup selection being:
 
/fs1
/fs2
/fs3
 
which results in 3 jobs, you can enable multistreaming (tic-box in the policy) and have a file selection list that looks like this:
 
NEW_STREAM
/fs1
/fs2
/fs3
 
this should result in 1 job with 3 file systems.
0 Kudos

Stumpr2
Level 6

‎03-10-2008 08:11 PM


Tomer Gurantz wrote:
Yes you can limit it by using streams...
 
i.e. in stead of your backup selection being:
 
/fs1
/fs2
/fs3
 
which results in 3 jobs, you can enable multistreaming (tic-box in the policy) and have a file selection list that looks like this:
 
NEW_STREAM
/fs1
/fs2
/fs3
 
this should result in 1 job with 3 file systems.

so the answer is still no
 
/fs1 /fs2 /fs3 will be just one stream /one entity/ one image
continuing on in oreder to get the 100 mount points
 
/fs5 /fs5 /fs6 will be just one stream /one entity /one image
 
/fs7 /fs8 /fs9 will be just one stream /one entity /one image
 
Ity doesn't matter how many file systems you include into a new_stream.
Each stream will be an entity unto itself /one image
 
There is no single job issued as the master of all streams. You are going to have an entry in the activity monitor for each and every stream regardless of how many file systems you group using the new_stream directive.  each new_stream entity becomes by definition an individual stream of joined file systems.
 
If one stream in the family fails then you only have to rerun that stream.
 
I hope that clears it up for you.
 

 


Message Edited by Stumpr on 03-10-2008 11:15 PM
0 Kudos

jplally
Level 3

‎03-11-2008 06:50 AM

Thanks Tomer, I will run some test with specifying streams and see how that looks. On a similar note, I've set up NOM to alert me if something fails, and because it seems to operate on a job (or stream or filesystem basis) instead of a client or  policy basis, I end up getting 96 alerts, one for each filesystem.
0 Kudos