Solved

Storage policy jobs preference

  • 23 July 2021
  • 5 replies
  • 208 views

Userlevel 4
Badge +13

Hi,

in our case there is a lot of running jobs accros multiple storage policies. Once the jobs reach more than 90% of completion, the throughput of the job falls and the speed of the job is very slow. However,some of the jobs, that have less than 50% in progress status, are running faster.

Is there any option to prioritize jobs, that are almost finished (with more then 95% of progress)?

I know there is possible to change priority of the single job, however, from my experiance it didnt increase speed too much. Maybe I should have used the highest priority, that I exactly dont know.

icon

Best answer by Mike Struening RETIRED 6 August 2021, 19:35

View original

5 replies

Userlevel 6
Badge +14

Hello drPhil,

Once the Job reaches 85% it usually (depends on Agent Type) would be in the Archive Index phase. It’s unlikely to see any “throughput” for that phase of the Job since the real processing of data would already have been completed. This is simply saving the Index Data to the Backup Media.

- Can you check the “Attempts” Tab of a recent Job and check how long it spends in this phase?

 

Best Regards,

Michael

Userlevel 4
Badge +13

Hello drPhil,

Once the Job reaches 85% it usually (depends on Agent Type) would be in the Archive Index phase. It’s unlikely to see any “throughput” for that phase of the Job since the real processing of data would already have been completed. This is simply saving the Index Data to the Backup Media.

- Can you check the “Attempts” Tab of a recent Job and check how long it spends in this phase?

 

Best Regards,

Michael


Hi @MichaelCapon, you are absolutely right. For the backup operation there are no throughput statistics respectively after reaching 85% it starts Archive Index phase.

 

However, I thought Auxiliary copy operation. For Aux copy jobs, sometimes Current througput is falling to zero when the job reaches more then 95%. Most likely there are not enough streams, on the other hand, it would be better to finish one almost finished job rather then move resources to the different ones.

Userlevel 7
Badge +23

Hey @drPhil , before I mark this answered, I wanted to add in that what is most likely happening is that your Aux Copy has X number of source streams.  Let’s say 50.  The earlier the stream, the more likely it was used for a Primary job which means those streams are bigger.  As the smaller streams finish in the Aux Copy, there’s less stream activity )since many of them are done) and therefore it looks like throughput is diminished.

That might not be the exact issue, but it is most likely at play.

Let me know if you have any further questions!

Userlevel 4
Badge +13

Hey @drPhil , before I mark this answered, I wanted to add in that what is most likely happening is that your Aux Copy has X number of source streams.  Let’s say 50.  The earlier the stream, the more likely it was used for a Primary job which means those streams are bigger.  As the smaller streams finish in the Aux Copy, there’s less stream activity )since many of them are done) and therefore it looks like throughput is diminished.

That might not be the exact issue, but it is most likely at play.

Let me know if you have any further questions!

 

Thanks @Mike Struening for the explanation! I admire such a deep knowledge of the Commvault software!

Userlevel 7
Badge +23

Happy to help!!

Reply