Forum Discussion

smith_jones's avatar
10 years ago
Solved

LIMITS JOBS PER POLICY

Hi,

I ran a policy containig some  paths in backup selections. Then after sometime I got that I have missed some paths to be added in backup selections of the same policy.The policy was running.

I opened the backup selections of the running policy and deleted the paths that I had specified before because they were already triggered. I added the missed paths and then saved the changes. I once again ran the policy (althogh it was already running but with some other paths defined in backup selections).

Now it is showing me as -

 

Info nbjm(pid=8520) starting backup job (jobid=1765) for client mailfiler1, policy NDMPBackup_SMESnapshots_mailfiler1, schedule Full 
Info nbjm(pid=8520) requesting STANDARD_RESOURCE resources from RB for backup job (jobid=1765, request id:{CA43BAD8-E0ED-4572-8400-1FD475C73BAE}) 
requesting resource mailfiler1
requesting resource mailbck1.NBU_CLIENT.MAXJOBS.mailfiler1
requesting resource mailbck1.NBU_POLICY.MAXJOBS.NDMPBackup_SMESnapshots_mailfiler1
Info nbrb(pid=6112) Limit has been reached for the logical resource mailbck1.NBU_CLIENT.MAXJOBS.mailfiler1

 

Also in the policy attribute the "Limits jobs per policy" option is not checked. I have blank tapes in the pool from where the backup is to be written.

 

Check the attachment.

How can I fix this?

  • GUI -> HOST Properties -> Master Server -> Select Master Server, right click , properties -> Global Attributes -> Maximum jobs per clients.

    Lift value to e.g 8 or 10.

  • Q) Now if maximum jobs per client is set to 1, does that mean it is applicable for all clients?

    A) If you are refering to the global setting, then yes.

    .

    There is not really an order of precedence, however there is implied 'scope' - by lack of precedence I mean that evaluation does not stop upon a condition or setting of any one maximum concurrent job/stream 'limit control' - instead all three 'limits' are always taken in to consideration.

    Of the three places where concurrent backup job/stream count can be controlled:

    - global setting - i.e. Master Server / Global Attributes / Maximum jobs per client : scope, applies to all backup jobs/streams for all backup policies for all backup clients.

    - policy setting - i.e. Policy attributes / Limit jobs per policy : scope, applies to all backup jobs/streams for this one backup policy only, for all backup client jobs/streams generated by this policy only.

    - client setting - i.e. Master Server / Client Attributes : scope, applies to all backup jobs/streams across all backup policies for this one backup client name only.

    ...then, whichever is present (set), the lowest of these will result in the maximum concurrent backup jobs/streams for any one backup client name.

    .

    N.B: none of the settings apply to any restore jobs at any time.

  • Case-1:

    As the "Limit jobs per policy" has lower value so it will be given more priority. Now tell me if all the clients are effected by this value or only the clients which are selected in the policy will be effected by this value?

    It's limit jobs per policy so what do you think will happen? It is set to 1 so only 1 job will be active for that policy at any one time.

    Case-2:

    As the "Maximum jobs per client" has lower value so it will be given more priority. And as suggested by you guys all the clients will be effected by this value regardless of the fact which clients are in which policy.

    Correct. Only one job will run per client, you could get up to 2 jobs running per policy (different clients). Be aware that if you have the same client in more than one policy that run at the same time then only one can run due to the "Maximum jobs per client"

    Case-3:

    So only the client I added in client attributes will be affected by the value of Maximum data streams and other clients will be affected by the value of "Limit jobs per policy" or "Maximum jobs per client", which one is lower.

    Correct.

     

    Just to throw something out there, there is another layer of complexity you may want to at least be aware of - you can also limit the overall number of jobs that can run at any one time by how you configure your STUs (max jobs/drive , max concurrent jobs)

     

  • Also user "sdo" -

    You were going to do some validation at your end. If you find something for me please share.

    Thanks.