It is not easy and quick to start a an immediate restore or backup job in a middle of backup window when all resources are taken by the scheduled jobs. It is possible to suspend and or kill the current running job and deactivate policies in order to prevent new jobs from starting. I propose to create an immediate priority job option and button for GUI, which will make sure that the job with immediate priority will get resources right away. That should be done without deactivating policies and suspending or killing current active and queued jobs. I think such job should be highlighted in the GUI. It should be a dynamic limit for such jobs based on server resources. The immediate priority should not be available on policy level. It would be also nice to change a regular job to immediate priority if necessary.
The only issue I see with this
Is that what if all of your drives are currently being used for backups. If you are able to give a job "immediate" priority, how does the system decide which backup to kill in order to use a drive?
One way to do this, assuming that there will not be an issue by killing any currently running backups, is to suspend the scheduling using nbpem. Then, select one backup to kill, kill it, and since you suspended scheduling, noting will hurry up and kick off in its place. Then you will have a free drive.
Dont forget to resume scheduling however.
Netbackup should figure out
Netbackup should figure out the jobs with lowest priority and smallest size. Than find a drive (for MPX jobs) which has largest number of such jobs and suspend those jobs. Of course, the drive should not be available for new jobs. The suspended jobs should be automatically resumed in configurable number of minutes after the restore is completed.
Would you like to reply?
Login or Register to post your comment.