Drive free time

[ronak_padhya]

Issue:

I have just 1 drive in my server and i have enabled multiplexing to 2 already, for using the drive for multiple backup jobs.

There are too many backup jobs in the server. And many of the jobs are failing because of error code "1: (196) client backup was not attempted because backup window closed ".

Now the issue is that as there are too many jobs in the server its difficult to find out when the drive is free so that i can re-schedule some failed jobs to the free time to decrease the queueing.

Question:

Is there any way i can find that out, for example:- for last day when the drive was free(no backup activity was going on)?

As an alternative is there any other way i can solve the above issue so that all the jobs work smoothly and get distributed evenly?

[RamNagalla · Accepted solution]

simply i would say.. open backup schedule window for 24X 7 and set the frequency based on your requirement..(may be daily once or weekly once.)

if you do this.. backups will not fail with 196, they will wait in Queue for Drive and use the Drive when ever it is avalible..

you can also set the Muliplexing value as 4 to increse the number of active jobs per Drive.

if you are jobs are not compleating even after the Drive is writing 24X7, then you need to look to increaes the capacity.

[sdo]

How about also prioritise your backup policies, so the most important get done first?

Then for any backup clients where it doesn't matter if the backups run during the day - you could increase the backup window, lower the priority - so that they get done last.

If you only have one storage unit, with one tape drive, in your environment, then surely the tape drive is free when no backups are running.

[sdo · Accepted solution]

Try increasing multi-plexing to 3, and then to 4 - but any multi-plexing (2+) can have significant impact upon restore times.  You have to be willing to take the hit on restore times.  I have seen environments go to multi-plexing of 8 - out of sheer desparation, just to get some backups.

Did no one do any sizing calculations for the environment?  These would clearly have caught any shortcomings before installation/implementation.

[sdo]

You also have to make sure that all backup policies and all schedules use the same volume pool and the same retention - otherwise they will not multi-plex together.

[GulzarShaikhAUS]

I totally agree with Ram and SDO.

In case you really need to see the drive utilization and you have opscenter installed then the report "Drive utilization" would give you graphical view of the drive utilization.

[Marianne]

Check that MPX is selected in backup schedules as well.
The default is 1 and will override STU setting.

[Nicolai · Accepted solution]

All suggestions above are good.

Also ensure you cut down volume pools and retentions as much as possible

And now my 0.02USD:

What about backup to a disk storage unit  and then schedule stagning to tape ?

You can either go for the DSSU opption in Netbackup or the AdvancedDisk/MSDP option with SLP duplication to tape. Scheduling to tape can then be open 24*7.

Make sure NUMBER_DATA_BUFFERS in Netbackup is set to 256 and SIZE_DATA_BUFFERS to 262144 (default in 7.5 and up).

Also consider to bite the dust and buy one more tape drives. You are maybe trying to postpone the unavoidable :)

[revarooo · Accepted solution]

I have another suggestion. Add more drives! What would you do if backup was running on your 1 drive and you needed to do an urgent restore? Ok you can cancel the backups, but what if that restore was going to take multiple hours/days to restore, you won't have any jobs running to tape and will get a backlog.