Tape Library - No concurrent jobs!

[CallwaveEmp]

Hi All -
I reinstalled Arcserve a few weeks ago, and upgraded to V7sp4. I am using an Adic Scalar 100 w/ 48 slots & 3 tape drives.

All 3 tape drives show up as being controlled by the library, however, I can no longer run more then 1 job at a time.

I hope this is a misconfiguration that I can quickly fix, not a bug. Can anyone help?

D

 

[jc536]

Have you configured more than 1 device in the Device manager? In Devices, you need to configure 3 devices (Groups) (i.e. for the 3 drives) and assign each device/group some tapes to use. By default, Arcserve only creates 1 group (Group0)for backup tapes. When you create more than 1 backup job to run simultaneously, each backup job needs to be assigned a different group of tapes in order to run. When creating a backup job, and selecting the Destination Tab, do you see more than 1 "Group" (Group0) for selection. If not then you must configure 2 more in the Device manager and assign each group some tapes.

 

[CallwaveEmp]

Yeah, I had multiple groups - but 2 jobs were set to share a group by mistake.

 

[jabus]

Did you install the TLO? Did you run Device Configuration and make sure it sees everything ok?

You do not need to create multiple device groups for the backups to use all the drives. If they are scheduled jobs (backups/restores/merges), and not Run Now jobs, all drives will be used.

There are some things that cannot be done while a backup, etc., is being run, such as Import/Export tapes.

We are using libraries with 6+ drives and 300+ slots, and they are all configured with a single device group, and we don't have any contentions with the backups.

 

[CallwaveEmp]

Yes to TLO, and Yes to Device Configuration.

A reboot seemed to fix the problem.

If you don't mind my asking, why use only 1 device group instead of multiple?

 

[jabus]

We have clustered servers in a SAN environment, and the primary server doubles as the Central DB server (with the dist nodes not keeping a local copy of the db).

There is a media pool defined for each cluster, which all nodes that belong to that cluster use said media pool. (We do non-GFS rotation jobs.)

Having to import tapes to specific slots can be error-prone, so have the ARCserve manage the media in the library (300+ slots) as a whole is easier on the admins. This works pretty good, with the exception of not being able to import/export while backup jobs are running (devgroup is locked).

 

[CallwaveEmp]

So you have 1 job per media pool, or you manually keep track of multiple jobs schedules?

(Thanks for the discussion - I appreciate it)

 

[jabus]

We have multiple jobs per media pool (each node in a cluster shares a media pool). You can ignore the useless warning when submitting a job about the "media pool in use by another job".