Queued jobs - free drives

[TA0]

Hello,

Windows 2012 Master

2 x Windows 2012 Media Servers

Tape Library with 4 x LTO6 drives

Storage Unit A: media server 1, four tape drives

Storage Unit B: media server 2, four tape drives

Both storage units are combined in a single storage unit group which is the policy storage for all my backup jobs. 'Media server load balancing' is configured on the storage unit group.

Multiplexing: 4

Max concurrent drives: 4

Allow multiple data streams: yes

Media multiplexing: 4

Maximum jobs per client: 4

I have backup jobs which always seem to favour one storage unit and I have queued jobs when there is a free storage unit and two free drives. Can someone point me in the right direction please?

Thanks

[Mark_Solutions]

When a job is queued what does it say on the detail of the waiting job? (or in the State section of the admin console) - that should give the best clue of what it is actually waiting for.

Does the second storage unit ever get used? If not then there may be a configuration error somewhere.

Storage unit groups can cause odd issues some times but please show us as much of your configuration as possible - the outputs of these would be a good start

bpstulist -g -U

bppllist policyname -U (for each policy running when jobs are queued)

[TA0]

Thanks for the reply.

I pulled this out of the Job Details earlier 'limit has been reached for the logical resources masterservername.NBU_CLIENT.MAXJOBS.backupclientname'

Attached is the other information you asked for:

Thanks again

[Andy_Welburn]

'limit has been reached for the logical resources masterservername.NBU_CLIENT.MAXJOBS.backupclientname'

So it's saying it has reached the maximum number of concurrent active backup jobs for that specific client - from your OP you must already have 4 active jobs/streams for (each of) the queued client(s).

[Mark_Solutions]

When the policy runs how many streams are created for each client?

Check the Master Server Host properties - Client Attributes section to see if the clients are listed in that location - if so do they have a max jobs setting against them

Also check on the Global Attributes tab to see what the Mximum Jobs Per Client is set to.

The NBU_CLIENT.MAXJOBS.backupclientname indicates that maximum number of jobs per client is restricted in one or both of the above locations

If you make a change to either / both then re-start netbackup on the master to ensure the changes are registered and try again

[TA0]

Thanks for the replies both.

Master Server Host Properties>Client Attributes: no clients listed

Master Server Host Properties>Maximum jobs per client: 4

Only one storage unit working out of two and only two drives running out of four. There only seems to be two streams running.

Thanks

[TA0]

Still haven't resolved this entirely...

This is an IBM TS3310 library with 4 x LTO6 drives. Could this be a multipath issue?

1. I've disabled the Data Path Failover and configured the library with SCSI Persistent Reserve and mirrored this in the NBU config
2. Control Path Failover is activated and I've configured all 4 tape drives as a potential Control Path - is this a good option?
3. My media servers have 4 x FC ports and the tape drives have x 2 my zoning is as follows:

• HBA 1 port 1 => tape device 1
• HBA 1 port 1 => tape device 2
• HBA 1 port 1 => tape device 3
• HBA 1 port 1 => tape device 4
• HBA 1 port 2 => tape device 1
• HBA 1 port 2 => tape device 2
• HBA 1 port 2 => tape device 3
• HBA 1 port 2 => tape device 4
• HBA 2 port 1 => tape device 1
• HBA 2 port 1 => tape device 2
• HBA 2 port 1 => tape device 3
• HBA 2 port 1 => tape device 4
• HBA 2 port 2 => tape device 1
• HBA 2 port 2 => tape device 2
• HBA 2 port 2 => tape device 3
• HBA 2 port 2 => tape device 4

So in the Windows Media Server Device Manager I can see x 16 tape devices and x 16 tape libraries. Is this going to casue me problems? Any recommendations? I can't see in the TS3310 docs any limitations on paths

Thanks

[Marianne]

If you use the wizard to config devices, NBU will use the unique serial numbers of each device (robot and tape drives) to create one device in NBU per physical device.

Please check output of these commands:

scan -changer

scan -tape

tpconfig -l

(these commands are in ...\veritas\volmgr\bin. Remember to open cmd with Run As Administrator)

[TA0]

Thanks for the respone

To confirm then, my zoning is fine and at the OS level seeing 16 libraries and tape devices is not a problem as NBU can handle it?

Is IBM Control Path Failover supported by NBU?

I've attached the output of the commands you asked me to run

Cheers

[Marianne]

Where did you type these commands? On which server?

scan sees no tape library and no tape drives.

tpconfig shows no devices configured in NBU.

All the zoning you mentioned and none of the devices respond to scsi-commands issued by scan command?

Not sure where 'Queued Jobs - free drives' fit in the picture if no devices is seen and nothing is configured!

[TA0]

I'll tell you what I run the commands on the media server   No excuses I was just being dumb - see attached

I notice I've got two dropped paths from the output which I need to resolve on IBM.ULT3580-TD6.004. Can't see any obvious issues at the moment. However, that's the point of the additional paths so I wouldn't think that it was related to the issues I am seeing. So at the moment the media server is seeing 14 drives and 14 paths, but as you mentioned Marianne, NBU sees 2 libraries and 4 drives.

Thanks

[TA0 · Accepted solution]

Hi,

Thanks for all your help

I seem to have resolved the issue. IBM Data Path Failover was configured on the problematic media server so I removed it and reconfigured the library, drives and Storage Unit. I've also resolved the path issue.

It doesn't explain all the issues I was seeing but it seems to be working fine now.

Thanks again