wierd behavior trying to use policies with tape copy NBU 10.3 (unpatched) with Windows 2K22 primary.

[KBfinder506]

Hi guys, I'm deeply sorry to bother you with so many issues with this combo; but I want to know there something left to be done before open tickets .

Now I'm facing a weird behavior with the policies with copy on their schedule. For any background:

• The primary server is a bare metal (new hpe proliant 380) also media server running on windows 2022. No staging disk (just a RAID1 with OS+NBU)
• The library is a HPE StoreEver MSL3040; config without partitions and all FC LTO9 drives are ok.
• Maximum concurrent write drives: 4
• Maximum jobs per client: 50
• All LTO9 media has already been initialized.
• Configure Multiple Copies
  • Copy 1 -> STU (any_available) -> Volume Pool (CDRsPool1) -> Retention (5 years) -> if fails (continue) -> media owner (any)
  • Copy 2-> STU (any_available) -> Volume Pool (CDRsPool2) -> Retention (5 years) -> if fails (continue) -> media owner (any)
  •  

The config of NBU has several pools which are paired. Like CDRsPool1 and CDRsPool2 (each of the pools has one tape). The policy has one schedule that makes copies. So when is running take the tape of CDRsPool1  and make a copy over the tape of CDRsPool2.

Jan 15, 2024 2:08:07 PM - Error nbjm (pid=16248)  NetBackup status: 96, EMM status: No media is available

^^^^ Each pool has 1 tape.
Jan 15, 2024 2:08:07 PM - Info nbjm (pid=16248) starting backup job (jobid=366) for client myserv.onmydomain.net, policy Respaldo_my-medcr01, schedule ComptelFULL
Jan 15, 2024 2:08:07 PM - Info nbjm (pid=16248) requesting STANDARD_RESOURCE resources from RB for backup job (jobid=366, request id:{1CA30367-8988-4C5D-BD97-39E2862F3A76})
Jan 15, 2024 2:08:07 PM - requesting resource  __ANY__
Jan 15, 2024 2:08:07 PM - requesting resource  __ANY__
Jan 15, 2024 2:08:07 PM - requesting resource  myserv.onmydomain.net.NBU_CLIENT.MAXJOBS.myserv.onmydomain.net
Jan 15, 2024 2:08:07 PM - granted resource  myserv.onmydomain.net.NBU_CLIENT.MAXJOBS.myserv.onmydomain.net
Jan 15, 2024 2:08:07 PM - granted resource  000012
Jan 15, 2024 2:08:07 PM - granted resource  HPE.ULTRIUM9-SCSI.000
Jan 15, 2024 2:08:07 PM - granted resource  myserv.onmydomain-01-hcart3-robot-tld-0
Jan 15, 2024 2:08:07 PM - Error nbjm (pid=16248) NetBackup status: 96, EMM status: No media is available
Jan 15, 2024 2:08:07 PM - estimated 0 kbytes needed
Jan 15, 2024 2:08:07 PM - Info nbjm (pid=16248) started backup (backupid=myserv.onmydomain.net_1705349287) job for client myserv.onmydomain.net, policy Respaldo_my-medcr01, schedule ComptelFULL on storage unit myserv.onmydomain-01-hcart3-robot-tld-0
Jan 15, 2024 2:08:07 PM - started process bpbrm (pid=4760)
Jan 15, 2024 2:08:08 PM - connecting
Jan 15, 2024 2:08:16 PM - connected; connect time: 0:00:00
Jan 15, 2024 2:08:18 PM - Error bptm (pid=4360) INF - no storage units available for use (213), cannot continue with copy 2 for index 1

^^^^ ???
Jan 15, 2024 2:08:18 PM - mounting 000012
Jan 15, 2024 2:10:14 PM - mounted 000012; mount time: 0:01:56
Jan 15, 2024 2:10:14 PM - positioning 000012 to file 9
Jan 15, 2024 2:12:07 PM - positioned 000012; position time: 0:01:53
Jan 15, 2024 2:12:07 PM - begin writing

In order to make an extreme test, I launch two different policies to use 4 drives (each policy should use two drives); both policies told me kind of same messages, policy 1 made the backup on tape 1 and canceled the copy con tape 2. But policy 2 made the backup con tape 2 but canceled the copy on tape 1.

Did you see any logic or point to check before opening tickets? By the way, there aren't any messages about fault on hardware, neither on the library nor the server. Also checked the inventory and the DB integrity all ok and fine.

[StoneRam-Simon]

I believe your problem is the residency being set to any available. I would try by setting the residencies to actual storage units so you can control which is being selected.

[KBfinder506]

Thanks a lot @StoneRam-Simon have already tried, same error. I will look if the current logging level shed light over the issue.

[davidmoline]

Hi @KBfinder506 

Two suggestions to try/look at. 

Firstly, using the current config, rather than preallocating media to a volume pool, put all the unassigned media in the scratch pool and let NetBAckup move them as required to the volume pool for each copy (this is the more normal way to manage media in NetBackup. 

The second would be to use an SLP to manage the copies, the first op would be the backup and the second operation would be duplication to the second tape/volume pool. 

Cheers
David

[davidmoline]

Hi @KBfinder506 

Another item comes to mind - does the media in the pool already have data on them? If so, are you trying to mix data retention on the one media (NetBackup doesn't like doing that, but can be configured to do so)? 

Check the media setting page on the master server properties and make sure everything is right.

David

[KBfinder506]

Hi people:

Well as some of you suggested; the main issue was that some backups had different retention times and I wasn't aware till dive a lot more.

About the suggestions mentioned, we are migrating from HPE Dataprotector (now Microfocus Dat{something}) and there's too much need about physical media control. Still we are trying to bridge both worlds in order to keep the workflow smooth as much is possible.

Thanks a lot to everyone. 

[StefanosM]

(now opentext DataProtector)