VMware BLIB Backup failing to tape, works to DeDupe Pool

[BoerLowie]

Hi,

i have a VMware policy where want to backup our File Servers to tape.  I have an SLP created which only has a backup to tape in it.  The Policy has BLIB enabled.

If i try to backup our VM(s) with this policy, i errors our with the following output:

6/2/2014 1:12:15 PM - Info nbjm(pid=5464) starting backup job (jobid=19323) for client SERVERNAME, policy VMWARE_FILESERVER, schedule Full  
6/2/2014 1:12:16 PM - estimated 0 Kbytes needed
6/2/2014 1:12:16 PM - Info nbjm(pid=5464) started backup (backupid=SERVERNAME_1401707535) job for client SERVERNAME, policy VMWARE_FILESERVER, schedule Full on storage unit hcart-robot-tld-0
6/2/2014 1:12:18 PM - started process bpbrm (13410)
6/2/2014 1:12:20 PM - Info bpbrm(pid=13410) starting bptm           
6/2/2014 1:12:20 PM - Info bpbrm(pid=13410) Started media manager using bpcd successfully       
6/2/2014 1:12:25 PM - Info bpbrm(pid=13410) xxx is the host to backup data from     
6/2/2014 1:12:25 PM - Info bpbrm(pid=13410) telling media manager to start backup on client     
6/2/2014 1:12:25 PM - Info bptm(pid=13417) using 262144 data buffer size        
6/2/2014 1:12:25 PM - Info bptm(pid=13417) using 64 data buffers         
6/2/2014 1:12:26 PM - Info bpbrm(pid=13410) spawning a brm child process        
6/2/2014 1:12:26 PM - Info bpbrm(pid=13410) sending bpsched msg: CONNECTING TO CLIENT FOR SERVERNAME_1401707535     
6/2/2014 1:12:26 PM - Info bpbrm(pid=13410) child pid: 13452          
6/2/2014 1:12:26 PM - Info bpbrm(pid=13410) start bpbkar on client         
6/2/2014 1:12:26 PM - Info bptm(pid=13417) start backup           
6/2/2014 1:12:26 PM - Info bptm(pid=13417) Waiting for mount of media id B00027 (copy 1) on server bedc-nbu0001. 
6/2/2014 1:12:26 PM - Error bpbrm(pid=13452) from client SERVERNAME: ERR - Missing block incremental notify script /usr/openv/netbackup/bin/bpstart_notify.VMWARE_FILESERVER  
6/2/2014 1:12:26 PM - Error bpbrm(pid=13452) from client SERVERNAME: ERR - bpstart_notify returned -1     
6/2/2014 1:12:26 PM - Error bpbrm(pid=13452) client SERVERNAME EXIT STATUS = 73: bpstart_notify failed     
6/2/2014 1:12:26 PM - Info bpbkar(pid=0) done. status: 73          
6/2/2014 1:12:26 PM - Info bpbrm(pid=13410) sending message to media manager: STOP BACKUP SERVERNAME_1401707535     
6/2/2014 1:12:26 PM - connecting
6/2/2014 1:12:26 PM - connected; connect time: 0:00:00
6/2/2014 1:12:27 PM - mounting B00027
6/2/2014 1:12:32 PM - Info bpbrm(pid=13410) media manager for backup id SERVERNAME_1401707535 exited with status 150: termination requested by administrator
6/2/2014 1:12:32 PM - end writing
bpstart_notify failed(73)

If i try to backup to a DeDupe pool, it works with the exact same settings!

[Nicolai]

What version of Netbackup are you on ?

[Mark_Solutions]

What version of NBU are you using?

Are you also using VMware accelerator? If so BLIB won't work to tape.

Can't see any other restrictions but i never do them to tape!

#Edit#

This tech note implies that it should be fine but maybe there are bugs, are you up to date?:

http://www.symantec.com/docs/TECH201453

[BoerLowie]

7.6.0.1 on the master (physical Win 2008 R2 server) & 2.6.0.1 on our 5230 appliance

[BoerLowie]

No, accelerator is offcourse not used because we're going to tape...

It should work but it doesn't...  I tried backing up different VMs but no change.

[Mark_Solutions · Accepted solution]

Maybe it is just the change of storage that causes it due to reference file somewhere

Start here: http://www.symantec.com/docs/TECH182631  and delete that file if it exists

Something seems to need re-setting - just not sure what!

[BoerLowie]

I checked /usr/openv/netbackup/bin on the appliance (as that is the client doing the VMware Backups) and no BPSTART**** files exist...

[Mark_Solutions]

I thought as much!!

I can only assume that the information is held in the catalog somewhere

Try making a copy of the policy so that it appears as a new one and then run a Full backup from there.

If that still fails then it may need something doing in vCenter itself to flush the Change Block Tracking in some way  - but dont know the deepest details on how it all works - try a new policy first

[BoerLowie]

Got it solved by creating a bpstart_notify.POLICYNAME on the appliance.  Followed this technote with support:

http://www.symantec.com/docs/TECH182631

So the same procedure as you mentioned, but the creation of the file did the trick!

[BoerLowie]

Tagged this post as the solution as this article guided me to the solution!

Thanks Mark!

[vic_medina]

Very goo the best