transfering BMR information hang

[santave]

Hi,

Previously client backup with BMR run completed succesfully.
Then its hanging when transfering BMR information to the master server on the last 2 days.
It just stuck on below log, then in few hours it will terminated by admin. Attached is the bmr log file.

Please suggest how to check and fix this issue?
 

01/26/2014 09:55:18 - begin Stream Discovery: Bare Metal Restore Save
01/26/2014 09:55:19 - collecting BMR information
01/26/2014 09:55:19 - connecting
01/26/2014 09:55:19 - connected; connect time: 0:00:00
01/26/2014 09:55:19 - transfering BMR information to the master server
01/26/2014 09:55:19 - connecting
01/26/2014 09:55:19 - connected; connect time: 0:00:00
01/26/2014 09:55:19 - started process bpbrm (pid=7602)

Thanks,
Santave

[RiaanBadenhorst]

Hi,

Dont let the admin cancel the job, let it fail so we can see the status code. What happens to the other jobs for this client?

[santave]

Hi,

Other jobs are not happening, as first job is transfering bmr info. bpbkar also not created yet.
BMR and standard backup for the other clients in the same policy is fine. Only 2 clients I have such issue.

[RiaanBadenhorst]

I doubt this has anything to do with BMR. Make a new policy for these 2 clients , turn off BMR on the policy and run it.

[santave]

I tried disabled BMR on the current policy, run backup, its completed succesfully.

[RiaanBadenhorst]

Hi,

Ok, turn it back on and then let the backup run and make sure it fails, don't stop it. Need to see the error codes.

[santave]

Any other way to check the hang process, before fails?
Its already 4 hours now and it stuck on  - started process bpbrm (pid=7602)
 

[RiaanBadenhorst]

What versions are you running?

http://www.symantec.com/docs/TECH163185

[santave]

Master, Media in  linux redhat 6.2, nbu 7.5.0.6

I read the technote,My BMR is not heavy load. I have bmr parameter also higher than on the technote.
The others clients able to run bmr and backup completed with no issue.
Only 2 servers, all linux clients, I have the problem currently.
 

[RiaanBadenhorst]

Kill the bpbrm process on the media server and restart the backup for those two clients then.

[Marianne]

Only 2 servers, all linux clients, I have the problem currently.

You gave us master and media server versions, but not the NBU and OS versions on the 2 problematic clients?

Seems troubleshooting should be concentrated on clients....

[santave]

The 2 clients have the same OS version and nbu, redhat 6.2, nbu 7.5.0.6
I have restarted nbu services on master, media and clients also, but no luck.

Do you think, the new bmr can't update the exisiting bmr image?
Should I delete old bmr image? (too bad, can't rename the old bmr image)

[RiaanBadenhorst]

Please get the logs

vxlogcfg -a --prodid 51216 --orgid 119 -s DebugLevel=6 -s DiagnosticLevel=6

The above will enable debug and diagnostic level 6 for OID 119: BMR Master Server Daemon (bmrd)  .

Herewith a list of BMR OID's:

119. BMR master and boot server (bmrd bmrdb)

121. BMR Save Configuration  (bmrsavecfg)

Repeat above vxlogcfg command for all required BMR processes (--orgid).

Command to collect logs on master

vxlogview -i 119 > <Path>\bmrd.log

Command to collect logs on client

vxlogview -i 121 > <Path>\bmrd.log

Also provide the bpbrm log from media server.

[santave]

Before I run vxlogcfg, how to see my current config?
After log generated, how to rollback my log config?

If bmr failed, bmw config will not updated.
But FS standard backup is completed successfully, when I disable BMR.
In such case, Can I use the old bmr config for OS restoration and new FS backup image for data?

[mandar_khanolka]

>> In such case, Can I use the old bmr config for OS restoration and new FS backup image for data?

It would work well with BMR recovery using "current" config. But in this case there should not be any change happened on your client system which can hamper BMR recovery. Changes like, disk/volume/fs/NIC etc.

BTW check bpbrm (media server),, bmrd (bmr master server) and bmrsavcfg (client) logs to trace this issue.

This is weird. I think problem is with process communication at media server. The logs would help to figure this out.

Thanks.

~Mandar

[santave]

Hi, Please find all the logs attached.

I tried, created new test policy with bmr enabled, backup selection = /tmp
Still same issue, log attached.
 

[mandar_khanolka · Accepted solution]

backup selection is the problem. in order to take bmr enabled backup you need to consider at least OS file systems backup and netbackup client directory backup.

Just backing up /tmp wont be sufficient to recover the system back.

If you just want to see BMR enabled backup as success then add below directory path in your backup selection list. NOTE: Until and unless you add ALL_LOCAL_DRIVES or at least OS file system (+ system state in case of windows) you wont be able to recover your system.

Add NB CLient dir path in backup selection list:

for windows client, for instance:

c:\program files\veritas\netbackup\baremetal\client

for unix client, for instance,

/usr/openv/netbackup/baremetal

Try bmr enabled backup now.

Thanks.

Mandar

[santave]

Do you mean, change backup selection to /usr/openv/netbackup/baremetal ? (create blank directory first)

note: I have another policy for this server that backup all_local_drives with BMR disable.