recovery without import fails

[ADDODD]

Hi,

I'm testing a DR plan.  I've got my main nbu environment set up to replicate the nbu catalog to the DR site.  At the DR site, i have a smaller nbu infrastructure (all nbu 6.5.4 on Solaris 10 (sparc)).

i have paused replication and mounted the catalog on the DR server.

I have created links for the clients I want to restore at /path/to/netbackup/db/images/<client>.

I then launch the Backup, Archive Restore app from the master server.  I specify the client in my production site as the source.  I specify a local client at the DR site as the target.  I can see backups for the source client listed in the browse box. I check a few folders for restoring and start the restore job.  The media server mounts the tape, but then fails (within about 30 seconds) with the error:

"The restore failed to recover the requested files"  (error 5).

I know that's a generic error, but that's all i get.

I have followed the "recovery without import" guide.  ON my DR media server, I have the FORCE_RESTORE_MEDIA_SERVER PRIMARY-MEDIA-SERVER DR-MEDIA-SERVER set in the bp.conf.

As far as I can tell, everything should be working.

any idea what I might be missing?

any idea where I can get more detail about why the restore fails?  (ie. which logs to check.)

(note: both sites have a master server and a separate media server.  name resolution works everywhere.  backups are all successful.  restore jobs in the DR site for the DR client work.  It's just restoring data from the Primary site at the DR site that fails).

thanks!

ADDODD

[Android]

Do these alternate targets/clients you are pointing the restore to have the NBU client software installed?

Do they have the proper entries in bp.conf for the proper master/media servers?

[J_H_Is_gone · Accepted solution]

1) tapes -do the tapes in your dr site have unique number from your tapes in your prod site?  They have to be different or it would try to pull the wrong tape.  (just a thought)

2) FORCE_RESTORE_MEDIA_SERVER PRIMARY-MEDIA-SERVER DR-MEDIA-SERVER
just double checking that this line has the real server names in it but you need an = sign

FORCE_RESTORE_MEDIA_SERVER = <media server in prod> <media server in dr>

3) when you put your prod tapes in your dr library you did write protect them and put them in a special spool so your DR server would NEVER try to use them for a backup.

4) if when you open the BAR and you can see the images info for the prod server on with your dr master chosen then you have that part just fine.

5) was replication up to date when you stopped it - meaning it had made current all the db/images files on the DR server?


I'll keep thinking but right now you have done it the way I have mine set up - just check the items I listed above.  (and once you get this to work - it is soooooo nice to not have to import tapes!!!)

[J_H_Is_gone]

i have paused replication and mounted the catalog on the DR server.

[Abesama]

If you want to look at debug log,

media server
/usr/openv/netbackup/logs/bpbrm
/usr/openv/netbackup/logs/bptm

DR client
/usr/openv/netbackup/logs/vnetd
/usr/openv/netbackup/logs/bpcd
/usr/openv/netbackup/logs/tar

I'd start with bpcd and tar log ...

Abe

[ADDODD]

I had the  FORCE_RESTORE_MEDIA_SERVER option specified in the bp.conf file on the media server without the equals sign.

I changed it so there was an equals sign (FORCE_RESTORE_MEDIA_SERVER = prod-server dr-server) and made sure the entries were on both the media and the master server at the dr site.

after that, it all worked.

i'm not exactly sure if it was the missing = sign or the fact that it was only specified on the dr media server (and not the dr master server).

anyway, it works now!

ADDODD