Restoring a tape from a standalone drive & master server to a different master server with robotic library

[markdavies1978]

Hi all,

I have a tape which has a label of "131109" which was has been taken from a different Master Server which has a standalone drive, I need to restore data from this tape at a different site with different master server. the differnt Master Server has a robotic tape library.

I have run an inventory & the robot has issued the media ID of "A00001" I guess I need to label this tape as "131109" in order to follow the 2 step import process?

I have checked the thread below but not sure i follow:

http://www.tek-tips.com/viewthread.cfm?qid=1522130

And advice?

I'm running version 6.0MP4 on Win Server 2003 by the way!

[Andy_Welburn]

delete A00001 from the library & then use vmphyinv to 'recreate' the volume using its internal label of 131109, then follow the phase I & phase II import procedures.

If you try & use the label functionality of NetBackup you could (will?) lose the data on the tape.

[rrosario]

if its the same version master server...copy the the client's images to the other master's master server...it would work...

[markdavies1978]

Andy,

Should I delete A00001 from the "Netbackup" Volume Pool & from the Robot | TLD & then run the vmphyinv command?

Cheers.

[Andy_Welburn]

(altho' logic dictates that you remove it from the robot). Then run the vmphyinv - you could probably restrict this to just the slot that the specific tape is in & hopefully it will read the internal label & you can go from there.

[markdavies1978]

Cheers, think I'm there.

[Andy_Welburn]

let us know!

[Marianne]

What is the make/model of the library? Some libraries will not load unlabelled tapes into tape drives, even if NetBackup is updated with the correct media ID and slot number.

If you know in which slot the tape is in the library, test if you can mount it using robtest.

If that works fine, you can simply delete A00000 and manually add a new media id (131109), select the correct density that matches the drive type, select the robot number and type in the slot number.

[markdavies1978]

16/11/2009 14:56:35 - Error bptm(pid=740) cannot read image from media id 131109, drive index 1, err = 23
16/11/2009 14:56:35 - Error bpimport(pid=5108) Import of policy IOMDaily, schedule Full (iomp3fiprt1_1258221621) failed, media read error. 
16/11/2009 14:56:35 - Error bpimport(pid=5108) Status = no images were successfully processed.     
16/11/2009 14:56:35 - end Import; elapsed time: 00:33:01
no images were successfully processed(191)

Any ideas?

[Will_Restore]

http://seer.entsupport.symantec.com/docs/319167.htm

[NavA]

As suggested by rrosario copy images from under /usr/openv/netbackup/db/images/<clientname>

find /usr/openv/netbackup/db/images/<client-name> -name *POLICYNAME* -depth -print

copy all the files that are listed by the above command [replace client-name & POLICYNAME with appropriate values], The files have to be copied to exact same location path directory etc.

I used this quite a few times and it worked. Offcourse both my master servers were UNIX. So, i am not sure if it will work for UNIX-WINDOWS.

Next step is to do inventory of robot but you already did that. So once you copy the images you should be able to see the from BAR GUI.

Hope this helps.

[Marianne]

You need to find the reason for the media read error.
What make/model tape drive were images written on? Firmware version on tape drives?
Compare with config at new site. Read errors could be due to different driver and/or firmware versions on tape drives.
What errors are you seeing in bptm log?
If you have VERBOSE entry in vm.conf, you will be able to see device-related errors in syslog on Unix and Event Viewer on Windows.

[markdavies1978]

wrobbins - Don't think it's an encryption issue as we use the same encryption keys, will check though - Thanks!

NavA - I have done this already, thanks for your input.

Marianne - My tape drives are HP. The old Master server has a standalone internal LTO Ultrium-2 drive. The new master is using a HP MSL2024 Tape Library with 2 LTO Ultrium-3 drives could this be the problem? They are using the same driver by the way. In the Event Viewer I have a few Event ID 7's stating "The Device, \Device\Tape0, has a bad block.

Thanks.
 

[Will_Restore]

Ultrium-3  should be able to read Ultrium-2 volume.  Could be a bad tape. 
Also, could you verify the SSL key per the doc I posted above?

[Marianne]

Have a look at this TechNote for troubleshooting info on Event Id 7:
seer.support.veritas.com/docs/241684.htm

Are normal backups and restores working fine in this environment?

[markdavies1978]

wrobbins - I'm using the same key as the old Master Server, also the location specified in the link you provided doesn't exist within my install. Thanks.

Marianne - Normal backups & restores work fine. I don't really follow the article you sent but I do have a similar error within the event log.

[markdavies1978]

Just wanted to bring this back up.

Running vmphyinv works fine but when I then run an inventory on my robot it tries to logically add the media id's of tapes which are already present back.

Any ideas?

Why does Netbackup want to add the media id's of already present tapes back into the system?

 

[Andy_Welburn]

a) are tapes in incorrect slots maybe? (i.e. a mis-match between what the library thinks its got where & what NB thinks)
b) does NB think they have different barcodes? (or maybe densities - doubtful, but just thinking aloud whilst having lunch!)

What exactly are you getting error or message-wise & when?

[markdavies1978 · Accepted solution]

Just to clarify...

I had (before I inserted my stand alone tape into the library) about 15 tapes in my library already.

I took several tapes out from last nights backup jobs & inserted my standalone tape.

I then ran the VMPHYINV command on the slot in question & the stand alone tape was logically added fine. I then ran the import etc.

While my import was running I ran a robot inventory (update volume configuration option) where it advesed me to add all the media which was currently present.

Make sense.

Thanks for your time...