Restore Issue..

[Shyam_Prasad]

Hi All,

This's regarding one restore issue.
when i Initiate a Restore for one host which is using NetBackup 5.1 version, it is countinously failing with following error::

11/26/2009 19:16:45 - mounting TF2556
11/26/2009 19:17:23 - mounted TF2556; mount time: 0:00:38
11/26/2009 19:17:23 - positioning TF2556 to file 547
11/26/2009 19:20:04 - Warning bptm (pid=16290) cannot locate on drive index 14, locate scsi command failed, key = 0x8, asc = 0x0, ascq = 0x5
11/26/2009 19:20:05 - Warning bptm (pid=16290) read error on media id TF2556, drive index 14 reading header block, No space left on device
11/26/2009 19:21:33 - Error bptm (pid=16290) ioctl (MTBSF) failed on media id TF2556, drive index 14, I/O error (bptm.c.6059)
11/26/2009 19:23:13 - Error bptm (pid=16290) ioctl (MTFSF) failed on media id TF2556, drive index 14, I/O error (bptm.c.6315)
11/26/2009 19:23:13 - Warning bptm (pid=16290) FREEZING media id TF2556, it has had at least 3 errors in the last 12 hour(s)
11/26/2009 19:23:14 - Error bptm (pid=16291) The following files/folders were not restored:
11/26/2009 19:23:14 - Error bptm (pid=16291) UTF - /C/RM_MountVol/ETFL17P32001A_SG1_BCV2/Gdrive/DATABASES/SG1/DB1/
11/26/2009 19:23:15 - Error bptm (pid=16291) UTF - /C/RM_MountVol/ETFL17P32001A_SG1_BCV2/Gdrive/DATABASES/SG1/DB1/ETFL17P32001A_SG1_DB1_ALL.edb
11/26/2009 19:23:15 - Error bptm (pid=16291) UTF - /C/RM_MountVol/ETFL17P32001A_SG1_BCV2/Gdrive/DATABASES/SG1/DB1/ETFL17P32001A_SG1_DB1_ALL.stm
11/26/2009 19:23:16 - restored from image shkg17p37101-bkup_1250768817; restore time: 0:07:22
11/26/2009 19:23:19 - end Restore; elapsed time 0:08:15
the restore failed to recover the requested files (5)


No issues found on Tape and Drive as well... tried to make the drive index 14 down and again submitted the restore...it is going to another drive and failing with the same error..!!! My master server version is 6.5.1. 

 and one more thing i need to add here is that
the data which i'm trying to restore has been backed up in the year 2006(where our master is in the version 5.1)
i checked the buffer settings and all...
but still not able to resolve...

can any one pls tak a look into this and advise...

--Thanks,
Shyam Prasad

[Marianne]

Why are you saying  'no issues found on tape'? The error messages says that the tape may not be healthy.
These errors are hardware related: 0x8, asc = 0x0, ascq = 0x5
You have proved that the problem remains with the media, regardless of drive being used.
What errors are you seeing in bptm log on the media server?

[Nicolai · Accepted solution]

Sense key 0x8 asc = 0x0, ascq = 0x5 is a "Blank check End-of.data.detected".

You are getting this error because the tape drive can't locate the backup image requested (it's like saying "go visit house 200", but at the end of the road you find house 100).  Replacing tape drive will only result in the same error message, the Sense key you are seing is a information message and not a error message.

Possible reasons:

• Overwritten by another Netbackup domain
• Media deassigned in error
• Possible corrupt image header in the Netbackup database.

SCSI sense keys (hp site)

[Shyam_Prasad]

Hi,,

Thanks for your reply...

But when i try using different media also..Restores are failing with the same error.
and also tried using different drives....but still no luck !!

and now how can i proceed to resolve this???

Thanks,
Shyam Prasad

[Nicolai]

You need to investigate why you are getting this "blank check". The tape drives are telling you they can't find what you are looking for.

Have you tried to restore a different client ?
Have you tried to restore from a different media ?
Have you checked the consistency of the Netbackup database ?
Try to run "Images on media" under Netbackup GUI - Report - Images on media
Do you have tapes with duplicate barcodes (think twice) ?

Do a bpimagelist -backupid shkg17p37101-bkup_1250768817

Does the media TF2556 get listed in the FRAG option ?

[Shyam_Prasad]

Hi Nicolai..

Thanks again for ue reply..

Yup...i tried all the above reason mentioned..

i tried to restore from a different media and from a different client and to different destination as well..
when i try all these things restores are successfull.

and no possibilities of having a duplicate barcodes...
and i checked bpimagelist command and found the Media for that backupimage.

Thanks
Shyam Prasad

[rj_nbu]

1. Your master server is at version 6.5.1, the media server which failed to read the tape is still at 5.1
2. Another media server was able to restore the files ? what is the netbackup version  on that media server
3. Did you use the same drive  ? I mean Are these drives shared between the media servers ?

[Nicolai]

Imported media ?

Has the media been imported from another Netbackup domain ?


T/N : Step by step guide to importing a tape using the NetBackup 6.x GUI

[sdo]

Shyam, you say you checked the buffer sizes, I take it that you mean SIZE_DATA_BUFFERS.
1) Is SIZE_DATA_BUFFERS the same on all NetBackup servers in the environment?
2) Can you be sure that the SIZE_DATA_BUFFERS of your current environment is the same as when the tape was written c. three years ago?  (e.g. lots of people had 64KB buffers for LTO1, but now use 256KB for LTO3 - and reading LTO1 on LTO3 is supported - but it might not be working because now your env is a 256KB env, but the tapes are 64KB.

N.B. Be careful with SIZE_DATA_BUFFERS.  If you do try testing to read with a different buffer size, ie. if you modify SIZE_DATA_BUFFERS on one media server temporarily - then you need to make sure that no backups are written by that media server whilst you have a temporarily modifed SIZE_DATA_BUFFERS.  And if you do change it temporarily, don't forget to change it back to its original setting.

You can confirm your current buffer size on a master/media server by grep'ing the bptm log for "data buffer size" .e.g.:
$ grep -i "data buffer size" log.112809
09:37:19.136 [21186] <2> io_init: using 262144 data buffer size
09:40:31.426 [22114] <2> io_init: using 262144 data buffer size
...which is 256KB.

[faridux]

According to your different tests, you may have to test the tape outside netbackup by using an unix media server

1 mount the tape in a drive
2 position to the image (mt -f device_path fsf 547)
3 read the tape with tar tvf or a dd command

If it fails at point 2 --> you may have a db image inconsistency as even the OS does not locate this position
If it fails at point 3 --> the tape could be damaged

I hope it helps

[Nicolai]

You don't need to care about SCSI block size when restoring. Netbackup auto detect the blocksize.