Phase 2 import not finding images from BackupExec import

[JasonP_76]

I am running NetBackup 7.5.0.4 and I have completed a phase 1 import of a whole lot of BackupExec Tapes (35 tapes in total) using :

vmphyinv.exe -rn 0 -slot_range 1 45 -d hcart2

It completes the physical inventory and I have done an intiate import for each of the tapes that were inventoried.

If I run a report of "Images on Media" I get a whole lot back.

In the C:\Program Files\Veritas\NetBackup\db\images\ there are a whole lot of directories with the server names, though inside each of those directories are a whole lot of lck files that are 0kb in size.

If I try run an "Import" from the catalog tree (Phase 2) I get the error :

INF - Found no images matching the selection criteria that were ready for phase 2 import.

Has anyone come across this or can give me an idea what is going on.

Thanks

[Mark_Solutions]

what version of backup exec were the tapes taken from?

There are a lot of restirctions - check the Admin Guide 1 for full details incase there is any incompatibility - Page 794 onwards in my copy

[CRZ]

..or in this TechNote:

DOCUMENTATION: Statement of support for the importing of Backup Exec images in NetBackup 7.x using the Backup Exec Tape Reader
 http://symantec.com/docs/TECH125836

[JasonP_76]

The version of Backup exec was 12. So looking at the docs it definitely looks like it is compatible.

The bpimport error is :

INF - Importing policy incremental, schedule win_full (lpkcochg01.old.local_1298579953), media id BE0027, created on 02/24/2011 20:39:13.INF - from host bkp.local, read invalid database name. database name size = 0x0, database name offset = 0x0c

INF - Import of policy incremental, schedule win_full (lpkcochg01.old.local_1298579953) failed, tar had an unexpected error.

[Mark_Solutions]

I did see one like this a while ago and seem to remember that it had been a glitch during the phase 1

On that occasion they cleared the temporary catalog files for that client and started again after which it worked

There can also be tape drive compatibility issues sometime (written to HP, read by IBM)

[JasonP_76]

How do I clear the temporary catalog files ? 

Do I have to then use the vmphyinv again ?

I am using the exact same library that was previously used to create the tapes, so there should be no hardware incompatibility.

[Mark_Solutions]

if that client(s) have never been backed up by NetBackup then under netbackup\db\images the only thing in that clients folders should be the phase 1 import data - do check carefully though

The vmphysinv is not needed as the NBU volume database is up to date for that part - just the Phase 1 again

I assume that you definately do have all of the tapes involved? If any fragments were missing a phase 2 could not work

[JasonP_76]

Ok so you mean delete all folders in there ? and remove all tapes from the library and re-inventory it ?

I initially used vmphysinv which is where the phase 1 details came from ? 

should I have used the bpimport -create_db_info ?

[Mark_Solutions]

Only delete the folder for the client to be restored (lpkcochg01.old.local) - and only if it has never been seen by NetBackup before otherwise you will delete its other known backups.

You do not need to remove the tapes - just do a phase 1 for the Backup Exec tapes via the catalog section of the admin console

[JasonP_76]

To add to this I ran bpimport and it is starting to do phase 2 but it keeps erroring with :

28/11/2012 16:05:19 - Info mtfrd(pid=3916) Restore started           

28/11/2012 16:05:19 - Info bptm(pid=2916) start            

28/11/2012 16:05:19 - started process bptm (2916)

28/11/2012 16:05:19 - Info bpdm(pid=2916) reading backup image          

28/11/2012 16:05:19 - Info bptm(pid=2916) using 30 data buffers         

28/11/2012 16:05:19 - Info bptm(pid=2916) spawning a child process         

28/11/2012 16:05:19 - Info bptm(pid=2916) child pid: 4328          

28/11/2012 16:05:19 - Info bptm(pid=4328) start            

28/11/2012 16:05:19 - started process bptm (4328)

28/11/2012 16:05:19 - requesting resource BE0027

28/11/2012 16:05:19 - granted resource BE0027

28/11/2012 16:05:19 - granted resource IBM.ULT3580-HH5.001

28/11/2012 16:05:20 - mounted

28/11/2012 16:05:20 - Info bptm(pid=2916) BE0027            

28/11/2012 16:05:20 - Info bptm(pid=2916) INF - Waiting for positioning of media id BE0027 on server scobkp.local for reading.

28/11/2012 16:05:20 - positioning BE0027 to file 23

28/11/2012 16:05:31 - positioned BE0027; position time: 00:00:11

28/11/2012 16:05:31 - begin reading

28/11/2012 16:05:31 - Error bpbrm(pid=1016) from client bkp.local: read invalid database name. database name size = 0x0, database name offset = 0x0

28/11/2012 16:05:31 - Error bpbrm(pid=1016) from client bkp.local: block type = dirb, common block attributes = 0x20080, file system type = 0xe, file system version = 0x3, logical block address (low) = 0x40, logical block address (high) = 0x0, control block identifier = 0x2, string type = 0x2

28/11/2012 16:05:31 - Info mtfrd(pid=3916) done. status: 182          

28/11/2012 16:05:31 - Error bpimport(pid=3148) from host bkp.local, read invalid database name. database name size = 0x0, database name offset = 0x0

28/11/2012 16:05:31 - Error bpimport(pid=3148) Import of policy incremental, schedule win_full (lpkcochg01.uk.chi-x.com_1298579953) failed, tar had an unexpected error.

28/11/2012 16:05:32 - Info bpbrm(pid=632) bkp.local is the host to restore to      

28/11/2012 16:05:32 - Info bpbrm(pid=632) reading file list from client        

28/11/2012 16:05:32 - Info bpbrm(pid=632) starting mtfrd on client         

28/11/2012 16:05:34 - Info bptm(pid=2916) EXITING with status 24 <----------

[JasonP_76]

I have started all from the beginning again as none of the tapes were importing and giving errors. All basically media read errors, which I find hard to believe seeing that these tapes have worked perfectly fine in the Backup Exec System a few weeks ago. 

I had to do a vmphyinv first as the the tape GUID's would be different from just a normal inventory. Then I am doing the phase 1 import from the catalog section as suggested. All the backups are password protected, and the password seems to be fine since it seems to be completing the phase 1 imports on the tapes without any failures.

When I do a phase 2 import is there anything I need to make sure of if these backups are password protected ? Do I have to give NetBackup the password again for the phase 2 imports ?

Thanks

[JasonP_76]

Ok after all of that it is still not working.

I have gone through the bptm log file and I can see this :

09:06:34.799 [6428.5692] <2> read_position: read position on drive index 0, is LBA 32688802

09:06:34.815 [6428.5692] <2> read_data: ReadFile returned FALSE, The media is write protected. (19);bytes = 0

09:06:34.815 [6428.5692] <2> is_possible_recoverable_error: not attempting error recovery, errno = 19

09:06:34.815 [6428.5692] <16> read_data: cannot read image from media id BE0033, drive index 0, err = 19

09:06:34.815 [6428.5692] <2> send_MDS_msg: DEVICE_STATUS 1 122 scobkp.uk.bats.com BE0033 4000122 IBM.ULT3580-HH5.000 2000005 READ_ERROR 0 0

09:06:34.815 [6428.5692] <2> log_media_error: successfully wrote to error file - 12/01/12 09:06:34 BE0033 0 READ_ERROR IBM.ULT3580-HH5.000

These are WORM tapes. Is there an issue in importing WORM tapes ? Is there a way to override this ?

[JasonP_76]

Now I think I have got a little further.

I think there is an issue with how NetBackup imports BackupExec Data from WORM tapes.

In the db/images directory under each of the servers that have been done under phase 2, there is a tmp directory. In that tmp directory is ALL the catalog files.

For some reason NetBackup does not add it to its database to complete the import for some reason.

Can someone please tell me how I will be able to import these individual catalog files ? Or moving them somewhere and firing off a command for NetBackup to update its DB ?

Thanks ....

Probably worth Symantec raising a bug report for this if they can.

[Marianne]

Have you logged a Support call with Symantec?

Please let us know what the outcome was...

[JasonP_76]

Hi Marianne,

I did end up raising a support call with symantec.

What came out was that actually NetBackup does NOT support encrypted tapes written in Backup Exec, which is really frustrating because I don't understand why the import process asks for the password initially used to create those encrypted backups in backup exec in the first place.

So now we have to restore the data to another location and re-back them up within netbackup.

Really poor on Symantec side to not have a way to export the key from backup exec and use it in NetBackup KMS if that is the case.

[CRZ · Accepted solution]

What came out was that actually NetBackup does NOT support encrypted tapes written in Backup Exec,

For what it's worth (and I'm sure you don't want to hear this and I apologise in advance), this was spelled out pretty plainly in the bullet list of what ISN'T supported in the previously linked Statement of Support:

Encrypted images are not supported

DOCUMENTATION: Statement of support for the importing of Backup Exec images in NetBackup 7.x using the Backup Exec Tape Reader
 http://symantec.com/docs/TECH125836

Unfortunately I don't think we were let on that these were encrypted images anywhere in this thread, so we weren't able to save you a couple months.  Sorry again!