Start a Conversation

Unsolved

This post is more than 5 years old

1047

December 28th, 2005 08:00

Cannot do a saveset recover

Networker 7.1.3 running on a HP-UX box.

My customer has a backup that passed their browse policy, and now want to restore that data. Well, we tried to do a saveset restore, as we could not do a browseable restore anymore. The system says that data was restored, but there's nothing there, only the root folder.

We took a look at the saveset in the volume, and it shows the saveset only with 2 files, and some few Kb in size, but if I look at the Backup Completion Report, it was 200Gb.

Then we run a scanner -m for the saveset, and in verbose mode it showed all the files we wanted. But then we tried the restore, but had the same problem.

We're now trying to run scanner with the uasm, but I don't know if it will work. And as it will take some time to complete, I would like to know if anyone has another idea.

Regards,

2 Intern

 • 

14.3K Posts

December 28th, 2005 09:00

Why don't you restore index file with nsrck command? Or simply run scanner -i against ssid on that tape.

Since you know ssid do something like:
mminfo -av -q ssid=$SSID -r name,client,ssflags,sumflags,nfiles,totalsize,sumsize

That should tell you the size and number of files inside.

With saveset recover when restoring file or directory you MUST use path exactly the same as it was used during the backup as it is case sensitive - this means that if you type one letter wrong (FoR ExaMplE C:\Temp instead of C:\TEMP or C:\temp then it won't work).

If you have space, restore whole ssid to some destination over the night and pick the file you need next morning if in hurry.

724 Posts

December 28th, 2005 09:00

That's the problem, the mminfo output is like:

ssid=1031450341 savetime=11/16/05 02:03:51 (1132113831) karina:/fila_1
level=full sflags=vF size=1664 files=2 insert=11/16/05
create=11/16/05 complete=11/16/05 browse=05/12/06 00:00:00 retent=05/12/06 00:00:00
clientid=f68d75b1-00000004-41f68f85-41f7c7c2-00230000-8a6ae64a
group: File_System;
Clone #1: cloneid=1132113831 time=11/16/05 02:03:51 flags=F
frag@ 0 volid=2625202496 file/rec= 68/10258 rn=0 last=12/26/05


As you can see, the files and size do not correspond to the backup, and also to the output of the scanner. I need the whole saveset, and that's why I'm trying the saveset restore, it should be faster then run 'scanner -i' and then do a browsable recovery.

Well, we're trying something like "scanner -S -x uasm -rv" to see if anything works, but I'm not so sure about this.

2 Intern

 • 

14.3K Posts

December 28th, 2005 15:00

Hm, according to above your saveset should be browsable. Do the query for the saveset again with mminfo and in report put nsavetime then use that value with nsrinfo for the client and see for files does it report (as it should be browsable until 12th May 2006 - same as retention).

724 Posts

December 29th, 2005 03:00

Another info I did not mention, that saveset had expired, but then the customer changed their browse and retention to use them to do browsable recover, but then it just showed nothing.

Well, over the night, we tried this:

- Removed the volume from media database
- Run "nsrim -X" and "nsrck -L3" (only to be sure that no more info about that tape exists).
- Did a "scanner -i" for the volume.

The volume and its data was rebuild, but the saveset we wanted is still with only 2k in size.

I don't know in which information to believe: the saveset completion report and the scanner output shows a lot of files in the saveset. But the volumes and the mminfo shows nothing.

I did hope that removing/rebuilding the volume in the database would work.

2 Intern

 • 

14.3K Posts

December 29th, 2005 04:00

Strange... never heard something like that, but if scanner is showing files (for that saveset) then data should be available. I would suggest to open this as a call with support and for your sake to run something like:
scanner -s save_set_id device_name | uasm -rv

Note that will recover your data from client to your server. You can do the same to your client and more specific info about scanner and uasm joint usage can be found in old bulletin 135 (somewhere at the bottom):
http://www.legato.com/support/websupport/tech_bulletins/index.htm?includefile=135.html

Bare in mind that if you recover this to server you must have same file system type as on client - otherwise it may fail (so no UNIX to Windows, no NTFS to FAT and things like that).

724 Posts

December 29th, 2005 11:00

We tried the scanner and uasm, but we always receive the "invalid save stream" error, enve using -x, as suggested somewhere in the KB.

I sent the case to Legato and will wait for their opinion.

2 Intern

 • 

14.3K Posts

December 29th, 2005 15:00

Sounds like corrupted stream which has been saved. However, perhaps it is related to immediate save technology. Try following:
- touch /nsr/debug/noimmediate (on backup server and storage node if using something else than server)
- restart daemons
- try restore again

If that won't change anything then I would really start to worry.

724 Posts

December 30th, 2005 04:00

Well, I sent the case to Legato, and they requested me a LOT of log files and command outputs, so I think that I need to worry, as it not seems to be easily solved.

I'll update this case when Legato give some info on that.
No Events found!

Top