Koozali.org: home of the SME Server

Restore failure

Ed Form

Restore failure
« on: April 07, 2003, 06:20:04 PM »
I have a big problem with my server which will not restore from tapes for which perfectly good backup logs were obtained. I've appended the one from last Wednesday as an example.

When I try to restore I get...

|------------------------------------------------
Reading from current tape position
|------------------------------------------------
At block 1
|------------------------------------------------
buffer -m 3m -s 10k -u 100 -t -p 75 -B -i /dev/nst0 | gzip -dq | restore -t \
 -v -b 10 -f -
|------------------------------------------------
buffer (reader): failed to read input: Input/output error

gzip: stdin: unexpected end of file

restore: Tape read error on first record
Verify tape and initialize maps
Input is from file/pipe
|------------------------------------------------
At block 34.
|------------------------------------------------

I've searched extensively for comments on similar problems and tried the various suggestions for repeating commands in order to advance to the correct start block but to no avail.

The backup device is an HP Surestor DAT 24 in apparently perfect condition, although one suspects otherwise in these circumstances. Watching a backup take place shows a quite normal sequence of data light flashing as the backup proceeds, so eveything looks kosher, but no restores!!!!

Anyone have any suggestions?

Ed Form

***********************************
backup Log File for Wed Apr 2

***********************************
Viewed at Mon Apr 7 14:01:56 2003.
|------------------------------------------------
| Doing level 0 backup of all using dump
| Retensioning tape...
| Rewinding & erasing tape...
| Creating index key 200304030200.56
| Tape #0
| Filesystems = / /boot
|------------------------------------------------
At block 1.
|------------------------------------------------
| File number 1, index key 200304030200.56
| Backup of: /
| Date of this level 0 backup: Thu Apr 03 02:01:02 2003
| Date of last level 0 backup: the epoch
|------------------------------------------------
| (dump -0 -b 10 -a -f - / | gzip -4) | buffer -m 3m -s 10k -u 100 -t -p 75 -B \
|  -o /dev/nst0
|------------------------------------------------
  DUMP: Date of this level 0 dump: Thu Apr  3 02:01:03 2003
  DUMP: Dumping /dev/hda3 (/) to standard output
  DUMP: Added inode 8 to exclude list (journal inode)
  DUMP: Added inode 7 to exclude list (resize inode)
  DUMP: Label: /
  DUMP: mapping (Pass I) [regular files]
  DUMP: mapping (Pass II) [directories]
  DUMP: estimated 3568734 tape blocks.
  DUMP: Volume 1 started with block 1 at: Thu Apr  3 02:01:22 2003
  DUMP: dumping (Pass III) [directories]
  DUMP: dumping (Pass IV) [regular files]
  DUMP: 11.25% done at 1338 kB/s, finished in 0:39
  DUMP: 17.26% done at 1026 kB/s, finished in 0:47
  DUMP: 29.86% done at 1183 kB/s, finished in 0:35
  DUMP: 47.21% done at 1404 kB/s, finished in 0:22
  DUMP: 61.07% done at 1452 kB/s, finished in 0:15
  DUMP: 76.48% done at 1516 kB/s, finished in 0:09
  DUMP: 92.25% done at 1567 kB/s, finished in 0:02
  DUMP: 100.00% done at 1508 kB/s, finished in 0:00
  DUMP: Volume 1 completed at: Thu Apr  3 02:42:11 2003
  DUMP: Volume 1 3748010 tape blocks (3660.17MB)
  DUMP: Volume 1 took 0:40:49
  DUMP: Volume 1 transfer rate: 1530 kB/s
  DUMP: 3748010 tape blocks (3660.17MB)
  DUMP: finished in 2449 seconds, throughput 1530 kBytes/sec
  DUMP: Date of this level 0 dump: Thu Apr  3 02:01:03 2003
  DUMP: Date this dump completed:  Thu Apr  3 02:42:11 2003
  DUMP: Average transfer rate: 1530 kB/s
  DUMP: DUMP IS DONE
Kilobytes Out 1117340
|------------------------------------------------
| Backup start: Thu Apr 03 02:01:02 2003
| Backup end:   Thu Apr 03 02:42:16 2003
|------------------------------------------------
At block 111735.
|------------------------------------------------
| File number 2, index key 200304030200.56
| Backup of: /boot
| Date of this level 0 backup: Thu Apr 03 02:42:27 2003
| Date of last level 0 backup: the epoch
|------------------------------------------------
| (dump -0 -b 10 -a -f - /boot | gzip -4) | buffer -m 3m -s 10k -u 100 -t -p \
|  75 -B -o /dev/nst0
|------------------------------------------------
  DUMP: Date of this level 0 dump: Thu Apr  3 02:42:27 2003
  DUMP: Dumping /dev/hda1 (/boot) to standard output
  DUMP: Added inode 8 to exclude list (journal inode)
  DUMP: Added inode 7 to exclude list (resize inode)
  DUMP: Label: /boot1
  DUMP: mapping (Pass I) [regular files]
  DUMP: mapping (Pass II) [directories]
  DUMP: estimated 4731 tape blocks.
  DUMP: Volume 1 started with block 1 at: Thu Apr  3 02:42:30 2003
  DUMP: dumping (Pass III) [directories]
  DUMP: dumping (Pass IV) [regular files]
  DUMP: Volume 1 completed at: Thu Apr  3 02:42:33 2003
  DUMP: Volume 1 4720 tape blocks (4.61MB)
  DUMP: Volume 1 took 0:00:03
  DUMP: Volume 1 transfer rate: 1573 kB/s
  DUMP: 4720 tape blocks (4.61MB)
  DUMP: finished in 3 seconds, throughput 1573 kBytes/sec
  DUMP: Date of this level 0 dump: Thu Apr  3 02:42:27 2003
  DUMP: Date this dump completed:  Thu Apr  3 02:42:33 2003
  DUMP: Average transfer rate: 1573 kB/s
  DUMP: DUMP IS DONE
Kilobytes Out 2600
|------------------------------------------------
| Backup start: Thu Apr 03 02:42:27 2003
| Backup end:   Thu Apr 03 02:42:39 2003
|------------------------------------------------
At block 111995.
|------------------------------------------------
| Rewinding...

Brian Read

Re: Restore failure
« Reply #1 on: April 07, 2003, 06:45:45 PM »
Ed

Not sure if this will help, but this is my shell script which I have sucessfully used to restore files and directories from tape in the past.

---------------------------------------------------------------------------------------------
# Restore script
# Put file names in /root/flexbackup/extract-list

cd /root
mkdir restore-list
cd restore-list
/bin/mt -f "/dev/nst0" rewind
flexbackup -list
/usr/bin/flexbackup -extract -files /root/flexbackup/extract-list
/bin/mt -f "/dev/nst0" rewind
---------------------------------------------------------------------------------------------

cheers

Brian

Ed Form

Re: Restore failure
« Reply #2 on: April 07, 2003, 07:45:14 PM »
Brian  Read wrote:

> Not sure if this will help, but this is my shell script which
> I have sucessfully used to restore files and directories from
> tape in the past.
>
> ---------------------------------------------------------------------------------------------
> # Restore script
> # Put file names in /root/flexbackup/extract-list... Snip

Thanks Brian,

I had already found the thread where you discussed this in the past and tried it out. It didn't work.

I'm taking the tapes to another colleague's machine later to check them out, but if anyone else has comments I would really be grateful.

Ed Form

Steve Bush

Re: Restore failure
« Reply #3 on: April 07, 2003, 09:44:36 PM »
I am using, and have used HP's DAT24 drive.  I have been able to backup and restore multiple times.  Initially I had a problem with an older version of SME that was related to the block size.
I had to fiddle with my restore script, setting the block size manually to get it working correctly.
Later versions, I changed my block size to 32k in /etc/flexbackup.conf, which helped speed up my backups.

Your backup and restore logs block sizes appear to match, so I don't know if that's the issue or not.

I also notice that your backup log shows it starting at block 1, mine starts at block 2.

Des Dougan

Re: Restore failure
« Reply #4 on: April 08, 2003, 06:18:14 AM »
Ed,

My restore script (which I've used and tested previously...) has an extra command:

[root@jeeves /root]# more flexbackup_restore
/bin/mt -f /dev/nst0 rewind
/bin/mt -f /dev/nst0 fsf 1
/usr/bin/flexbackup -extract -files /root/extract-list
/bin/mt -f /dev/nst0 rewind

I believe what you are seeing in your restore is that you are looking in the tape header for your data, hence the "fsf 1" (forward skip file) command to get to the datafile.

Good luck!

Des Dougan

Ed Form

Re: Restore failure
« Reply #5 on: April 09, 2003, 12:47:03 AM »
In article , ddougan@telus.net (Des Dougan) wrote:

> My restore script (which I've used and tested previously...) has an
> extra command:
>
>
> [root@jeeves /root]# more flexbackup_restore
>
> /bin/mt -f /dev/nst0 rewind
>
> /bin/mt -f /dev/nst0 fsf 1....    Snip

Thanks Des and everyone else who replied. None of these ideas worked, but the solution was, fortunately, straightforward.

The server in question is still running the last 5.6 beta with all of the updates that were made available for it and there is a bug in the Flexbackup configuration of that version. I ran the tape in an identical HP drive on a 5.6 Update 2 machine and was able to restore the backup without problems. I'll bring the machine with the
problems up to the latest spec this weekend.

Ed Form