Topic: tape backup error

[henry44]

I searched the forums with no joy.
I am getting this error message with every run of the backup to tape routine.

buffer (writer): write of data failed: Input/output error
bytes to write=32768, bytes written=-1, total written    3190208K

flexbackup version 1.2.1 (http://flexbackup.sourceforge.net)
/etc/flexbackup.conf syntax OK

 $keyfile not found in config: default=00-index-key
 $buffer_fill_pct not found in config: default=75
 $buffer_pause_usec not found in config: default=100
|------------------------------------------------------------
| Checking 'buffer' on this machine... Ok
| Checking /bin/sh on this machine... bash1
|------------------------------------------------------------
| Doing level 0 backup of set all using tar
| All sets = full (1 tapes)
| Rewinding & erasing tape...
| Found index key 201005070100, next file is 2
| Deleting record for 201005070100 file 0
| Deleting record for 201005070100 file 1
|------------------------------------------------------------
| Creating index key 201005080100
| Backup set "full" (/)
| Tape #0
|------------------------------------------------------------
SCSI 2 tape drive:
File number=1, block number=0, partition=0.
Tape block size 0 bytes. Density code 0x19 (DLT 10GB).
Soft error count since last status=0
General status bits on (81010000):
 EOF ONLINE IM_REP_EN
|------------------------------------------------------------
| File number 1, tape index 201005080100
| Backup of: /
| Date of this level 0 backup: Sat May 08 01:00:14 2010
| Date of last level 0 backup: the epoch
|------------------------------------------------------------
| cd "/" && /tmp/collectexit.14103.sh find . -depth -xdev ! -type s ! \
|   -regex "news/articles" ! -regex ".*~"$ -print0 | \
|   /tmp/collectexit.14103.sh tar --create --null --files-from=- \
|   --ignore-failed-read --same-permissions --no-recursion --totals --label \
|   "level 0 / Sat May 08 01:00:14 2010 tar+gzip from larry-server" \
|   --sparse --atime-preserve -b 64 --file - | /tmp/collectexit.14103.sh \
|   gzip -4 | /tmp/collectexit.14103.sh buffer -m 20m -p 75 -s 32k -t -u \
|   100 -B -o "/dev/nst0"
| [ ! -e /tmp/exitstatus.14103 ]
|------------------------------------------------------------
buffer (writer): write of data failed: Input/output error
bytes to write=32768, bytes written=-1, total written    3190208K

ERROR: non-zero exit from:
buffer -m 20m -p 75 -s 32k -t -u 100 -B -o /dev/nst0
gzip -4
tar --create --null --files-from=- --ignore-failed-read --same-permissions --no-recursion --totals --label level 0 / Sat May 08 01:00:14 2010 tar+gzip from larry-server --sparse --atime-preserve -b 64 --file -
find . -depth -xdev ! -type s ! -regex news/articles ! -regex .*~$ -print0


ERROR: exiting
|------------------------------------------------------------
| Backup start: Sat May 08 01:00:14 2010
| Backup end:   Sat May 08 04:10:28 2010
|------------------------------------------------------------
SCSI 2 tape drive:
File number=1, block number=99694, partition=0.
Tape block size 0 bytes. Density code 0x19 (DLT 10GB).
Soft error count since last status=0
General status bits on (21010000):
 EOT ONLINE IM_REP_EN
|------------------------------------------------------------
| Rewinding...
| Compressing log (all.0.201005080100.gz)
| Linking all.latest.gz -> all.0.201005080100.gz
|------------------------------------------------------------
Backup terminated: backup failed - status: 256

I ran a cleaning tape and changed tapes with the hope that the problem is in the device, but that did not fix things. I have no experience with debugging this kind of error, but I am fine with following directions. If there is someone that can translate the error message into a way to fix the issue, I would be grateful.

Thanks..........

Henry

[pearless]

I've have these errors in the past. 
The message is rather generic, so things to check:

(1) Are you exceeding the capacity of the tape?
(2) Have you run diagnostics on the tape drive itself?
(3) Is there enough room on your disk for mbuffer to use -> try disabling buffering to see if that helps.

Cheers
Douglas.

[henry44]

Thanks for your reply.

I have not run any tests on the drive itself. I think that the amount of data being backed up may have exceeded the capacity of the drive. I have a larger capacity drive on the way and will try that drive. I have about 60gb of available space on the drives.

Henry

[smeghead]

.. is the tape write protected?

[henry44]

Nope, I checked. It is a case of the common "it worked, then it didn't" syndrome.

Henry

[CharlieBrady]

I've have these errors in the past. 
The message is rather generic,

No it isn't. It's quite specific for Input/Output errors.

so things to check:

(1) Are you exceeding the capacity of the tape?

That will generate a specific error message about "no space left on device".

(2) Have you run diagnostics on the tape drive itself?
(3) Is there enough room on your disk for mbuffer to use -> try disabling buffering to see if that helps.

mbuffer does not use space on the disk.

I'm fairly sure that the error message is specific to a problem involving tape and/or tape drive.

[henry44]

So it looks like I should check for something gone wrong physically with either the drive or the tapes?

I get the same error message with both of the tapes I use, so this leads me to think the drive has developed some kind of mechanical problem. Would this be a reasonable assumption? If so, I can work with this idea to see what might have gone wrong. I have had drive belt issues in the past, but I didn't see this error when those problems developed. When the drive belt failed, the drive just stopped working. Maybe the belt is slipping enough to prevent the drive from functioning properly. I'll give that idea a look.

Thanks...............

Henry

[henry44]

I solved the error problem by installing another tape drive. I rebooted after the hardware install and everything seemed fine. The back-up routine seemed to run without errors and reported the appropriate amount of data backed-up. However, the drive is identified as the old drive with the old drive tape capacity. How do I get the server to correctly identify the new hardware and tape capacity?

[cactus]

I solved the error problem by installing another tape drive. I rebooted after the hardware install and everything seemed fine. The back-up routine seemed to run without errors and reported the appropriate amount of data backed-up. However, the drive is identified as the old drive with the old drive tape capacity. How do I get the server to correctly identify the new hardware and tape capacity?

Please log it as a bug in the bugtracker as IMHO this should just work. Please post back a reference to the bug for future readers here.

[henry44]

Thanks. I'll do that.

Henry

[CharlieBrady]

However, the drive is identified as the old drive with the old drive tape capacity.

Where do you see the drive and its capacity identified?

[henry44]

I see it in the e-mail I get after the back-up routine runs.

Here is the snippet I am referring to:

SCSI 2 tape drive:
File number=1, block number=0, partition=0.
Tape block size 0 bytes. Density code 0x81 (DLT 15GB compressed).
Soft error count since last status=0
General status bits on (81010000):
 EOF ONLINE IM_REP_E

Line 3. The part right after Density code 0x81.

The tape drive is a Exabyte VXA-2 drive (SCSI LVD) with an 80/160gb capacity. I have  230m tape in the drive (80gb native/160gb compressed).

While I didn't try to restore from tape, the e-mail notification reports the correct amount of data backed up with no errors. Therefore I assumed the data was successfully archived. Is this reporting error a problem?

Thanks............
Henry

[henry44]

Should I file a bug report? What I think I need to know is how to change the density code for the tape drive. That would fix the problem.

Henry

[CharlieBrady]

What I think I need to know is how to change the density code for the tape drive. That would fix the problem.

'mt densities' will show you the available density settings.

'mt -f /dev/st0 setdensity 0xnn' and 'mt -f /dev/st0 defdensity 0xnn' set the density code.

[henry44]

Great, thanks for the help. I have more question that should bring this to issue to a close.

Would you like me to submit a bug report?

Thanks for your help.

Henry