Failure of ontape Archive.

[LogicallyLJ]

Currently we are getting an archecker failure on restores taken on a clients major system.

However it shows that the backup was successful in the logs. They are using ontape with an Ultrium drive.
They are running HP11.11 with Informix 9.4 HC4.

They have sent us the Ultrium tape to attempt a restore here.

We are aware that there is a problem with archecker
validation, but if this was the problem we should be able to restore the tape on a different system here. We cannot.

The onconfig settings are the same and these have been ruled out. The resotre is attempted using an ontape -r command.

It progresses successfully to the point where it asks if you would like ot continue, y is selected, it then askes if you would liek to backup logical logs, no is selected, program then ends blaming corruption, as shown below.

Continue restore? (y/n)y
Do you want to back up the logs? (y/n)n
Physical restore failed - The archive backup is corrupted.


Program over.


At an equivalent time we get the folloing error in the syslog.

Jan 25 18:41:43 snss1a vmunix: SCSI TAPE: dev = 0xcd002040 I/O error during close

I would like to rule out a tape device problem or tape problem, mainly to get IBM support to actually focus on the problem, if there is one.

Any help much appreciated.

The device can be written to and read from.

Please find attached the complete output from the failed restore along with the output from the failed archecker.

Thanks

LJ

[Steve Lewis]

Its a known defect, see
http://www-1.ibm.com/support/docview.wss?rs=630&context=SSGU8G&dc=DB520&dc=D600&dc=DB530&dc=D700&dc=DB500&dc=DB540&dc=DB510&dc=DB550&q1=The+archive+backup+is+corrupted&uid=swg21208667&loc=en_US&cs=utf-8〈=en

Or search for the error message
"The archive backup is corrupted"
at
http://www-306.ibm.com/software/data/informix/ids/support/

where it gives a workaround, which is to use a larger block size in your backup.

I don't know if they can get the data off the tape - maybe using a couple of dds you might get something which you can write to another tape, but I doubt it.
Anyway congratulate your customer for actually trying to restore its tapes once in a while. I hope they don't need this one for real.

[LogicallyLJ]

Hello.

thanks for your help, however, IBM do not feel this is a related issue.

1) Apparently the bug mentioned is fixed in IDS p.4 HC4, which we are running.
2) It applies when there are over 50 chunks to a dbspace, we have a maximum of 32 in a dbspace.
3) It was in reference to TAPLEBLK sizes of 16. Ours are set to 512, so should not suffer from this problem.

Does anyone else have any further ideas?

Kind regards
LJ