permanent AutoRAID disk errors

[Vassily Gorbounov]

Dear colleagues,

May be you could help me cause nobody from Russian Hewlett can resolv this problem.

May be you've already encountered such situation or just know how to deal with it.

I have AutoRAID 12H in the following configuration:
1. 8 disks by 9 GB (slots A1-B4) (from the start - approx. 3 years)
2. 4 disks by 18 GB (slots A5-B6)(installed approx. a year ago)
3. Dual-Controller 96MB with firmware version 62H (upgraded approx. half of the year ago)
4. Firmware on all drives upgraded to the latest version approx. half of the year ago.

Problem:
ALL drives report about huge amount of read errors. Newly inserted drives (18GB or replaced 9GB drive) immediately started to report about errors. I've attached logprint command output for one of the drives as an example.

Actions taken:
I've upgraded all that's possible - firmware for drives, firmware for controllers, patched the OS, upgraded os from 10.20 to 11.00 (disk errors isn't the reason for upgrade), replaced one of the controllers and switched primary to it - no success. Errors continue to appear at approx 1000 a day for the drive (vary by drive).
HP serviceman said that he haven't seen nothing similar in other customer installations.

I would be very grateful if you would give me a hint on this situation. What's wrong - SCSI config, main host SCSI controllers, autoraid backplain.

May be you know something I could try to localize the problem?

With Best Regards,
Vassily Gorbounov
Unilever CIS

[harry d brown jr]

Have you tried replacing the scsi cables, and scsi cards on your host? Also, check the cables, and disk drive bays for bent pins. You will need a very good flashlight. Also, have you replaced the power supplies?

If all of this fails to fix the problem, tell Hp to replace the ENTIRE unit, as it could be a backplane issue.

live free or die
harry

[Insu Kim]

It could be a midplane in the AutoRAID or HBAs on host side.

By the way,
I wonder whether you update firmware to the latest, HP62 with the required AutoRAID patch.
One of the features is enhanced logging facility that is very useful for troubleshooting.

In the document,
HP62 adds logging for data unavailable events. If the array detects one or more blocks of data that have become unavailable due to multiple failures the location will now be logged. This will provide the ability to pinpoint the location of the lost data so that just the affected file or files may be restored.

Please be aware that logprint is much better than any other commands supplied for AutoRAID maintenance as far as troubleshooting concerned.

# logprint -a

And finally, logtool in STM (Support Tools Manager) will be useful as well.
Try to see if there is anything producing enormous errors.
If something associated with AutoRAID is there, you should think why it's getting them.

Hope this helps,