NetRaid Predictive failure

[Robert Robbins_1]

I recently installed NetRaid Assistant on a Netserver LC II, NT4 Server sp6. The server has three 9.1 Gb drives configured as one Raid 5 logical drive.
The server has been running in production for several years but it did not have the NetRaid assistant until now. The NetRaid Assistant logs immediately started reporting the following error:
NOTIFY:Check Condition on Ch 0 ID 0 with the following sense key - Wed Jun 30 13:31:42 2004
70 00 01 00 00 00 00
0a 00 00 00 00 5d 00
Predictive failure detected : Ch 0 : Id 0 - Wed Jun 30 13:31:42 2004

I have also saw this error once:
Predictive failure detected : Ch 0 : Id 0 - Thu Jul 01 10:35:01 2004
Notify message : Predictive failure detected - Thu Jul 01 10:35:01 2004
Battery backup:Module missing - Thu Jul 01 11:01:42 2004

I replaced the drive and did a Rebuild and Rescan but the error is still being reported. A Consistency check before and after I replaced the drive completed OK.

The server itself seems to be running OK. The system event logs does show that 6 days ago
"The device \Device|ScsiPort1, did not respond within the timeout period" and ScsiPort1 is the NetRaid adapter. Otherwise, all seems normal.

Does anyone know what the NetRaid Assistant errors I am seeing mean?

Thanks in advance

[Darren Webb]

If the server has been in production for a while do you update the drivers / bios to the latest versions?
http://forums1.itrc.hp.com/service/forums/questionanswer.do?admit=716493758+1089129089369+28353475&threadId=100515

[kris rombauts]

Hi Robert,

the disk that was at SCSI id=0 indeed once has send this 'predictive failure' because it probably crossed it's threshold of bad blocks it encountred on the media as per the disk drive manufacturer's specifications
(not Netraid spec's)

There is a 'limitation' of the Netraid Assistant in a way that it will read the Netraid's error buffer every time you start the Netraid Assistant and each time you will see the error added again to the log file, even if you replace the disk with a good one. There is a way to clear it but without a full backup i would not do it since it concerns clearing the Netraid config etc etc.

So if you replaced that disk at ID=0 which was probably bad after all and which would likely fail in the future, then you should be in good shape.

Ignore the message untill you see a additional one on ID=0 or a different type of message from the same disk or from different SCSI id of course.

If you install the Netraid monitor service then you will get notified in the Windows eventlog if disk problems or array problems occur, so keep an eye on that one.

It's always a good idea to schedule the weekly consistency check (via the Netraid service) so that your raid5 health check is done on a weekely basis, it will then try to correct any issue and reconstruct data before you hit the bad spot.

Netraid Assistant and Netraid monitor service are both available on the public HP web pages.
Take a look at the readme on how to enable the consistency check on a day/time that best suites you.


The timeout error in the Systemlog from the Netraid driver can occur once in a while when very busy and is not directly related with the above errors unless they occur frequently and Netraid logs also shows errors, if not don't worry.

HTH

Kris

[Robert Robbins_1]

Thank you, I think you just saved me from spending a lot of time chasing a non-existent problem.