Rebuild problem on NetRaid-1 si with LC2000

[José Caeiro]

Hi,

I've got on LC2000 server running Windows 2000 with a NetRaid-1 si card, I've got 3 18.2Gb disks on a Raid 5 configuration... 2 days ago one of those disks failed. at the same time there was a message in the system event viewer:

The device, \Device\Scsi\mraid35x1, did not respond within the timeout period.

I replaced the failed disk, but the rebuild failed somewhere around 60%...
I've updated the NetRaid firmware to F.02.05 version.
I also updated the BIOS on the LC2000 to 4.06.30 PV.
I've instaled the LC2000 windows 2000 drivers downloaded from HP.

Afterwards I tried, once more, to rebuild the failed drive but the rebuild process always fails around 60%..
There's the following message in the adapter log:

Rebuild Started on CHNL 0 TARG 1. - Fri Sep 07 23:07:00 2001
Error on Rebuilding CHNL 0 TARG 1 - Sat Sep 08 00:26:44 2001
Notify message : DRIVE STATE changed in Ch 0 ID 1 to FAILED - Sat Sep 08 00:26:50 2001
NOTIFY:Check Condition on Ch 0 ID 0 with the following sense key - Sat Sep 08 00:26:57 2001
f0 00 03 01 5e a6 b4
18 00 00 00 00 11 00


Any ideas?

P.S.: by the way, the drive that failed the first time wasn't out of order. It is working on another server...

[Paul R. Dittrich]

Bad cable / bad connector / bad drive cage

I'd start with the cable since it's the most easily damaged and the most easily replaced.

[Marino Meloni_1 · Accepted solution]

Hi
When a rebuild fail always at the same point, even after replacing a hdd, you have a problem on one of the other hdd.
this problem may be phisical or logical, if you check the log of the two disk and loot at media error, you could discover if one of these has a bad block, if all log are at 0 error than probably you have a logical problem.
anyway, the only think to do is a complete backup of your data and a reinizialisation of your array, after you can restore data.
to avoid this operation, it is recomanded to run every one ore two weeks the consistency check wich adjust these kind of error and avoid your situation in case a rebuild is need.
in fact in a rebuild task, the netraid controller check on the on-line disk the data and the local parity (the parity of the single block) if this is not valid, the controller cannot know wich data to record on the rebuilded disk and then fail the operation.

bye
Marino

[Tomas Lofgren]

Hi,
Marino is absolutely correct.The answer to why the rebuild fail is in your question (adapter log).There you see that CHNL 0 TARG 1 is the drive where the rebuild fails.If you look below you see that you have a Check condition on CH 0 ID 0 which is probably why the rebuild fail,you have bad blocks or parity error on that drive (which is online) and the only way now is to backup the data,reconfigure the raid and restore the data.To avoid this in the future see Marinos note about consistency check.
Just some more info.

[José Caeiro]

Thanks Marino and Soze... Unfortenatly you were right. Shortly after writting this message the ID 0 drive failed and, of course, the Logical drive failed. I still managed to backup most of the information a now the server is once again operational...
By the way: Is there any way to launch the consistency check on a bacth file? someway I could schedule it to run during the nigth?

[Marino Meloni_1]

Just instal vers 2.02 of netraid assistant, you will found a scheduled consistency check every two weeks at saturday midnight.
if you want to schedule with command line you need to download a file from hp wich let you schedule the task when you want,
http://www.hp.com/cgi-bin/cposupport/swindex.pl?p=hpnetraid-10478&l=en&c=Utility
bye
marino

[Dan Rod]

Hi,

I got the very same problem. The only difference is I am running WIN NT on my LC2000 server instead W2K. One of the disks failed, Tried to Reinitialize and failed at 14%. replaced it with a new disk, same result, failed at 14%. Tried replacing some pieces of hardware and nothing. the message:

Rebuild Started on CHNL 0 TARG 1. - Wed Apr 30 17:13:58 2003
Error on Rebuilding CHNL 0 TARG 1 - Wed Apr 30 17:16:10 2003
Notify message : DRIVE STATE changed in Ch 0 ID 1 to FAILED - Wed Apr 30 17:16:16 2003
NOTIFY:Check Condition on Ch 0 ID 0 with the following sense key - Wed Apr 30 17:16:21 2003

Time Stamp Date = On Apr,30 2003 At 17:16:4

The CDB = 28 00 00 50 bb 80 00 00 80 00

Sense Data = f0 00 03 03 50 bb 85 18 00 00 00 00 11 00 00 80 00 35

So, according to Meloni, Keyser and Local Support people We should format and rebuild the all array, because the OS could get faulty. I was looking for a different way to go, because the all redo process will take about 14 hours due to the apps we have to reinstall after the reinitialization. Any last minute advice?
By the way, where can I download the update for netraid assistant? Thanks in advance for all your help.