help with raid array drive replacement lp200r

[ashley mallory]

Hello,

I have a failed drive in a raid5 array. the server is a lp2000r , the controller is HP NetRaid 1M

I have tried replacing the failed drive with two other new drives but the rebuild fails.

The replacement drives match the exact size of the failed drive.


Any suggestions?


This is from the log:


The CDB = 28 00 00 56 d7 80 00 00 80 00

Sense Data = f0 00 03 03 56 d7 84 0a 00 00 00 00 11 00 e4 80 00 82
Rebuild Started on CHNL 0 TARG 3. - Fri Dec 28 10:09:58 2007
Error on Rebuilding CHNL 0 TARG 3 - Fri Dec 28 10:12:44 2007
Notify message : DRIVE STATE changed in Ch 0 ID 3 to FAILED - Fri Dec 28 10:12:48 2007
NOTIFY:Check Condition on Ch 0 ID 2 with the following sense key - Fri Dec 28 10:12:53 2007

Time Stamp Date = On Dec,28 2007 At 10:9:20

The CDB = 28 00 00 56 d7 80 00 00 80 00

Sense Data = f0 00 03 03 56 d7 84 0a 00 00 00 00 11 00 e4 80 00 82

[e4services · Accepted solution]

From what I get, the server is still running, yes?
So the only recovery from these data errors in the degraded array is to backup the entire array, delete the array and re-configure it and then restore. The error is usually because there is some data error in the array. Verifying the data on a regular basis can avoid such a situation, but unfortunately you can not start a verification when the array is running in a degraded mode, but you can not get back to because you can not rebuild.

[ashley mallory]

thanks for the reply... although I'm not very happy about the answer

the server is still up and running... just degraded.

how can I avoid this in the future?

[e4services]

Check Consistency. This check the consistency of the data (the paired bits store on at least 2 drives) to make sure everything is the same. Checking the NetRAID manual, "Verifies the correctness of the redundancy data in the selected logical drive; this option is only available if RAID level 1, 3, or 5 is used. Check consistency should be run if the system shuts down irregularly, e.g., if the system hangs or suffers a power failure. It is also recommended to run consistency checks every 2-4 weeks to ensure that bad blocks on the disk drive are mapped out. (See the â Check Consistency Feature.â ) Check
consistency will always remedy inconsistencies. Check consistency will only fail because firmware could not complete parity and data writes due to a power-failure or a system hang, or because the logical drive degraded or failed due to the physical drives going offline."
Why you can not run it in the situation you are in.

Using the NetRAID Assistant you can setup the NetRAID monitoring tool to load in as a Windows service and you can schedule the Consistancy Check at a set interval, "megactrl -start" command starts the monitor and The default consistency check happens every Sunday at midnight starting from the date of installation.

[e4services]

if the NetRAID Assistant utility being used is from an L.19.xx or older Navigator CD. That CD may have come with the server. Those versions will run consistency checks automatically at midnight every Sunday.

If you download the latest NetRAID Assistant from our web-site, it will be necessary to also download the Monitor Tool from the same download page, in order to schedule consistency checks. The newest version (5.07), will default to 6:00pm on Sundays.

All versions will default to use 50% of the resources of the NetRAID controller, and drives to run. So, if there is a backup scheduled for the same time, it could take a lot longer. Users may notice slower service during the consistency check. That is why Sunday at 6:00pm is the default, and is usually a good time to run one.

The Consistency Check acts as a pass/fail utility. If it finds parity stripe corruption, it will report it; if there is no corruption, everything is fine and the utility simply ends.

[Sean T. Craig]

Hi Ashley,

Likely, there is an error on one or more of the active drives in the array. Check for media & non-media errors on all the drives, either through the NetRAID Assistant or through Express Tools. Next, shut down the server and connect your drives to the embedded SCSI controller and run a Verify on any drives that had errors. If and when it finds an error, choose the option to remap all the bad sectors. When the verify is complete, reconnect the drives to your 1M controller and rebuild.

Let us know how it goes,

Sean T. Craig Sr. C.E.T., M.C.P.

[ashley mallory]

Sean, Thanks for the reply... However, I have already done an image backup, reconstructed the array and restored the image. So far everything looks good.

I'll keep your instructions for if /when it happens again.


Thanks again,
Ashley

[e4services]

Glad it worked out Ashley. You would need a backup anyway if you follow scanning the drive for errors as stated above, the image you have, and in rebuilding you corrected any drive errors just as you would have in scanning the drive, so essentially you did the same thing.