RAID 5 SCSI, Drive failed on an old Netserver LC 3

[SGrensted]

1. all of my old 9.1 GB disks are manufactured by Seagate even though they are identified as HP disks. If I simply install the same model of Seagate drive will there be a problem?
2. I did install a new Seagate drive. When I select the replaced drive and choose to rebuild, within seconds the rebuild quits with ERROR. Any suggestions?

[r siva]

Dear SG,

Can you give the Error Detais.

[SGrensted]

At this point I am investigating where I can get more information on what the error is. I am using HP NetRaid Express Tools Ver B.02.02 I access the tools during post. It is not software. I am not sure that there is any information on the error. I will keep looking, but if you can advise me here, please do so. Thank you.

[kris rombauts]

Can you collect the raid.log file from the Netraid Assistant utility.

You have to open the Netraid Asisstant utility to be sure the last controller error log is copied into the utilities log file.

That log file might indicate the SCSI sense keys that the drives send back when the rebuild stops. Most of the time those SCSI sense keys are indicating a media error (bad block on a disk) from experience and it could well be that the original disk has a bad block and cannot be read successfully so that means that you will never be able to rebuild onto any disk (because the source is bad so to say).

HTH

Kris




Kris

[SGrensted]

I have installed the software utility. Here is what I see in the log. It appears that one of the other two drives has an error.

Check condition on CH 0 ID 5 with the following sense key

f0 00 03 00 00 39 8f
0a 00 00 00 00 11 00

[kris rombauts]

f0 00 03 00 00 39 8f
0a 00 00 00 00 11 00

is returned by disk at CH 0 ID 5

This SCSI check condition means a medium error occured while accessing to that disk at ID=5. In this case it reports a Unrecovered Read Error


"MEDIUM ERROR. Indicates that the command terminated with a nonrecovered
error condition that was probably caused by a flaw in the medium or an error in the recorded data"


So the rebuild fails because there is a read error on a disk that contains the remaining stripes in the RAID5 to rebuild the array.

Most of the time these errors are being corrected if you let the system check the consistency of the array. There is a Netraid consistency check that by default does a weekly consistency check of the whole array. You can download this and it is best to use that in the future to try to limit this kind of issues.


So the only safe solution now is to backup data and re-install this server.

HTH

Kris

[SGrensted]

Kris,

Thank you for the information. Do you have a source that interprets these key errors?

It sounds to me like you not implying that one of the remaining disks in the array has a physical error, but rather that there was a consistency error in the data on the array. (I don't think I want a flawed disk in the array.)

I already ghosted the image of the partition. Since it is a Windows 2000 server I want to send the ghost image back to the array after I setup the array from scratch, otherwise the computers that logon to this server will need to be reconfigured as well.

Any comments?

[kris rombauts]

http://docs.hp.com/en/J6373-90030/apas01.html#ciadehfd

Often a reformat fixes this again and it is not a real media problem that cannot be fixed (i.e. one that needs a hard disk replacement).


Kris