netRAID error

[David Wilson_16]

I have a netRAID warning apprearing in the application log as follows:

Event Type: Warning
Event Source: NetRAID.Log
Event Category: None
Event ID: 3202
Date: 18/08/2004
Time: 10:41:08
User: N/A
Computer: CONTROL
Description:
Adapter 0 Channel 0 Target 0: Extended ErrorEvent: [Time:MM-DD-YY:hh:mm:ss: 08/17/2004: 10:34:27] [10-bytes CDB: 28 00 02 5c 8d 00 00 00 80 00] SenseData: errCode=0x70, valBit=1 segMent=0x00 SenseKey=03 ILI=0 EOM=0, FMRK=0 Address=0x025c8d39 ASL=0x18,cmdSpec=0x0000, ASC=0x11,ASCQ=0x02 FRUC Code + Sense Key Specific=0x80008000

and also

Event Type: Warning
Event Source: NetRAID.Log
Event Category: None
Event ID: 3201
Date: 18/08/2004
Time: 10:41:08
User: N/A
Computer: CONTROL
Description:
Adapter 0 Channel 0 Target 0: Media Error Count=4, Other Error Count=0

the server is not hanging as i have previously seen in discussions but simply throwing this error message up three to four times daily. Should i be considering a new disk ASAP?

any help would be appreciated!!

[Greg Carlson]

David,

Install Netraid assistant and see if you have check condition sense key errors also for the hdd. You also need to verify fw of the hdd and controller. If you already have netraid assistant installed. Look in the netraid log to see if you have errors reporting on ID 0. Most likely you have media errors on the hdd which are not being remapped correctly. Verify all other hdds for media errors before replacing any hdds though, because if you have more than one hdd with media errors, you will fail rebuilds. Let us know if you have this scenario.

hth,
Greg

[kris rombauts]

David,

your disk at SCSI id=0 has media problems as shown by the SCSI sense key in the events you posted.

SenseData: errCode=0x70, valBit=1 SenseKey=03 , ASC=0x11,ASCQ=0x02



If you get a couple of those per day then i would plan the disk replacement asap to avoid you run in to bigger problems like disk failure and make the array non redundant.

As Greg correctly pointed out, make a consistency check first and then check the logs (and also make a backup) because if another disk in the array happens to have a media error then your rebuild (after disk ID=0 is replaced) will never be succesfull since more then one stripe is missing.

If you do have multiple disks with media erors, a full backup/restore is the only solution anyway, it's just that you know upfront that the rebuild won't suceed.

Check your full eventlog under Windows for any entry from source = Netraid.log that is not from the disk at ID=0 or start Netraid Assistant and open the log viewer and check which disks have logged sense keys that are media related.

HTH

Kris

[David Wilson_16]

Guys thanks for the prompt responses about my errors, i have taken all of yesterday to investigate the problem and here are my findings.

I have a raid 5 array with three disks, two of the disks are producing the errors not simply one as i thought. Nightmare.........

I have run consitency checks and the event log filled up with errors from both drives, they are drive 0 and drive 1.

I assume from the posting that the server shall have to be rebuilt with new disks?

Thanks for the feedback.
David

[kris rombauts]

Hi David,

would be good to copy the raid.log file contents and eventually the Systemlog entries and attach them here in the forum.

Raid.log is the file you view with the Netraid Assistant Log Viewer and is normally located in c:\Netraid or alike (depends on version).

If the sense keys are indeed media errors (and not something else which is why we need to see the SCSI sense keys) then a array rebuild will be needed unfortunately.

Kris

[David Wilson_16]

Can you give me your email address and i can mail the event viewer logs and the netraid logs.

my email is david@certum.co.uk

[kris rombauts]

David,

it's clear that even since the start of this log file (1 July) the disk id=0 already had media issues and as of 4 July I see disk 1 with media issues also (and probably from before already). Both absolutely need to be replaced since they report unrecoverable read errors when a SCSI read command 28Hex was send by the Netraid controller.

CDB = 28
Sense Data = f0 00 03 02 19 89 63 18 00 00 00 00 11 02 00 80 00 80



The consistency check however seems to complete succesfull (still now but for how long ???) so if you can swap both disks I would give it a try and rebuild one by one. Start with disk at id=0 since that one has the most problems, then when finished, replace disk id=1, then run a consistency check again.

If one or the other rebuild does fail the only solution is to reinitialize the raid 5 after the disk 0 and 1 are replaced by new ones, this is still required since the media errors wonâ t be fixed with a initialize, this is a pure disk hardware problem they have.


Remember that even after you replaced both disks, when you open the Netraid Assistant again later and do a 'view log' the controller will still add the last sense keys it ever received to the logfile , this is a bug (feature ?) and I described this in this forum before, have a read here :

http://forums.itrc.hp.com/service/forums/questionanswer.do?threadId=628577

So donâ t worry about these as long as no new ones appear in the Windows Application log then you are ok.



good luck
regards,