Compaq raid 3200 controller..imminient failure

[joeblk]

Hi all,

I have a compaq smart array 3200 raid controller. This is configured in raid 5 array running w2k. The machine has 12 x 18.2 gb disks. The problem is that at boot up the machine has the foll msg: 1720 slot 5 drive array..imminient failure.
Question:
1: How do you identify the failing disk? (there are no amber lights on the physical machine - all the lights are green)
2: I have a replacement 18.2gb disk, how do I do this? My understanding is that I can remove the failing disk (once I'm able to id it) and just put the new one and it should rebuild automatically..is this correct?

Can someone please assist.

Thanks
JB

 

[ntinlin]

Easiest way to do it is through the insight management agents if you have installed them. Go to http:\\servername:2301 and then login if required. Default user is Administrator with password of administrator.

You should be able to see the failing drive in there somewhere and there will be a control to flash the lights on that particular drive to identify it.

Neill

 

[TheLad]

Agreed. And you are correct, on a RAID5 setup if you remove the disk and replace it, the system will restripe as normal. However since the disk has not actually failed yet, it is risky removing the disk as it will cause the server to force itself into interim mode (the state whereby a disk has failed) - saying that, this is what the server would do if the disk did actually fail but I would still suggest that you remove the disk at a time when the server is less busy than normal.

-----------------------------------------------------
"It's true, its damn true!"
-----------------------------------------------------

 

[joeblk]

Thank you both for your replies. I will check to see if the management agent is installed. Otherwise I'll see if I can convince the client to wait until the drive actually fails.

Will keep you updated.

Thanks
JB

 

[TheLad]

You might be better off swapping the disk as it could cause problems with the server if there is a physical problem with the disk. You know how a PC gets when the hard disk is on it's way (starts making a ticking noise and the PC hangs whilst the HDD either writes the data or fails), well the same applies with a server and you could cause problems (including corruption)with the logical volume if one disk is having read/write problems.

-----------------------------------------------------
"It's true, its damn true!"
-----------------------------------------------------

 

[joeblk]

OK, I have checked and the server does not have the management agent installed. What I then did was I removed a physical disk in slot 5 and then restarted the machine and got the following msg: drive controller 1 (slot 5) Hard drive 1 - scsi bus 2, drive 104. Array accelerator parity read error: 12.

Then I put back the physical disk and it took > 3 hours to recover the logical drive; and then it's back to the initial stage.

 

[TheLad]

So you put the faulty disk back into the server? Not wise in my opinion. I have seen servers corrupt as a result of putting a faulty disk back in...

-----------------------------------------------------
"It's true, its damn true!"
-----------------------------------------------------

 

[joeblk]

I don't think it was the faulty disk. At the moment it all seems to be working OK though the imminient faillure msg is still there. I simply do not seem to be able to diagnose the faulty disk even using the diagnostic CD that came with the server.

As it stamd, I have decided to leave it as it is and only change it when the disk actually fails. So in order to be ready for that, can anyone confirm or advise the best way to replace the disk when it fails please.

My understanding is that while the machine is online, I can simply remove the failed disk and replace it with a new one and RAID would rebuild automatically.

Please advise me what to do, so that I am ready for when it actually fails.

Thanks
JB

 

[NortonES2]

When the disk fails pull it out and replace with a new disk, the RAID will then rebuild automatically.

-------------------------------

If it doesn't leak oil it must be empty!!

 

[newBPhoneGuy]

i believe you have to remove the disk from the array before you actually pull it out

 

[TheLad]

Can you give some more detail on that? I have never known of any option to remove disks from an array?

-----------------------------------------------------
"It's true, its damn true!"
-----------------------------------------------------

 

[newBPhoneGuy]

sorry about that but what i meant to say is that you need to set the the drive to offline before you remove the actual drive but if you have a hot spare configured i dont think you need to do that but its always safer to set it offline before you remove and once its set off line you can remove it and if the drive is configured correctly it should start rebuilding by its self..........i believe you can set the status of the drive to off line through the raid manager software..... im sorta new to this stuff but that is what i was tought to do......

 

[newBPhoneGuy]

sorry about that but what i meant to say is that you need to set the the drive to offline before you remove the actual drive but if you have a hot spare configured i dont think you need to do that but its always safer to set it offline before you remove and once its set off line you can remove it and if the "NEW" drive YOU PUT IN is configured correctly it should start rebuilding by its self..........i believe you can set the status of the drive to off line through the raid manager software..... im sorta new to this stuff but that is what i was tought to do......

 

[newBPhoneGuy]

lad i appologize you only need to offline the drive with a dell or an IBM i went ahead and tried it with the compaq and you are right there is no option to off line the drive