ML370 G2, Hot Swap disks. Removal process.

[Paul Taylor]

Hi all,

I have 9 Compaq ML370 G2 1.4Ghz servers in Service. All have the 5i Raid controller.
All servers run Win 2000 SP4 fully patched, and run the PSP 7.40b.
Firmware Maintenance CD 7.40 has also been run on all my ML's

I have one server that has a drive which keeps dropping out of the raid. When i reboot this server it prompts to rebuild the disk which dropped out of the raid. So i re-build the raid and it carrys on fine for a few days and then drops out again.

Ive just had some new drives arrive, so my question is can i just pull the drive out live and put a new one in ??. Currently the drive is Green but it oviously has an issue.

Normally when i get a dead drive i shut the server down and swap it out. I do this because i had a bad experiance many years ago with an older Series 1 ML370. I pulled out a dead drive while the server was running and Windows crashed with a blue screen. This has forever put me off Hot-Swapping. It was a mission critical server and i got my arse kicked due to the downtime. ( We have all been there ! )

So, to Hot Swap, or Not to Hot Swap, that is the question.

Regards, and thanks to all,
Paul Taylor.
IT Manager,
Lo-Q plc.

[Don Eilenberger]

My experience with an ML370-G3, with the Smart-Array 642 Raid controller was - I could swap out one of the 3 drives whenever it was necessary.

In my case - I keep 4 drives in the bay - one hot spare. When one failed I just pulled it and replaced it with the hot spare. When the replacement arrived I installed it as the hot-spare.

No fuss - no blue screen.. but of course YMMV. Also running W2000-Server, SP4.

HTH,

[Hense, Klaus]

Hallo,
you can change the drive online with no problem if all drives of the array have the green light on.
A better way is the plug in all server a HotSpare drives.
When a disk is corrupt automaticly the sparedrive takes his place and rebuilt the array.
When arriving the new drive you change it with the defekt drives and automatic the new drives rebuild with the array. After this the spare drive is ready for the next crash...
All this you can make without reboot !

[e4services]

Verify the data after one of the rebuilds. The drop out is more likely a RAID problem that is data related and not a physical drive error. If the drive had physically failed, you would see a failure light on the drive itself.