Compaq Array - Ready for Rebuild?

[BrianH1]

Have a problem with a Compaq 2500 with a RAID5 array.

One of the drives in the array failed, and this was replaced with a new one. Normally the array would rebuild, but in insight manager the status is "Ready for Rebuild" and shows that there were errors on another disk a good few thousand service hours ago.

When I restart the server, I now get two options
F1 - Attempt Data Recovery
&
F2 - Continue without attempting data recovery

For now, I have selected F2 and the server starts up fine, but the array still won't rebuild.

Am I right in thinking that option 1 will force the array to rebuild?

 

[TheLad]

There may be two solutions to this problem:

a) Check firmware of RAID controller and upgrade if necessary

b) Rebuild server as logical volume is corrupt

I have had both solutions, both on Proliant 3000 machines

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

 

[Petje]

As far as I have seen in my servicing compaq servers history.. you should have chosen option 1.. this would have rebuilded the raid set...



Petje
A+, MCP, MCSE on NT4.0 and windows 2000 with messaging specialty and CCEA

 

[BrianH1]

Thanks Both.

The plan is to back the server up, then restart it and choose option 1.

The hope is that as Petje suggests, it rebuilds fine, but I'll be armed and prepped to rebuild if I need to (media, replicas removed etc.)

 

[terry712]

sounds like best plan

you say replicas - is this a netware server?

if so what does cpqonlin say -

 

[BrianH1]

It is NetWare, 4.11. CPQONLIN gives the basic options delete array etc. Tried a newer version and it caused the server to ABEND.

Had some error writing to VOL1 and TTS shut down last night at about midnight while my backup to TSM was running.

 

[TheLad]

Yeah, CPQONLIN has a habit of abending servers, never quite found out why but it could be firmware related if it is the newer version that abends the server (RAID card might need the latest firmware).

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

 

[BrianH1]

I've got the firmware flash for the raid card which I'll be using before I try to get the array to rebuild. When my backup is running now, the server goes mental and throws out error reading from FAT, shuts TTS down and then restarts.

I have downtime for Saturday, so with any luck I can back it up then, maybe to tape if the TSM backup is still bonkers. Then I plan to flash the SMART-2/P controller with the latest firmware and then get the array to recover (see option 1 above)

Wish me luck...

 

[terry712]

sounds a good plan

must confess though cpqonlin has always been rock solid with me - never even seen it abend a server
except a vinca one - but smile at them could cause an abend

 

[BrianH1]

Quick update. Spent 10 hours on site on Saturday, most of it trying to get a good backup.

Tried the recovery option but the array rebuild was still queued and wouldn't start.

Ran the DAAD and looks like the disk our hardware contractor put in the server as a replacement for the failed drive is faulty! Have raised a call for them to replace this disk, will keep you posted...

 

[BrianH1]

It's been a long time since this problem and it was indeed due to the replacement of the failed disk with a faulty reconditioned unit.

This being the case I imagine you'd see the same "Ready for rebuild" message on any Compaq Proliant server or HP proliant server.

1st Rule of Tech Support - Never believe the user.
2nd Rule of Tech Support - Don't assume that new hardware is fault free!

 

[technome]

Raid arrays will often fail a disk which will test out fine with scsi test utilities or run great on a standard SCSI adapter. When returned, they often end up as reconditioned or recerts.