Dell PE4400 HD & PERC3di craziness 1

[muddawg1976]

Hello all, long time reader, first time poster. I recently purchased a used poweredge 4400 on ebay. Dual 866MHz xeons, 2GB mem, perc3di RAID, 8- 18.1GB scsi hard drives, all standard Dell eqmt.
When I recieved the server, all was fine. I initialized the drives, set the containers, etc. Installed Win2000 server, installed raid driver, setup went excellent. I actually had the server ready to be deployed the next day. I left it running overnight and when I came in the next morning I had a BSOD. I powered off the server and on reboot all containers & volumes were gone. All 8 drives spin up with a green light, until the controller tries to access the drives. Then one drive will flash amber. The interesting thing is, if I swap the drive it shows up as fine. The only recognized drive is the drive that flashes amber. Swapping the drives around results in all of the drives flashing amber.

The only way got the config utility to recognize the drives was to remove the key, mem, and unplug the battery. They all showed up again for one reboot and then they are gone. BTW, BIOS, ESM, & adapter have all been flashed with most recent updates during my troubleshooting.

Any help on this would be greatly appreciated. I will be following this thread closely, and I look forward to hearing any suggestions or answering any questions.

 

[DanZeMan]

Sounds to me like your 3di may be going south.
We've had so many problems with 4400/3Di that Dell ended up replaceing two of the worst offenders with 4600's - probably not going to happen in your case.

Some of the things that we have had to do with our 4400's to resolve disk issues:
Reseat all connections from mobo to backplanes
Backplane replacement (assuming that they are on the 1x2 and 1x6 internal backplanes)
Replacement of keys, RAID memory, mobo
Replacement of power distribution system board

It may be difficult to find dell replacement parts, not sure what 3rd part channels there are available, but you may need to try and find another mobo since the Di is build in....

What info did your BSOD give you?

 

[muddawg1976]

Thanks for the reply DanZeMan. I guess I should update my own post so others can see my solution to this particular problem (for the moment at least!).

I tried almost every trick in the book to solve this problem. I reseated every connection I could find, moved drives around, rebooted with the drive that had "failed" removed, etc.

What finally worked for me was to let the server sit unplugged from any power source with the PERC battery unplugged, and the raid key and memory removed from the system. I left it this way for a day.

After the 24 hours were up, I reconnected the battery, key, and memory, then I plugged the backplane scsi cable into the second channel scsi channel on the mobo. The troublesome drive was still removed from the drive cage. I booted it up......surprise, surprise there were all of the drives still initialized as the original containers minus the "failed" drive. Windows 2000 booted up with no errors, and has been running ever since with no problems.

Once windows was running again, I installed array manager and reconditioned the 3di battery.

The interesting part to the story is this; I bought 4 new replacement drives, put in one of the new drives into the trouble slot and, of course, it is working fine.

The "failed" drive that caused all of the problems was moved to a Dell 2300, reformatted, and now it is working fine on another system.