Windows 2k Server RAID 5 not booting

[cscheck]

Hello all,
I have been very pleased with the amount of great advice I have recieved from the members of this community so I come to you again in dire straits. I have an HP LH 3000 Netserver that was given to me not long ago. It has 12 SCSI drives configured as one container in a RAID 5 configuration. This last weekend the server was running a bit slower than usual so I decided to give it a reboot after applying a few patches. These patches were nothing that should have caused a problem as they had been tested on other systems with the same configuration (minus drive size) without any problems.

I have my entire life on this server and now it will not boot up. I know I probably screwed myself when I failed to install a tape backup system but I was informed that RAID 5 was safe. I am a "server" noob, I primarily work on workstations. The first time the system went to reboot it went through the POST test just fine, but failed to see ID 8 on Channel 0. At the suggestion of a co-worker I pulled all the hot swap drives, and reseated them all to make sure it wasn't a connection issue. This caused the SCSI utility to detect all the devices properly, however it still won't boot. The goes through all the post testing, and jsut about the time you would normally get the boot screen it goes to a black screen with a cursor blinking in the upper left corner of the screen.

I have tried ERD's but I can't get the server to boot from the floppy drive. I have even tried the windows 2000 server CD and I cannot get it to see that to boot from either. In fact when I do get to the Boot menu it will not acknowledge my selection of any 1 thru 5 option. I tried to select the CD-ROM and the flexible disc and it just stares back at me. The only thing I can do is hit enter which simply takes me into the BIOS setup.

Is there anyone who can help me get any information off of these drives? I have school, work, and my entire life backed up on this server, and I cannot lose then.

Please and thank you in advance.

 

[velis]

Huh, this seems like a very serious problem.
Usually such servers "run slower" because the RAID is having problems with one of the drives, which you apparently confirm saying the BIOS didn't detect ID 8.
The problems you describe about the boot process seem to suggest it was your server that got faulty (MB, CPU or RAM), not the RAID setup.
I might dare to claim you'll be fine if you move the RAID to another machine. Anyway it would be worth a try, but i fear the RAID card itself isn't exactly a common PCI type, but rather one of the more advanced (server) variants, which is going to make your life a bit harder.

You should be aware though, that RAID 5 is not entirely safe, especially when 12 drives are involved. Only 1 (one) can fail at any given time for the array to still be usable. With 12 drives, you have quite a big chance of multiple simultaneous drive failure. So backup would still be in order.

I guess the best course of action would be to take the server to HP support and have them repair it / get the data from the array, depending on whay you intend to do with the server.

EDIT: StorageReview forums might be a bit better place to ask these questions

 

[cscheck]

Unfortunately this is running into an on board controller. I fear that I may have lsot all the data reagardless because I know that changing RAID controllers can really screw up the info on the drives. I will call HP even though this was purchased second hand. I jsut wish there was an easy way to extract that little bit (4 Gb) of data that I desperately need. Thanks for the help velis, and keep the ideas coming everyone. I will try anything at this point.

 

[velis]

It is not absolutely necessary for the data to be unreadable by ALL controllers. Maybe HP can tell you which controller would be compatible with the one you have. If not, I'm sure they have such a machine available so they can move the disks to that machine to extract the data on the array.

 

[TGS]

Could it be that the container was rebuilding the failed disk to the hot spare? Don't tell me you pulled that hot spare when it was on...?

 

[cscheck]

I have been unable to "rebuild" any of the drives. By entering the HP SCSI Utility tool its tells me that all drives were online, and that a drive must be in a failed state to rebuild it. No drives were in a failed state. I only had a message about the system no detected the device on ID 8 one time, and after a shutdown>pull drive>reseat drive, that error did not return. The LED's on the drives have never gone red, which leads me to believe velis' suggestion about the MB or Controller card being bad.

 

[Brazen]

Find out what chip the onboard SCSI controller uses, you may be able to find an add-in card with that same chip and maybe boot it up using that. Also, you could try booting to an IDE drive in that same system, just in case the issue is not the SCSI devices at all, but something else on the mobo or other hardware is failing at that point in the boot process.

 

[cscheck]

Good advice, I am going to take the server over to a friends house tonight and try to boot from a cd with the scsi drivers on it, and I will also try to boot from an IDE drive. Thanks,

 

[cscheck]

Well the drives were initialized at some point. bye bye data... What a crappy experience with SCSI. I think I will stick to the PC servers for now. Thanks for the advice everyone.

 

[GreyMittens]

There is nothing wrong with scsi or raid, and not to sound insensitive but..

"I know I probably screwed myself when I failed to install a tape backup system but I was informed that RAID 5 was safe. "

That was the problem and it could have just as likely occured on IDE or SATA drives in your workstation if the drive bombed and you had no backup.

 

[Smilin]

The LH 3000, if I remember right, has a kind of goofy RAID controller in it. It is basically two SCSI controllers bridged with another device. The symptoms of not booting to CD or floppy indicate something other than a failure of the controller or drives.

I wouldn't "fool with" this too much before calling HP. Don't try to perform any drive maintenance until they confirm it truly is a problem with the drives. As bad as it is now, your data is likely intact, so don't make it worse.