I run this Compaq ProLiant 1600 server:
- 2x Intel Pentium III 550mhz (Katmai, 512kb each)
- 512mb SDRAM (PC100/133)
- 4x 9.1gb ULTRA-2 SCSI 10k RPM
I'm having some major issues with the RAID array and the hard disks.
My server had been on for about a week (24/7). I woke up in the morning to find it had BSOD'd. Pretty unusual, as it's usually so well behaved, but it is Windows so it's bound to happen every so often, nothing you can do.
Anyway, I restarted the server and the drives started spinning up. Except this time, 2 of the drives presented errors. I should point out at this stage that Drive 4 has had a 'drive failure' light on for about 3 months, and still worked perfectly well (passed disk tests). Now drives 2+4 weren't spinning up and drive 2 had no lights on at all.
I switched around the caddies, tried changing the order of drives etc... but to no avail. (I restored it to it's original order btw.)
I went into Windows setup and, while it recognised both logical drives on the RAID array (there's one of about 4gb for Windows and one of about 30gb for backup - it's a 'Vital files' backup server for my network.) However, it said it couldn't access either of them.
So I thought I might just take out the two drives presenting failures and try the two remaining good ones in a smaller array. However, when I did that, drive 1 presented a failure light. Now I have 2 drives presenting failure lights, one drive presenting nothing at all, and 1 drive that still works.
Sounds like the controller or the backplane might have gone bad. Not much to do now, except maybe try one disk at a time. You might also try to eliminate the backplane by using a scsi cable directly to the disks, but since they are hotswap disks you may not be able to find a cable to fit them.
I didn't actually set the array up myself (factory) but as far as I can tell it's RAID 0.
It's 4 physical disks set to 2 logical drives of different size - one 4gb and one 30gb - although whether or not it's one drive with 2 partitions or 2 drives I'm not sure, although it does say 2 logical drives so I guess it's 2 drives.
It's more than likely a raid 5 [added: Oh, no, if you have 34g of disk space then it's not a raid5, so you're probably right that it's a strip set]. Yes, the backplane is the board at the back of the drive cage that the disks slide into.
Oh, and you do know that your array is now toast, so once you get the problem sorted you'll have to reload the OS? As soon as you started moving disks around and taking disks out then you very likely blew the array. I'm just saying that in case you intend to recover that old array. If you just want to start over then cool.
Pain: like I said, it's a backup server, and as long as I've got the originals it's cool, I'm not really bothered about losing the data. I wasn't hotswapping the drives if that helps.
Rippleyaliens: Nope, not the latest updates (I try to steer clear of firmware updates until it's neccessary) but I'll try. I can't actually boot the system to run the Compaq array manager. I'm not sure what version of startsmart I have (or what startsmart is, to be honest!)
I'll try firmware - keep the suggestions coming in this is great, thanks.
I think that's the best, and probably the only think you can do right now...to update all the firmware, etc.
It won't matter if you hot swapped the disks or not, if they were written to after you moved them around or removed some of the disks completely, then the array is gone. If nothing was written to them during that time then it may still work, but since you have all your data I would certainly start over from scratch when/if you get it running and reformat/reload.
When the system starts up it says (after RAM/CPU initialization):
Adaptec AHA-2940 present
(this is what I use for my HP DDS2 drive)
then the array controller initializes:
COMPAQ SMART ARRAY 221 (Rev. A v. 4.16) 2 Logical Drives
1780 Slot 4 Drive array failure
The following SCSI drive(s) should be replaced:
SCSI Port 1: SCSI ID(s) 0, 4
I looked for some files on the HP website in relation to my server. I found:
A SmartStart bootable (I think) CD-ROM, and some bootable firmware floppies for the 5300 array controller, which it says is compatible with my server, though mine says "SMART ARRAY 221" but I guess it just won't work if it's the wrong file.
If you still want to play with this then go into the array control menu and force the failed disks back on-line.
Assuming that this machine has been running fine for years, I have a hard time believing that it is due solely to outdated firmware. I would still check for any updates and apply them first. I would also open the machine up and blow all the dirt out of the backplane fans, from around the disks, etc.
Spooky. I'll try changing the SCSI ID of the last remaining failed drive and try adding it to my new array. I went for RAID 0 by the way, I figured I need all the disk space I can get with the low capacity drives and I've just proved that it can afford to be down for a few days.
Well, of the four drives one was showing it was failing, so I used the Compaq StartSmart tools on bootable CD-ROM to reset all system settings and drives and then built a new array out of the 3 remaining good drives. I'm currently exploring the possibilities of Linux with the help of linux_0 but it's not working too good at the moment so I may just have to switch back to Windows.