Unsolved
This post is more than 5 years old
3 Posts
0
21471
Replace failed disk - 600SC with CERC ATA100/4-Channel RAID Controller
My company purchased a PowerEdge 600SC nine years ago and it's been a reliable Windows 2003 server until this past weekend when a disk in the mirrored array failed. We're a small operation and have no real business need (nor the funds) to upgrade the server to something more modern so I'd like to replace the failed IDE disk with another to buy us time until we can afford to upgrade, perhaps in 2013.
The task seemed simple enough--from what I understand, I should be able to shut the system down, swap the failed unit with a new one, power up, and the RAID controller should handle rebuilding. I ordered two new drives (just in case) and had them sent via overnight delivery. Upon arrival earlier today, I performed surgery to pull out the failed unit and install one of the new drives. Upon power up, however, the CERC controller complained the new drive was "Failed." OK, maybe I got a dud. Power down. Pull drive. Replace with 2nd new drive. Power up. Unfortunately, I got the same result--the RAID controller indicated the 2nd new drive was also in a Failed status.
I double checked the jumper on the new drive--cable select, just like the existing (good) disk.
I then tried moving the IDE ribbon cable to different port connectors on the controller card--there are four. No success. Next, I tried each port with a different ribbon cable. No good. I then purchased a third IDE drive locally and swapped it in. Same result. The array controller appears to recognize each of the replacement drives--the drive capacity shows up correctly when I view the physical drive object. But the array controller continues to show each of the replacement drives as "Failed" when the system is powered up. My solutions so far:
1. Swap out the failed disk with a new disk. Repeat with three separate IDE drives including a brand new unit with a larger (320GB) capacity.
Result: Array controller shows each replacement drive as Failed.
2. Swap out the ribbon cable.
Result: Array controller shows replacement drive as Failed.
3. Move ribbon cable to different ports on controller card. Tried all four ports.
Result: Array controller shows replacement drive as Failed. Note--good drive shows up as "online" in all cases, regardless of array card connector port.
It would appear the controller card is working, given that the good drive continually shows up as "online" regardless of connector port used on the array controller card. I know that IDE drives are ancient technology, but it still seems unlikely that I'd get three bad drives. (Right?) In all three cases, the array controller card software recognizes the drives but indicates a "Failed" status whenever the system is powered up.
To the best of my knowledge, I shouldn't have to do anything special to the new drives to get them working with the controller--I should theoretically just swap out the bad drive with a new one, power up, and the array controller should rebuild the mirrored array using data from the good disk. But maybe I'm missing something obvious.
Oh, one last thing--I've tried a manual rebuild. The status changes to "Rebuilding" briefly, then "Ready" and then "Failed." The whole process takes about 10-15 seconds.
I'm running out of ideas. If anyone has any suggestions, I'd sure appreciate hearing them. Thank you.
DELL-Chris H
Moderator
Moderator
•
8.5K Posts
0
October 10th, 2012 07:00
Jerry,
Did you ever select Rescan Drives in the controller BIOS?
You find this under Array Configuration Utility in the Cerc BIOS. After you select Rescan the drive should start rebuilding to the array.
If that fails then you can go to Manage Arrays, select the Virtual Disk and hit CTRL-S, then select the replacement drive. That will set it as a hotspare and should initiate it rebuilding to the array.
You are correct, the drive should normally rebuild after your power up, but if the controller is out of date or something it can cause issues.
Let me know how it goes.
jerryspaeder
3 Posts
0
October 10th, 2012 17:00
Thanks for the suggestions, Chris. I scheduled some server down time this weekend and will test your ideas. I also ordered a spare controller card to have on hand just in case. I'll let you know the results.
jerryspaeder
3 Posts
0
October 12th, 2012 18:00
I swapped out the RAID controller card and introduced new IDE ribbon cables. I now have a good drive running Windows 2003 and three new spare drives. Each of the three new drives continue to show up as failed regardless of controller card port. I have a good backup so I went as far as resetting the configuration on the card using the BIOS utility--deleting the virtual/logical disk and recreating it. Still no luck. Windows 2003 still starts up and I can view all four disks attached to the card using the Windows Array Manager. Two of the disks are associated with the logical disk--one online and one failed. I set one of the two spare drives as a Global Hot Spare and watch it take over. A rebuild commenced and promptly failed within 10-15 seconds. I then assigned another drive as a Global Hot Spare. Same story--the rebuild started and failed. I have now rotated through each of the 3 disks--setting them as spares and watching them attempt a rebuild only to fail.
My next test: Create a new virtual disk and assign two of the drives. Success. (Interesting.)
Connections:
Channel 0 - Used in original virtual disk (Status: Online)
Channel 1 - Used in new virtual disk (Status: Online)
Channel 2 - Used in new virtual disk (Status: Online)
Channel 3 - Used in original virtual disk (Status: Failed)
I was then able to create a partition on the new virtual disk and format it NTFS. After setting up a new virtual disk using the array manager, I deleted it. I then assigned the disk on Channel 1 as a Global Hot Spare. A rebuild commenced and failed:
Channel 0 - Used in original virtual disk (Status: Online)
Channel 1 - Used in original virtual disk (Status: Failed)
Channel 2 - Not in use
Channel 3 - Not in use
I then created yet another virtual disk, this time using the physical disks on channels 2 and 3. I was (again) successful in creating a virtual disk:
Channel 0 - Used in original virtual disk (Status: Online)
Channel 1 - Used in original virtual disk (Status: Failed)
Channel 2 - Used in another new virtual disk (Status: Online)
Channel 3 - Used in another new virtual disk (Status: Online)
To recap:
My conclusion: All hardware is functioning. Could something be on the original disk that is preventing the controller card from rebuilding? Is there some special header on the disk that could have become corrupted? If so, can it be fixed without losing data?