Hi, I've recently encountered a problem with one of my Linux RAID boxes, and I was hoping that someone on the list may have seen this before and have an answer for me. I've always liked the ICP Vortex controllers, but I've been bitten by what appears at first glance to be a bug of some description this time. I have a machine with a GDT6537RP 3 channel RAID controller, and it was configured with 4 Fujitsu disks attached as to channel A with ID's 0,1,2 and 3. This system was working fine, but I have recently purchased a new case for it, and a bunch of hot-swap drive caddies. So, yesterday, I remounted the motherboard et al in the new case, and swapped the drives over into how swap caddies. At the same time, I changed the allocation of drives to channels, such that the new config has drives on Channel A, ID 0 and 1, Channel B, ID 0 and 1, and Channel C, ID 0. The fifth drive is an identical drive that was to become a Hot Fix drive. In the process of moving the drives over, one failed (the first drive in the array). Rebooting the machine popped up a message about activating FAIL mode for the array, but then the machine booted normally. The failed drive is seen by the controller as missing (it's the power connector on the drive that's damaged. After booting the machine, I started the GDTMON program from the console, which gives an option to replace a failed array drive. Doing this appeared to work - the disk took about an hour to rebuild, and the array showed itself in the ready state afterwards. So, I rebooted the machine again, and it seems that the controller has failed to integrate the new disk into the array - it's presenting it to the host operating system as a second RAID host drive, which causes the machine to fail to boot, as LILO is confused. Going into the BIOS setup routine for the GDT controller and removing this new host drive got the machine booting again, with the array in FAIL mode again. I went through this rigmarole twice with the same results. I also added the new drive as a pool hot fix drive using the GDTSETUP options, but the controller isn't using this newly designated hot fix drive to rebuild the array as I'd hoped. I now have a machine with no redundancy on the array, and I'm rather nervous... I also have no idea how to convince the controller to regenerate the array, as the provided "replace drive" option goes through all the motions, but just doesn't work properly. If anyone has seen anything like this before, or if anyone from ICP is reading this list and can help, and information would be greatly appreciated. The machine in question is mission critical, and I don't want to be without the redundancy for any longer than absolutely necessary. Mike. P.S. Please CC any replies to my work email address - [EMAIL PROTECTED]
