Failing RAID1 Array... or was it? [solved]

[Algenon]

I thought I would share this peculiar problem I encountered in case any one else has the same symptoms.

My system is a Gentoo 2006.1 AMD64 (upgraded from 2005.0) running RAID + LVM. I have a pair of 250GB disks configured as RAID1 which are housed in removeable drive caddies. The system has worked extremely well over the years without any major problems. The 250GB disks are recent upgrades (October 2006).

Out of curiosity, I sometimes remove the dual drives and insert a single drive when I want to try out another distro. Having done this on one occasion, on replacing the dual RAID disks, the system reported /dev/hdc failure. Using mdadm -a to rebuild, brought the disk back to life, but it failed again a couple of days later. The disks are only a couple of months old, so I was surpised that there was a hardware failure, and a couple of days later, /dev/hda failed. mdadm allowed me to rebuild without any loss of data.

Still not beliving that the drives were faulty, I suspected the RAID software and was about to embark on building the system as non-RAID when I also noticed that the machine was failing to find the boot sector (obviously no RAID or OS was running at that point). Then I realised what the problem was.

The drive caddies were suspected of making poor electrical contact, and so I wired the drives directly onto the IDE ribbons without the caddies and now the system has been stable for well over a week now.

So, the morale of this tale is buy good quality caddies (or don't use them at all). In order to continue to run alternative OS's I have descovered VMware (downloaded via portage) and although not the subject of this thread, I would just like to say what a fabulous piece of software this is.

Regards,

Algenon.