-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 For quite a while now, I've been running a software RAID5 using external firewire disks on a Debian Unstable system. The exact architecture was four 250 GB disks on one firewire controller, for a total of 750 GB of usable space.
Last night, when I was trying to unplug something under my computer desk, I accidentally unplugged one of the external disks (/dev/sdd). No problem, I thought, I can just re-add it with mdadm. The problem is that when I tried, using "mdadm /dev/md0 --add /dev/sdd1", the rebuild would kick off and then fail after a short time, marking all four drives as faulty. So I rebooted, running the RAID in degraded mode, fdisked and mkfsed /dev/sdd1 again, and tried again. Same result. I replaced the firewire PCI card. Same result. I reformatted another firewire drive I had lying around and tried to add it, in case the problem was the drive. Same result. Every time I try to add a replacement volume to the RAID, the rebuild fails and I get this message: "Buffer I/O error on device md0, logical block 1545" I tried running "fsck -C -p -f /dev/md0" and got a whole bunch of the same error in the logs, with different logical blocks. Does anyone have any suggestions for cleaning up my degraded RAID set so that I can put another drive in the pool? All of my data seems to be okay (although you'd better believe I've backed the irreplaceables up again to be sure), but I really don't want to deal with another disk failure. - --Matt -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.4.9 (GNU/Linux) Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org iEYEARECAAYFAkhefY8ACgkQ6wBqCH7pnjJ/OgCeN1tEyE2cDN5GZ4s5juJcidiU on4AoLt7OTKWNed5o8d+wi9DfcX/9gQU =6vB3 -----END PGP SIGNATURE----- -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
