On Saturday 15 December 2007 11:00:12 Tim Hempstead wrote: > Hi > > Woke up this morning to find that my server is making some noises > which are no normally associated with healthy disks. Looking on the > system everything still looks fine in mdadm but there are some errors > against sdg in dmesg indicating a media error and this combined with > the noises its making makes me think its on the way out. The server > is running 10.3 (32bit) and the disks are all 500gb SATA drives. Also > the raid array does not hold the base OS filesystems just data > filesystems. > > Now sdg is part of a 6 disk RAID5 array (md0) using Linux software > raid. I am assuming that the process to replace the disk is something > like this: > > 1). Make sure I have an up to date backup of everything important off the > server 2). Fail the disk in the Raid array by using a mdadm command? (as > its not showing in there as having issues yet so I'll need to force it) 3). > Remove disk from array using mdadm command?
mdadm --manage /dev/md0 -f /dev/sdg to fail it, and mdadm --manage /dev/md0 -r /dev/sdg to remove it > 4). Establish which physical disk is sdg (I have it down to 2 already i > think) 5). Shutdown server > 6). Replace disk with another of the same size (fortunately i have a > spare). 7). start up server > 8). Add disk to array (as new spare?) > 9). re-sync the array onto the new disk 8) and 9) is just one step. mdadm --manage /dev/md0 -a /dev/sdg assuming it gets the same device name, which I think it should Also, don't forget step 10) Buy a new spare, for the next emergency Anders -- Madness takes its toll -- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
