"kstratton---" == kstratton--- via WLUG <wlug@lists.wlug.org> writes:
I am looking googling mdadm for a raid 1 drive. In particular I am making notes of how to mount a raid1 drive (for emergency recovery). Is the expected location of the raid array at /dev/mdXX or /dev/md/mdXX? Is it one of those two is historical or is it one of those thing that just does not matter?
So if you are talking about mirrored drives, you should be able to start the array with one drive missing. So you won't have any problems getting your data back, because it hasn't really gone away. If you look at the output of /proc/mdstat, you will see something like this (using one of my one RAID1 pairs as an example): $ cat /proc/mdstat Personalities : [linear] [raid0] [raid1] [raid10] [raid6] [raid5] [raid4] [multipath] md0 : active raid1 sdj2[4] sdi2[3] 185545656 blocks super 1.2 [2/2] [UU] bitmap: 1/2 pages [4KB], 65536KB chunk So /dev/md0 has two disks, /dev/sdj2 and /dev/sji2. In the worst case, I could boot the system from a LiveCD and just do: mdadm -A -a auto to assemble automatically any arrays it finds. A good reference to all this is the Linux RAID Wiki at: https://raid.wiki.kernel.org/index.php/Linux_Raid but the big big big thing I want to stress is that you *never* run the 'mdadm --create ...' command when trying to recover a broken RAID array of any type unless you are absolutely certain you know what you are doing. DO NOT. Get help first. Post to the Linux RAID mailing list. But DO. NOT. try to create an array to recover it. It's almost certainly the wrong thing to do. I'd be happy to talk Linux RAID at the meeting, but as most of you know, I'm happy to talk period. See you thursday! John