Good morning, On 12/12/18 4:18 PM, Alex Sommer wrote: >> Von: "Phil Turmel" <philip@xxxxxxxxxx> >> An: "Alex Sommer" <alex@xxxxxxxxxxxxxxx>, "linux-raid maillist" <linux-raid@xxxxxxxxxxxxxxx> >> On 12/12/18 2:48 PM, Alex Sommer wrote: [trim /] >> 5) mdadm --stop again, then retry the assemble leaving out the "m", like so: >> >> mdadm --assemble --force --verbose /dev/md2 /dev/sd[ijkl]1 > > ~ # mdadm --stop /dev/md2 > mdadm: stopped /dev/md2 > ~ # mdadm --assemble --force --verbose /dev/md2 /dev/sd[ijkl]1 > mdadm: looking for devices for /dev/md2 > mdadm: /dev/sdi1 is identified as a member of /dev/md2, slot 4. > mdadm: /dev/sdj1 is identified as a member of /dev/md2, slot 1. > mdadm: /dev/sdk1 is identified as a member of /dev/md2, slot -1. > mdadm: /dev/sdl1 is identified as a member of /dev/md2, slot 3. > mdadm: no uptodate device for slot 0 of /dev/md2 > mdadm: no uptodate device for slot 2 of /dev/md2 > mdadm: added /dev/sdl1 to /dev/md2 as 3 > mdadm: added /dev/sdi1 to /dev/md2 as 4 > mdadm: added /dev/sdk1 to /dev/md2 as -1 > mdadm: added /dev/sdj1 to /dev/md2 as 1 > mdadm: /dev/md2 assembled from 3 drives and 1 spare - not enough to start the array. Well, darn. I was hoping the "spare" status was an artifact of a device held in a partially assembled array. But no. At this point, --create --assume-clean is the next step. You *must* specify chunk size and data offset matching the --examine reports for your drives. You must place the good devices in the order shown by their roles, not trusting the device names. Per the last assembly attempt, sdi1, sdj1, and sdl1 are roles 4, 1, and 3, respectively. Based on the old data in the superblock, sdk1 is likely to be role 0, but it could be role 2 if any names have changed. If you know from drive serial numbers/smartctl that sdm is truly the failing drive, then leave it out. Use "missing" in its place in the command. Do an fsck -n before mounting to verify that the role order you've re-created is correct. Phil
