we have an old linux host that it lost it's boot disk and won't boot up anymore. What we did is we moved those luns to another linux host. There are 9 luns all together. from the other linux host, we did a rescan:
unfortunately, we don't have any backup for that server that died. It was a legacy system and we were supposed to decomm this server last week but we realized we need to archive some files from it but it was too late.
output from "pvs --segments" does not provide any hints ? We swing devices from one system to another often when we do system refreshes/backup (moving snapshots). They are graceful though (unmount, vg de-activate, vg export/ vgimport, vg-activate, mount). I am just wondering if lvm header got corrupted or something.
i don't know what else I can do. Usually, after rescan and you do a pvscan, it should show all the VGs with all the volumes. in this case, a VG showed up with only 3 volumes and I don't know what happened to the 6 of the luns.
i don't have that info actually because the original server died before I got to it but our DBAs said there are 2 other mount points that starts with Ora*. I am just hoping after presenting those 9 luns to another host of similar OS and LVM, the VGs will just come up.
Did you match the HLU/ALU numbers to match what it was in the old Storage Group or did you just add a new host to the existing Storage Group? Maybe the order of the HLU numbers changed (I know nothing about Unix)? Did you check that the new host has connection to each LUN in the Storage Group from all paths?
dynamox
11 Legend
•
20419 Posts
•
87439 Points
745
0
Posted May 10th, 2015 19:00
assuming the host was being backed up, anything of use in /etc/lvm directory (if you can restore that) ?