Stale NFS handles? Try lsof|grep MD instead?
Trevor Cordes trevor@tecnopolis.ca wrote:
Very weird problem here:
I umounted the filesystem on a md (RAID1) device.
Then:
mdadm --stop /dev/md1 mdadm: Cannot get exclusive access to /dev/md1:Perhaps a running process, mounted filesystem or active volume group?
Huh?
lsof /dev/md1 <nothing!>
I remounted the fs and tried lsof <fs> and it was empty too.
I can't seem to find any reason why mdadm --stop should not allow stopping this array. Any ideas? Anything other than lsof that can help find out why? _______________________________________________ Roundtable mailing list Roundtable@muug.mb.ca http://www.muug.mb.ca/mailman/listinfo/roundtable
On 2013-03-16 Adam Thompson wrote:
Stale NFS handles? Try lsof|grep MD instead?
No NFS on this box.
However, you gave me an idea to look in ps:
root 361 2 0 16:08 ? 00:00:00 [md1_raid1] root 768 2 0 16:08 ? 00:00:00 [jbd2/md1-8]
So those might be keeping md1 "busy"? Or are those the very things that mdadm --stop should halt? The md1_raid1 is familiar but I've never seen jbd2/md1-8 before.
-
Adam Thompson -
Trevor Cordes