Hi, Experts,
we have an cephfs cluster 16.2.* run with multi active mds, and we have some
old machine run with ubuntu 16.04 , so we mount these client using ceph-fuse.
After a full mds process restart, all of these old ubuntu 16.04 clients cannot
connect to ceph , `ls -lrth` or `df -hT` hang on the client node.
and we can see mds log said `evicting unresponsive client *** after waiting **
seconds during restart`
also checking the client node, no ceph-fuse process, (ps -ef | grep ceph or ps
-ef | grep fuse).
but there is a remount process (ps -ef | grep mount) in the client node.
root 176516 1 0 01:38 ? 00:00:00 mount -i -o remount /data
cannot kill this process using `sudo kill -9 176516`.
we really need experts help us how to save this client out of hang. we cannot
restart this client node due to there are some other critical service.
Thanks a ton.
zx
_______________________________________________
ceph-users mailing list -- ceph-users@ceph.io
To unsubscribe send an email to ceph-users-le...@ceph.io
