Hello,We are running a Octopus cluster however we still have some older Ubuntu 16.04 clients connecting using libcephfs2 version 14.2.13-1xenial.
From time to time it happened that the network was having issues so the clients lost the connection to the cluster. But the system still thinks the mount is running and it has to be restarted manually. In some cases we even had to restart the complete machine because it would refuse to unmount.
I also attached a dmesg log so you can see the systems behaivior. How do you deal with such issues? Kind regards, Julian Fölsch -- Julian Fölsch Arbeitsgemeinschaft Dresdner Studentennetz (AG DSN) Telefon: +49 351 271816 69 Mobil: +49 152 22915871 Fax: +49 351 46469685 Email: julian.foelsch@xxxxxxxx Studierendenrat der TU Dresden Helmholtzstr. 10 01069 Dresden
[1366901.940605] libceph: mon2 10.144.0.4:6789 session lost, hunting for new mon [1366937.780384] ceph: mds0 caps stale [1367164.851326] ceph: mds0 hung [1367819.440486] libceph: mds0 10.144.0.3:6801 socket closed (con state OPEN) [1367950.511242] libceph: mds0 10.144.0.3:6801 socket closed (con state CONNECTING) [1368016.048263] libceph: mds0 10.144.0.3:6801 connection reset [1368016.048578] libceph: reset on mds0 [1368016.048588] ceph: mds0 closed our session [1368016.048589] ceph: mds0 reconnect start [1368016.223200] libceph: mds0 10.144.0.3:6801 socket closed (con state NEGOTIATING) [1368016.752349] ceph: mds0 rejected session [1368024.244740] libceph: mon0 10.144.0.2:6789 session established [1369816.744074] libceph: mds0 10.144.0.3:6801 socket closed (con state OPEN)
_______________________________________________ ceph-users mailing list -- ceph-users@xxxxxxx To unsubscribe send an email to ceph-users-leave@xxxxxxx
