Yeah, the message

[ 171.485542] ceph: No mds server is up or the cluster is laggy

can be misleading sometimes. But I'd recommend to create a new thread since the original issue is resolved. Zitat von Jacek Rużyczka via ceph-users <ceph-users@ceph.io>:

...

It seems like removing packages did the trick, and there are up OSDs.

Apparently yes. I've already placed some test files in the cluster, and the test went well. I've also added the cluster as a mount in /etc/fstab. The mount itself was performed after mount -a, but not on startup.

There is only one thing I've just found in the syslog at startup:

[ 150.710312] ceph: loaded (mds proto 32) [ 150.745343] libceph: mon1 (1)10.20.0.12:6789 socket closed (con state V1_BANN ER) [ 151.016656] libceph: mon1 (1)10.20.0.12:6789 socket closed (con state V1_BANN ER) [ 151.520015] libceph: mon1 (1)10.20.0.12:6789 socket closed (con state V1_BANN ER) [ 157.642865] libceph: mon3 (1)10.20.0.14:6789 session established [ 157.645794] libceph: client274103 fsid 8aad3073-39a1-11f1-bf6e-f2704a1efa9b [ 157.646023] libceph: mds0 (1)10.20.0.11:6801 socket error on write [ 157.919274] libceph: mds0 (1)10.20.0.11:6801 socket error on write [ 158.426046] libceph: mds0 (1)10.20.0.11:6801 socket error on write [ 171.485479] ceph: mds0 reconnect start [ 171.485542] ceph: No mds server is up or the cluster is laggy

This contradicts the output of ceph -s, which still says, HEALTH_OK.

Am Fr., 24. Apr. 2026 um 00:38 Uhr schrieb Anthony D'Atri via ceph-users < ceph-users@ceph.io>:

...

_______________________________________________ ceph-users mailing list -- ceph-users@ceph.io To unsubscribe send an email to ceph-users-leave@ceph.io

_______________________________________________ ceph-users mailing list -- ceph-users@ceph.io To unsubscribe send an email to ceph-users-leave@ceph.io