[SOLVED] cluster ceph very slow when one node is offline

@tabakoff, are you talking about the same cluster?

Thanks for your help. We did our pool 4/3 size, but it still doesn't worked like we expected.
size 3 / min_size 2 is the default. Usually three copies are enough.

We have enabled igmp and igmp snooping (is there anything in common with our problem) on our switch (Extreme Networks Summit X670-G2-72x) and then did this:
Corosync 2.x uses multicast, while Corosync 3.x doesn't. And Ceph doesn't use multicast either. So this seems not related to any ceph issues.

We disable ens2f1 (cluster interface 10G) and put the command on other node:
About which cluster interface are you talking about? Corosync / Ceph?