Ceph plus docker swarm?

Question

PostgreSQL

Good afternoon
I have a task to build a fault-tolerant postgresql on docker swarm.
As volume storage, I'm trying to use a ceph cluster of 3 virtual machines.
Everything is more or less running.
When I check for a crash in one of the ceph nodes, my mount on the docker swarm machines hangs for 1-5 minutes.
Is this normal behavior for cephfs? In the logs, I see reconnection to other mon nodes.
I was counting on the failure handling in a few seconds.
It happened on gluster.

Answer 1

No, the behavior is not normal, but the cluster needs to be built differently - such tasks will be solved at the postgres level.
In addition, ceph is best put on real hardware, virtual machines kill it even more and so low performance.
well, three nodes is very little.
According to your problem - did you specify all the monitors during the mount?