I’ve setup 3 node cluster which is working fine,If i’m taking one node offine(removing) ,I’m expecting all the container running in faulty nodes should get moved to other active node,however this is not happening,Is there any limitation in LXC/LXD?
There’s no such feature at the moment.
Ok,Is there any work around to do this ? I mean Clustering does mean fault tolerant ,Is there any way i can achieve this?
would you tell me in which case i can use LXC/LXD clustering and can make use of this feature?
Am looking for some kind of fault tolerant with LXC/LXD clustering not sure if this is possible with the existing feature of LXC clustering ?
You can use the REST API or the command line client to program any automatic failover logic you wish to have. In order to have the state of your containers survive after a node dies, you’ll most probably want to use ceph as backend for your storage pool.