On Wed, 2022-04-20 at 12:09 -0400, john tillman wrote: > Here's the scenario: > I have a two node cluster setup. (details to follow) > I start with both nodes booted. > I can reboot the active node and the resources move. >
By default, corosync's two_node mode enables wait_for_all, which means that when a node first starts up, it has to see the other node at least once before it will claim to have quorum. This is the only safe mode of operation, otherwise two nodes that get severed from each other and then reboot will both claim quorum. However you can override that by setting wait_for_all to false in the corosync configuration. This is mainly useful when you are intentionally and manually bringing the cluster down and want to bring up a single node. Once the node is up you can set wait_for_all back to true for safety, so if it reboots unexpectedly it will require manual intervention again to get going. -- Ken Gaillot <[email protected]> _______________________________________________ Manage your subscription: https://lists.clusterlabs.org/mailman/listinfo/users ClusterLabs home: https://www.clusterlabs.org/
