Hi Community, I am testing High Availability (HA) for Cloudstack VMs using KVM and NFS. To simulate a host failure, I am triggering a kernel panic. When the command is executed, the host status remains "up" for 3-4 minutes before transitioning to an "Alert '' state and not marked as "Down ". As a result, the VMs are not being migrated and still appear to be running on the downed host server.
My questions are: 1. Can we reduce the detection time for a downed host to about 30 seconds? 2. What could be the reason for hosts entering an "Alert" state after a kernel panic, which is preventing VM HA from functioning properly? Alex.
