GitHub user prashanthr2 added a comment to the discussion: Host in Alert state after CloudStack Upgrade to 4.20.2.0
@Shivamkashyap08 The pending host.call_plugin tasks on the Xenserver host suggest the issue is on the hypervisor/toolstack side. Please check /var/log/xensource.log on the affected host around the same timestamps to see what those plugin tasks are waiting on or failing against. As a recovery step, you can try xe-toolstack-restart on the affected host. Restarting the toolstack does not affect running VMs, but it does restart the management stack, temporarily disconnect the host from management, and cancel in-flight management tasks such as stuck plugin/API operations. If the pending tasks are cleared after the toolstack restart, check whether the host reconnects cleanly to CloudStack. If it does not, you can then consider unmanaging and managing the cluster/host from CloudStack to force re-registration and reconnection. I would keep that as a secondary step after reviewing xensource.log and trying the toolstack restart, because unmanage/manage changes CloudStack’s control state, while guest VMs can continue running even when management connectivity is disrupted. GitHub link: https://github.com/apache/cloudstack/discussions/13017#discussioncomment-16585403 ---- This is an automatically sent email for [email protected]. To unsubscribe, please send an email to: [email protected]
