Knowledge Base
Troubleshooting and solution know-how — search, or ask AI Assistant.
Management node and browser times differ because of clock driftWhen management node and browser times differ, stop the management node service, configure an external NTP source, synchronize the local clock, and write it to the hardware clock.Updated: 2026-09-17Management node CPU usage is high after zops tasks become stuckTroubleshoot persistent high CPU usage on a management node by identifying accumulated ansible-playbook processes under zops, then restart the zops service.Updated: 2026-09-17Resource operations fail due to lvmlock result -28 on SharedBlock storageDescribes how to diagnose resource operation failures after SharedBlock storage reports lvmlockctl result -28, apply temporary recovery steps, and upgrade to a version with the long-term fix.Updated: 2026-09-17VM HA does not trigger while a host is unresponsive but its storage heartbeat remains aliveTroubleshoot why VM HA does not trigger after a host becomes unreachable and determine whether a responsive storage heartbeat prevents the HA trigger conditions from being met.Updated: 2026-09-17Troubleshooting intermittent "Cannot assign requested address" errors when a VM accesses object storageTroubleshoot intermittent `Cannot assign requested address` errors caused by a Linux guest opening many short-lived connections to an object storage endpoint.Updated: 2026-08-17Troubleshooting snapshot deletion when control-plane and data-plane chains are inconsistentResolve a "could not find image ... in chain" failure when deleting a volume snapshot by reconciling the libvirt/QCOW2 backing chain with the VolumeSnapshotVO records and removing the stale DB row.Updated: 2026-08-17