VMware Troubleshooting: Stuck Snapshots, Consolidation Alerts and Datastores That Fill Up
Why snapshots balloon, how to consolidate safely, and what to check when a datastore hits 100 percent.
Snapshot problems
A previous snapshot deletion left delta chains behind - often after backup jobs or a crash. Right-click the VM, Snapshot, Consolidate, but do it in a change window: merging deltas spikes I/O and needs datastore headroom. If it stalls, check datastore latency and free space, then retry. Repeated alerts after backups point at the backup proxy timing out on snapshot removal - raise the timeout or fix in-guest VSS.
Large or old snapshots take a long time to merge; do not cancel mid-merge. Watch tasks and events in vCenter and disk latency in esxtop. Keep snapshots under 72 hours old and no deeper than two or three deltas. On NFS datastores also check for locking issues and saturation.
Datastore capacity
Free space fast: delete old snapshots first (the biggest hidden consumer), then storage vMotion the largest VMs off. Thin-provisioned VMDKs and growing vmware.log files also consume space. Set alarms at 80% and keep 10-15% headroom for snapshots and swap files.
Check for unresolved SCSI reservations, snapshots counted on the array side, and thin versus thick provisioning differences. Rescan storage adapters after array-side changes, verify the multipath policy is round robin, and confirm all paths are active with esxcli storage core device list.