Check ESXi datastore capacity and free space
A storage alert deserves a closer look before you start moving or removing files. First, confirm which datastore is affected and what the host can see.
Symptoms
- A datastore capacity alarm appears in vCenter.
- A VM task fails with a storage-space error.
- A datastore unexpectedly shows zero capacity or is unavailable.
Before you start
This guide uses ESXi 7.x/8.x command-line inspection. Use an authorised host administrator account and the ESXi Shell or SSH. If needed, follow Enable SSH on ESXi. The command below lists information without changing the datastore.
Resolution
1. Identify the datastore in the client
In the vSphere Client, select the affected datastore and record its name, capacity and free space. Note which host and VM were involved in the failed task. Keep the original error message too.
2. Check the host's filesystem view
esxcli storage filesystem list
Match the Volume Name and UUID to the affected datastore. Read Mounted, Type, Size and Free together. Size and free-space values are in bytes; divide by 1073741824 to express them in GiB. Broadcom documents these fields in its ESXi datastore listing example.
3. Decide what to investigate next
| Observation | Next step |
|---|---|
| Mounted with little free space | Review capacity growth, recent tasks and storage requirements with the workload owner. |
| Plenty of free space, but a VM task fails | Confirm the task targets this datastore and inspect its exact error. |
Zero capacity or Mounted is false | Investigate datastore accessibility and mount state before treating it as a full volume. |
Broadcom describes a case where unmounted VMFS datastores show zero capacity. That symptom needs storage investigation; the display alone doesn't identify the underlying cause.
Check ownership and dependencies first. Avoid deleting VM disks or snapshot files directly from the datastore browser or shell. Storage cleanup and snapshot consolidation need their own planned procedure.
Verify
Run the listing again and refresh the datastore view in the client. Confirm that you're comparing the same datastore UUID. If the values disagree, capture both readings and the host name for further investigation.
When finished, disable SSH if you enabled it for this check. This inspection helps narrow down a storage problem; it doesn't measure latency, physical array headroom or vSAN policy health.