About this article: This article is reviewed against the cited sources. This is source research, not lab validation. Check the stated versions and your own requirements before making changes.
Hyper-V pauses a virtual machine when the volume that holds its virtual hard disk files, or its checkpoints, has no free space. The Microsoft-Windows-Hyper-V-VMMS-Admin log records Event ID 16060 with the description <virtual machine> has been paused because it has run out of disk space on '<path>', and the machine shows a state of "Critical - Paused". Microsoft's resolution is two lines long: "free disk space on these hard drives or move these VHD files to a new location." The safe order is below.
What the error means
Microsoft's page states the cause as: "The hard drives that store these VHD files or snapshots of these virtual machines are out of disk free space." The page describes the scenario on Windows Server 2008 and 2008 R2 with System Center Virtual Machine Manager (VMM), where the VMs also fail to connect from the VMM console with "Unable to connect to the Virtual machine Manager Server vmm. The attempt to connect timed out". The page's header applies it to "Supported versions of Windows Server". Event 16060 and the cause are the useful parts on a current host.
A related case applies to Cluster Shared Volumes (CSV). If the CSV runs out of space, the VMs on it pause, but "the Virtual Machines on the other nodes of the cluster that do not own the CSV are not affected." Before that, the node that owns the CSV logs a warning, Event ID 16050, with the description <virtual machine> is about to run out of disk space on '<path>'. Microsoft explains: "The node that owns the Cluster Shared Volume keeps checking the disk space and warns. When the warning is received, the Virtual machine's go into the paused critical state." Microsoft adds that a node that does not own the CSV "will not have any ways to determine if it is running out of space on CSV volume", so it lets its VMs keep running. While a VM is paused, the services inside it are unavailable until it resumes, so treat the pause as an outage even though the virtual machine has not been shut down.
Likely causes
| Cause | What you see | How to confirm |
|---|---|---|
| The volume holding the VM files is full | Several VMs on one volume are "Critical - Paused" together; Event ID 16060 | Get-Volume shows almost no SizeRemaining on that volume |
| Checkpoint files have filled the volume | The VM folder holds large .avhdx files; the VM has checkpoints |
Get-VMSnapshot -VMName lists checkpoints; the folder shows .avhdx files |
| A Cluster Shared Volume is full | VMs owned by one node pause; VMs on other nodes keep running; Event ID 16050 earlier | Get-ClusterSharedVolume and the volume's free space; the owner node's event log |
| Other data shares the volume | The volume fills although the VM files did not grow | A folder-size check shows non-VM files |
Prove the cause before changing anything
All read-only. Run these on the Hyper-V host (on the node that owns the volume for a CSV).
- Record which VMs are affected and their state:
Get-VM | Where-Object {$_.State -ne 'Running'}
- Find the events. Microsoft names the log
Microsoft-Windows-Hyper-V-VMMS-Adminand event IDs 16060 and 16050:
Get-WinEvent -FilterHashtable @{ LogName='Microsoft-Windows-Hyper-V-VMMS-Admin'; Id=16060,16050 }
The path in the description names the volume that is out of space.
- Check the free space on that volume (replace
Ewith its drive letter):
Get-Volume -DriveLetter E
The output includes SizeRemaining and Size.
- For a CSV:
Get-ClusterSharedVolume
The cmdlet "Gets information about Cluster Shared Volumes in a failover cluster." Then find which node owns the volume, using the cluster tools you already use, and read the events on that node.
- List checkpoints for each affected VM:
Get-VMSnapshot -VMName VMName
- In File Explorer, open the VM's folder and note the size of each
.vhdxand.avhdxfile. Do not change anything.
Example (illustrative): a three-node cluster at contoso.com.au has 14 VMs on a CSV, C:\ClusterStorage\Volume1, owned by hv01. Event ID 16050 appears on hv01 on Monday evening; by Tuesday 06:00 Event ID 16060 appears. Six of the 14 VMs run on hv01 and show "Critical - Paused". The other eight, running on hv02 and hv03, keep working. Six of 14, about 43 per cent, are paused. Get-VMSnapshot shows a checkpoint on fs01 taken six weeks ago, and its .avhdx is the largest file on the volume.
Fix
Choose the least destructive action that frees enough space. Start with files that are not VM files.
Free space on the volume
Save first: copy anything you are unsure about to another volume, and write down what you removed.
Risk: deleting files in the next step cannot be undone.
Remove files you can name that are not part of any VM: old installers, exported copies, abandoned backups. Do not delete .vhdx, .avhdx or configuration files of a VM.
Delete old checkpoints properly
Microsoft's checkpoints page says "Deleting checkpoints can help create space on your Hyper-V host" and warns: "You should not delete the .avhdx files directly." When you delete a checkpoint, "Hyper-V merges the .avhdx and .vhdx files for you." Do the three save-first steps below before the delete command.
- Record the checkpoints and save the output to a file on a volume that has free room, not the full one (create
D:\Exportfirst if it does not exist):
Get-VMSnapshot -VMName VMName | Out-File D:\Export\checkpoints-before.txt
- Export the VM to a different volume that has more free space than the VM's files occupy. Microsoft's cmdlet page describes
Export-VMas exporting a virtual machine to disk, with a Snapshots folder, into the folder given by-Path:
Export-VM -Name VMName -Path D:\Export
Do not export to the full volume. Microsoft's cmdlet page does not say whether Export-VM works on a VM in the Critical-Paused state; if the export fails, or if no other volume has the room, take a backup with your backup product instead, and confirm it can be restored. Do not go on to the delete without one of the two.
- Check that the volume still has free room for the merge (
Get-Volume, as above).
Only then run the delete. Read the mark first.
Disruptive step: deleting a checkpoint removes that restore point and merges its changes into the disk. Impact: you can no longer revert the VM to that point, and the merge takes time and disk activity. Back-out: the checkpoint itself cannot be recreated, so the restore path is the export or backup from step 2. Microsoft's Import-VM page says the cmdlet imports a virtual machine from a file, and that -Copy -GenerateNewId copies the files and gives the imported machine a new unique identifier, which lets it sit beside the original. Use the exported configuration file from the export's Virtual Machines folder (it is named with the VM's ID):
Import-VM -Path 'D:\Export\VMName\Virtual Machines\VM-ID.vmcx' -Copy -GenerateNewId
Replace VM-ID with the file name you find there. Microsoft's Export-VM page says the export's Snapshots folder holds the associated snapshots, so the imported copy should hold the checkpoint as it was at the export; check that after the import. The delete command is:
Remove-VMCheckpoint -VMName VMName -Name "checkpoint name"
Microsoft's page does not say how much free space the merge needs, so make some room by the first step before you start the merge on a nearly full volume.
Move the VM's storage
Microsoft's other option is to move the files. The Move-VMStorage cmdlet "moves the storage of a virtual machine", and its first example moves all files of a VM to one destination. Do not run it yet. First record where the files are now, and save the record to a volume that has room (not the full one). The Format-List page shows Format-List -Property *, which lists every property of the VM, including its file locations:
Get-VMHardDiskDrive -VMName VMName | Out-File D:\Export\vmname-disks-before.txt
Get-VM -Name VMName | Format-List -Property * | Out-File D:\Export\vmname-locations-before.txt
Read the saved file and note the path of the configuration, checkpoint and smart-paging locations, as well as each disk path.
Risk: the move copies the whole VM, so it needs free space on the destination and time, and it puts load on both volumes. Impact: the VM's disk activity and the host's storage load rise while it runs. Microsoft's cmdlet page does not say whether the VM must be off, or whether a paused VM can be moved, so try it first on a VM that can be spared. Back-out: the single-folder command below puts every file in one folder, so it can reverse the move only if the files all sat in one folder before. If the saved records show disks, checkpoints or configuration in different places, move them back one kind at a time with the cmdlet's -VHDs, -VirtualMachinePath and -SnapshotFilePath parameters, using the source and destination paths from the saved records; Microsoft's page shows the -VHDs form with SourceFilePath and DestinationFilePath for each disk. The move itself, with the example's form and your own destination, is:
Move-VMStorage "VMName" -DestinationStoragePath D:\VMName
To move everything back to one folder on the original volume, the same form applies with that volume as destination:
Move-VMStorage "VMName" -DestinationStoragePath E:\VMName
Moving back works only if the original volume has the space. If the original volume is the one that is full, do not move back until space is freed.
Before you move anything, check that the destination volume has more free space than the VM's files occupy, using Get-Volume as above. A move to a volume that is nearly as full as the source only moves the problem.
If the full volume is a Cluster Shared Volume
Work on the volume, not on one node: the pause hits the VMs owned by the node that owns the CSV, and moving a VM to another node does not create space on the volume. Free space or move VM storage off the CSV as above. After that, check that each node's VMs are in the state you expect, including those on nodes that were never paused.
Add space
Adding space to the volume or the CSV is a storage task for whoever owns the storage, and falls outside the commands here. Microsoft's page names no command for it.
Resume the virtual machines
After space is free, check the state again. If a VM is still paused, Resume-VM "resumes a suspended (paused) virtual machine".
Risk: a VM resumed onto a volume that is still almost full pauses again. Check Get-Volume first, and resume the most important VM before the others so you can see that it holds.
Resume-VM -Name VMName
Confirm the fix
Get-Volumeshows a comfortableSizeRemainingfor the volume.Get-VMshows each VM as Running.- No new Event ID 16060 appears after the resume.
- The applications inside the VMs answer, and any clients that dropped have reconnected.
- For a CSV, the event log of the owning node is quiet for several hours.
What to watch afterwards
- Free space on every volume that holds VM files. Microsoft's page lists Event ID 16050 as the warning that precedes the pause on a CSV, so it is worth alerting on.
- Checkpoints that stay for weeks. Microsoft's checkpoints page says they are stored as
.avhdxfiles beside the.vhdxfiles. - Growth of the VM files over time, so the volume is enlarged before it fills.
- Backup jobs that create temporary checkpoints on a volume with little room.
Read next
Related articles you may want next.
- TroubleshootingHyper-V virtual machine fails to start: General access denied error (0x80070005)Hyper-V VM won't start: 'General access denied error' 0x80070005 or 'failed to initialize'? Add the VM ID to the file permissions with icacls, safely.Read article
- TroubleshootingHyper-V backup checkpoint fails: Error -0x800423F2, error code 32768 and VSS writer events 8229 and 12293Hyper-V backup says it could not create the backup checkpoint (-0x800423F2) or VSS events 8229 and 12293? Prove the cause, then fix it safely.Read article
- Backup & RecoveryDefragmenting a Windows server made the next backup huge: what about shadow copies?Defrag made the next incremental backup huge? Why it happens, how to confirm it, what is documented on shadow copies, and a safe schedule.Read article
Risks and limits
These commands have not been run or lab-tested.
- Microsoft's paused-state page describes Windows Server 2008 and 2008 R2 with VMM. It gives two event IDs, a cause and a two-line resolution, and no free-space threshold. The remaining detail here comes from Microsoft's checkpoint and cmdlet pages.
- Microsoft's paused-state page does not describe how the pause is cleared on a current host beyond freeing space.
Resume-VMcomes from its cmdlet page. - A merge after deleting a checkpoint can need space and time. Do not start one with no free space.
- Moving storage while a host is under load can slow other VMs.
Get-ClusterSharedVolumeshows state; Microsoft's cmdlet page does not name the property that holds the owner node.- Check the current Microsoft pages before relying on event descriptions.
Sources
- Virtual machines enter the paused state due to low disk free space — Microsoft Learn, page dated 2026-02-12, accessed 20 November 2026
- Using checkpoints — Microsoft Learn, page dated 2025-08-14, accessed 20 November 2026
- Get-VM — Microsoft Learn, date not shown, accessed 20 November 2026
- Get-WinEvent — Microsoft Learn, date not shown, accessed 20 November 2026
- Get-Volume — Microsoft Learn, date not shown, accessed 20 November 2026
- Get-ClusterSharedVolume — Microsoft Learn, date not shown, accessed 20 November 2026
- Get-VMSnapshot — Microsoft Learn, date not shown, accessed 20 November 2026
- Format-List — Microsoft Learn, date not shown, accessed 20 November 2026
- Move-VMStorage — Microsoft Learn, date not shown, accessed 20 November 2026
- Get-VMHardDiskDrive — Microsoft Learn, date not shown, accessed 20 November 2026
- Export-VM — Microsoft Learn, date not shown, accessed 20 November 2026
- Import-VM — Microsoft Learn, date not shown, accessed 20 November 2026
- Out-File — Microsoft Learn, date not shown, accessed 20 November 2026
- Resume-VM — Microsoft Learn, date not shown, accessed 20 November 2026
This article was originally published on .