Disk exhaustion on a Windows VPS rarely arrives quietly. SQL Server reports error 9002 (“The transaction log for database is full”), Windows Update fails with 0x80070070, or IIS starts returning 500.50 because the request-tracing log directory is full. On KVM-based plans the fix does not require a reinstall or a restore from backup: you grow the virtual disk at the provider, then extend the volume inside Windows. The whole process takes about thirty minutes plus a reboot.
Two preconditions decide how this goes. The disk must be GPT rather than MBR if you intend to exceed 2 TB, and the unallocated space must sit immediately after the volume you want to grow. Both are checkable in a single command, and both are fixable — but you want to know before you start, not halfway through.
Diagnose Before You Buy More Storage
Run these three commands from an elevated PowerShell session. Get-Volume shows free space per volume, Get-Disk | Select-Object Number, FriendlyName, Size, PartitionStyle shows the partition table type, and Get-Partition -DriveLetter C | Get-PartitionSupportedSize tells you exactly how much room you could extend into if unallocated space existed.
Then find what is actually consuming the drive. The usual offenders on a Windows VPS are C:\Windows\SoftwareDistribution\Download after a failed update cycle, C:\inetpub\logs\LogFiles with per-site W3C logs that were never rotated, the SQL Server backup directory sitting on the system volume, and C:\Windows\Temp filled with installer residue. Get-ChildItem C:\ -Directory | ForEach-Object { [pscustomobject]@{Name=$_.Name; GB=[math]::Round((Get-ChildItem $_.FullName -Recurse -Force -ErrorAction SilentlyContinue | Measure-Object Length -Sum).Sum/1GB,2)} } | Sort-Object GB -Descending | Select-Object -First 10 gives a top-ten list in under a minute.
Reclaim Space That Is Wasted by Definition
Two built-in operations routinely free 5–15 GB on a server that has been through several update cycles. The first is component store cleanup, which removes superseded Windows Update payloads from WinSxS:
Dism.exe /Online /Cleanup-Image /StartComponentCleanup /ResetBase
The /ResetBase switch prevents uninstalling previously installed updates, which is acceptable on a stable production host but should be skipped if you are mid-way through troubleshooting a bad patch. The second is the standard cleanup profile, which clears update cache, old Windows installations, and delivery optimisation files: cleanmgr /d C: /sagerun:1 after configuring the profile with cleanmgr /d C: /sageset:1.
If those recover enough space, stop — you have avoided both the cost of additional storage and the reboot. If they do not, move IIS logs and database backups to a second disk first, because they will consume any new space you buy within months.
Step 1: Grow the Virtual Disk at the Provider
The virtual disk is a file on the host, so no amount of work inside Windows will enlarge it. Depending on the panel — SolusVM, Virtualizor, Proxmox, or a provider-specific interface — locate the disk or storage section and raise the size, for example from 60 GB to 120 GB. Shut the VPS down cleanly first with shutdown /s /t 0. Some platforms support online resize, but a full stop removes any risk of the guest filesystem caching a stale block count.
After booting, Windows Storage Management should see the larger disk immediately. Verify with Get-Disk: the Size column should reflect the new capacity while Get-Partition still reports the old partition size, with unallocated space after it. If the disk size has not changed, the guest may need a SCSI rescan — Update-HostStorageCache, then re-check. On some KVM setups you must also detach and reattach the volume from the panel’s device list.
Step 2: Extend the Partition Inside Windows
With contiguous unallocated space in place, extension is a single command. Modern Windows Server can extend the system volume online, without a reboot:
Resize-Partition -DriveLetter C -Size (Get-PartitionSupportedSize -DriveLetter C).SizeMax
Confirm with Get-Volume C. If the command fails with “The parameter is incorrect” or “There is not enough space available on the disk”, the unallocated space is not adjacent to C: — typically because a recovery partition sits between them. Check with Get-Partition -DiskNumber 0. The recovery partition is usually a few hundred megabytes and can be deleted with reagentc /disable followed by Remove-Partition, then recreated at the end of the disk with reagentc /enable after the extension.
For an MBR disk that needs to cross the 2 TB boundary, convert first. mbr2gpt /validate /disk:0 /allowFullOS checks feasibility, and mbr2gpt /convert /disk:0 /allowFullOS performs the conversion. Both require a full backup beforehand and a UEFI-capable boot configuration.
The Second-Disk Alternative
When the panel has no resize option, or the plan’s disk is already at its maximum, attach an additional data disk instead. This is often cheaper per gigabyte and has a side benefit: logs and backups live on a volume that can be snapshotted independently of the system drive.
After attaching, initialise and format it: Get-Disk | Where-Object PartitionStyle -eq 'RAW' | Initialize-Disk -PartitionStyle GPT -PassThru | New-Partition -AssignDriveLetter -UseMaximumSize | Format-Volume -FileSystem NTFS -NewFileSystemLabel "Data" -AllocationUnitSize 65536 -Confirm:$false. A 64 KB allocation unit is the right choice for SQL Server data files and large IIS logs; 4 KB is better for a volume holding thousands of small files. Our article on NTFS allocation unit size for SQL Server covers the trade-off in detail.
Prevent the Next Outage
Resizing is a fix, not a policy. Set IIS log rotation to a fixed schedule and a size cap in applicationHost.config, move SQL backups to the data disk with a full/differential/log retention ladder, and monitor free space with a scheduled task that writes to the event log when a volume drops below 15 percent. Windows has a built-in mechanism for the last part: a performance alert or a simple PowerShell task running Get-Volume -DriveLetter C | Where-Object SizeRemaining -lt 15GB.
Also account for the page file. If Windows manages it dynamically, it will try to grow into free space during a memory spike and fail if the volume is 99 percent full, producing out-of-memory errors that look like an application bug. Set a static page file sized to 1.5x RAM and put it on the data disk if that disk is SSD-backed.
| Symptom | Likely cause | Fix |
|---|---|---|
| Resize-Partition fails, space shown as unallocated | Recovery partition sits after C: | reagentc /disable, delete partition, extend, re-enable |
| Cannot extend past 2 TB | MBR partition table | mbr2gpt /convert after full backup |
| Get-Disk still shows old size | Guest has not rescanned | Update-HostStorageCache, or detach/reattach volume |
| Space refills within weeks | IIS logs or SQL backups on system volume | Move to data disk with retention policy |
Planning the Right Starting Size
Most resize requests come from instances provisioned at 40 or 60 GB for a workload that needed 100 GB from day one. A realistic minimum for a Windows Server host running IIS and SQL Server Express is 80 GB for the system volume plus a separate 100 GB data volume. Adding the data disk at provisioning time costs less than a mid-life resize that requires a maintenance window and a reboot.
If you are choosing a plan now, factor in SQL Server’s growth rate and the Windows component store’s steady expansion of roughly 1–2 GB per year of patching. Our Windows VPS plans list disk sizes and allow additional data volumes, and our team can advise on a layout before you migrate.



