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.
When the System log records a VolSnap event 25, the volume's shadow copies have been aborted because the shadow copy storage area, also called the diff area, could not grow in time while the volume was being written to heavily. The event refers to the shadow copies of the volume in the plural, so assume all of them are gone; a backup that was reading one of them would also lose its source (the event text does not say this, so treat it as a likely consequence). Events 33, 35 and 36 are close relatives with different causes: a user-set size limit, a failed growth, or a limit that blocked growth. Find which one you have before you change anything, because the fix for each differs and some fixes delete the shadow copies that remain.
What the error means
A shadow copy is not a full copy of the volume. Microsoft's description of the diff area is that it is "the storage space" allocated on a volume "for maintaining the snapshots", and that a snapshot is "a block-level set of information that represents the differences between the current content and content from a previous point in time". The integrity of existing and new snapshots depends on the integrity of that area (Diff Area Integrity, Windows Server 2008 R2). When writes arrive faster than the area can be extended, or the area has nowhere to extend to, the system cannot keep the snapshot consistent.
By default, Windows protects the live volume, not the shadow copy. The VSS_PROTECTION_LEVEL_ORIGINAL_VOLUME level is the default, and it "specifies that I/O to the original volume must be maintained at the expense of shadow copies". Shadow copies might be deleted if a write to the original volume occurs and the integrity of the shadow copy cannot be maintained, for example because of a failure to write to the storage area or to allocate memory (VSS_PROTECTION_LEVEL). Event 25 is that trade-off made visible.
The events, from Microsoft's table
All rows are from the Windows Server 2008 R2 reference (version 6.1), source VolSnap:
| Event | Message (shortened from the 2008 R2 reference) | Meaning |
|---|---|---|
| 24 | Insufficient disk space on volume %3 to grow the diff area for shadow copies of %2; all shadow copies are at risk of being deleted | Storage volume is full |
| 25 | The shadow copies of volume %2 were aborted because the diff area file could not grow in time. Consider reducing the IO load on this system | Growth too slow for the write load |
| 33 | The oldest shadow copy of volume %2 was deleted to keep disk space usage for shadow copies of volume %2 below the user defined limit | Routine trimming against a maximum size |
| 35 | The shadow copies of volume %2 were aborted because the diff area file failed to grow | Growth failed |
| 36 | The shadow copies of volume %2 were aborted because the diff area file could not grow due to a user imposed limit | Maximum size reached |
| 38 | A user imposed limit prevented disk space on volume %3 from being used to grow the diff area; all shadow copies at risk | Limit about to cause loss |
| 41 | The shadow copy storage on volume %3 did not have sufficiently large contiguous blocks | Fragmentation or free space |
Note that event 33 is the milder one. It says the oldest shadow copy was removed to stay under the limit, which is the system working as configured, and is not in itself a failure of growth.
The wording "were deleted because shadow copy storage could not grow in time" is the form quoted in a 2019 Microsoft forum thread about Windows Server 2008 through 2016 (Data Loss Event ID 25). Message text differs between Windows versions, so match on the event ID and the VolSnap source. The event 25 reference cited here is an archived page.
Likely causes
| Cause | What you see | How to confirm |
|---|---|---|
| Heavy writes outrun growth of the storage area (event 25) | Event 25 during a busy window such as a backup, import or batch job | Event time matches a known write burst; vssadmin list shadowstorage shows used space close to allocated |
| Maximum size reached (events 33, 36, 38) | Oldest copy trimmed, or copies aborted | The association's maximum is set and used is at or near it |
| Storage volume full or cannot extend (24, 35) | "Insufficient disk space" or "failed to grow" | Free space on the storage volume |
| Storage on the same busy disk | Event 25 on a heavily used file server | The storage volume is the volume being shadow-copied |
| Fragmented or full storage volume (41) | No sufficiently large contiguous blocks | Free-space layout of the storage volume |
| I/O failure or dismount (14, 16, 40) | Copies aborted because of an I/O failure or a dismounted volume | Disk and volume events at the same time |
Prove the cause before changing anything
These are read-only. The commands are taken from the documentation and have not been run.
- List the events. In PowerShell,
Get-WinEvent -FilterHashtable @{ LogName='System'; ProviderName='volsnap'; Id=14,16,24,25,33,35,36,38,40,41; StartTime=(Get-Date).AddDays(-7) }filters by log, provider, event IDs and start time. The manual says keys and values are case-insensitive andIDtakes an array (Get-WinEvent). Note the time of each event. - List the storage associations.
vssadmin list shadowstorageshows which volume stores the shadow copies of which, and how much space it uses. The Learn page says to run it "to find out where shadow copy storage resides and how much space it consumes, such as before you change its size".vssadmin list shadowstorage /on=C:lists the associations located on one volume (vssadmin list shadowstorage). Do not use/forand/ontogether. - List the copies that remain.
vssadmin list shadows /for=C:lists existing copies of a volume (vssadmin list shadows). - Compare with what ran at the time. Backups, database loads, antivirus scans and file-server bursts are common examples of write bursts. This is advice: Microsoft's pages do not list them as causes.
- Check the minimum size. If
MinDiffAreaFileSizeis set underHKEY_LOCAL_MACHINE\System\CurrentControlSet\Services\VolSnap, note it. It is a per-computer minimum. When unset, the minimum is 32 MB for volumes under 500 MB and 320 MB for larger ones, and "shadow copies might not function correctly" if the minimum exceeds the configured maximum (Registry Keys and Values for Backup and Restore).
Fix
Warning (data loss): changing where the storage lives deletes the existing shadow copies, and resizing "may cause shadow copies to disappear". If you rely on those copies, copy out what you need first.
Cause: growth cannot keep up (event 25)
Microsoft's resolution is to put the diff area on a volume on a different disk, which removes the contention from high I/O and "also provides better performance", recommended for heavily used file servers. The archived procedure uses the Configure Shadow Copies dialog: select the volume whose copy failed, choose Settings, disable shadow copies first if they are enabled, pick the new volume under Storage area, and enable again. The page cautions: "When you change the location of the Diff Area, all previous shadow copies are deleted" (Event ID 25). Where possible, also schedule write-heavy jobs away from shadow copy creation. A practical second step: Microsoft's page does not mention it.
Cause: maximum size too small (33, 36, 38)
Increase the maximum for the volume. The archived procedure for event 33 is to type a larger value for the maximum size of the storage area in the same dialog (Event ID 33). On the command line:
vssadmin resize shadowstorage /For=C: /On=D: /MaxSize=900MB
The syntax is vssadmin resize shadowstorage /for=<ForVolumeSpec> /on=<OnVolumeSpec> [/maxsize=<MaxSizeSpec>]. MaxSizeSpec must be 1 MB or more, in KB, MB, GB, TB, PB or EB (bytes if no unit); with no value there is no limit. The page's other examples are /MaxSize=UNBOUNDED and /For=C: /On=C: /MaxSize=20% (vssadmin resize shadowstorage). Choose a size from your own change rate, not from the example. An unbounded setting lets the shadow copy storage consume the storage volume, which then leads to events 24 and 35. Microsoft's page does not state this consequence; it follows from the storage volume filling.
Cause: storage volume full or cannot extend (24, 35)
Move the storage to a larger volume. For event 35 the archived page says the new volume must be an NTFS volume and uses the same Settings procedure, with the same warning that previous shadow copies are deleted (Event ID 35). Free space first if you can.
Cause: I/O failure or dismount (14, 16, 40)
These point to a disk or volume problem. Fix that first; changing shadow copy settings will not help.
Example (illustrative): FS01 keeps shadow copies of D: on D: itself with no size limit. A nightly import at 01:00 writes heavily and event 25 is logged at 01:12; the morning check finds no shadow copies. vssadmin list shadowstorage earlier showed used space of 212 GB against 212.3 GB allocated. The administrator adds a dedicated S: volume on another disk, moves the shadow copy storage there with a 300 GB maximum, and moves the import to start after shadow copies are taken. The names and numbers are invented.
Confirm the fix
- Create a shadow copy and confirm it is listed. The archived procedure uses Create Now, and
vssadmin list shadows /for=D:shows the result. - Run the heavy job again at its normal time and check the System log for any new
VolSnapevents. - If a backup uses VSS, a successful run alone is not proof. Restore a file from that backup to a scratch location and compare it with the source.
- Check
vssadmin list shadowstorageagain and note the used space against the maximum.
What to watch afterwards
- Alert on
VolSnapevents 24, 25, 35, 36, 38 and 41 and on a rising count of event 33. - Track used against maximum for each association over a month, so you see growth before it becomes a deletion.
- Treat shadow copies as a convenience for recent file versions. See the related article on why Previous Versions and shadow copies are not a backup.
- If shadow copies also fail for writer reasons, see the related article on diagnosing VSS writer failures.
Read next
Related articles you may want next.
- Backup & RecoveryPrevious Versions and shadow copies are not a backup: what they are good for, and where they stopUnderstand local Windows shadow-copy dependencies, automatic deletion, administrative access and why independent retained backups still matter.Read article
- Backup & RecoveryVSS Writer Failed? Diagnose It Before the Next BackupWindows VM backup failed at the VSS stage? Read each writer's state, phase and last error, find the owning service, and decide the fix before the next run.Read article
- TroubleshootingWindows Disk Full, VMware Space Free: Map the LayersMap Windows volumes to virtual disks and datastores, then identify where capacity is missing before approving an expansion or filesystem cleanup.Read article
Risks and limits
- No command in this article has been run or lab-tested.
- The event table is from archived Windows Server 2008 R2 pages (version 6.1). Event numbers and meanings are likely stable, but message text varies between Windows versions.
- The 2019 forum thread's author calls relocating the storage "a band aid and NOT a good solution, as that could also incur failures", and says
MinDiffAreaFileSize"does nothing to alleviate the issue". Microsoft staff did not offer a fix in the thread. Treat relocation and sizing as mitigation, and test under load. - The thread also describes changing the volume protection level through the VSS management interface. Per Microsoft's page, the alternative level makes all I/O to the original volume fail if the write to the shadow copy storage cannot complete. A later poster in the thread reported volumes going offline after trying it. This article does not recommend it.
vssadmin resize shadowstorageis documented for Windows client and server; the page's own warning is that resizing "may cause shadow copies to disappear".
Sources
- Diff Area Integrity (Windows Server 2008 R2) — Microsoft Learn archive, 27 January 2011, accessed 16 November 2026
- Event ID 25 — Diff Area Integrity — Microsoft Learn archive, 27 January 2011, accessed 16 November 2026
- Event ID 33 — Diff Area Integrity — Microsoft Learn archive, 27 January 2011, accessed 16 November 2026
- Event ID 35 — Diff Area Integrity — Microsoft Learn archive, 27 January 2011, accessed 16 November 2026
- vssadmin resize shadowstorage — Microsoft Learn, 5 March 2020, accessed 16 November 2026
- vssadmin list shadowstorage — Microsoft Learn, 7 September 2026, accessed 16 November 2026
- vssadmin — Microsoft Learn, 18 May 2018, accessed 16 November 2026
- vssadmin list shadows — Microsoft Learn, 18 May 2018, accessed 16 November 2026
- Registry Keys and Values for Backup and Restore — Microsoft Learn, 2 November 2026, accessed 16 November 2026
- VSS_PROTECTION_LEVEL enumeration (vsmgmt.h) — Microsoft Learn, 5 December 2018, accessed 16 November 2026
- Get-WinEvent — Microsoft Learn (PowerShell 7.6 view), 12 August 2026, accessed 16 November 2026
- Data Loss Event ID 25 "The shadow copies on Volume E: were deleted because shadow copy storage could not grow in time." — Microsoft Learn archived forum thread, 2 February 2019, accessed 16 November 2026
This article was originally published on .