Short answer: “Failed to open I/O device [\\?\GLOBALROOT\Device\HarddiskVolumeShadowCopy17]” means Windows deleted the shadow copy Veeam was reading in the middle of the backup. Check the System event log for volsnap Event ID 25 (“the shadow copy storage could not grow in time”). Run the job again once the computer is quiet, and if it keeps happening, give the shadow copy storage more room or move it to another volume.
A car dealership’s nightly Veeam Agent backup for one office computer failed with 0 bytes transferred and the error “The system cannot find the file specified. Failed to open I/O device”. The disk was healthy and had 170 GB free, so it looked like a broken backup agent. The event logs told a different story. The computer had shut down unexpectedly and stayed off for 15 days. When it was turned back on, Veeam started the missed backup 62 seconds after boot, right in the middle of Windows Update, antivirus and other startup work. Windows ran out of room for the snapshot and deleted it mid-backup. Here is what we found and how to fix it.
Fixed by Kenn Cruz, IT Support Specialist at GCIT. Written up by Elliot Munro, CISO.
Symptoms
The Veeam Agent for Microsoft Windows job report shows the job in Error with 0.0 B read and transferred, and details like:
Error: The system cannot find the file specified. Failed to open I/O device [\\?\GLOBALROOT\Device\HarddiskVolumeShadowCopy17] (read-only: [1], async mode: [1]) Failed to open emulated disk. Failed to open disk for read. Failed to upload disk. Agent failed to process method {DataTransfer.SyncDisk}

The shadow copy number at the end changes from run to run. Other forms of the same problem include “The device is not ready” and “Shadow copy was not found”. In the computer’s System event log, at the same time as the failure, you will usually find:
The shadow copies of volume C: were deleted because the shadow copy storage could not grow in time.
That entry comes from the volsnap source with Event ID 25.
Affected Software
- Application: Veeam Agent for Microsoft Windows (workstation and server editions), also seen with Veeam Backup & Replication guest processing
- Platform: Windows 10, Windows 11 and Windows Server
- Related Services: Volume Shadow Copy Service (VSS) and the volsnap driver
Cause
Veeam doesn’t back up the live disk directly. It asks Windows to take a shadow copy (a point-in-time snapshot) and reads from that. While the backup runs, Windows keeps a record of every change to the disk in a shadow copy storage area, which grows as needed.
If the disk is very busy, that storage area can’t grow fast enough to keep up. When that happens, Windows deletes the shadow copy to protect the volume and logs volsnap Event ID 25. Veeam suddenly loses the snapshot it was reading, so it reports that the I/O device can’t be found.
Common triggers are:
- A missed or catch-up backup starting just after boot, when Windows Update, antivirus scans and other startup tasks are all hitting the disk (our case).
- Shadow copy storage capped at a small size. In our case it was limited to 5% of the drive.
- Very low free space on the volume.
- Heavy disk activity during the backup window, such as large file copies, database maintenance or defragmentation.
Resolution
Step 1: Confirm the shadow copy was deleted
- On the affected computer, open Event Viewer > Windows Logs > System.
- Filter on the source volsnap and look for Event ID 25 at the time the job failed.
- Check what else was happening. A recent unexpected shutdown (Event ID 6008) followed by a boot just before the job is a strong sign of a catch-up run colliding with startup activity.
Step 2: Run the backup again
- Once the computer has finished starting up, run the job again from the Veeam Agent tray icon (Backup Now) or the Veeam console.
- Confirm it completes with a new restore point. Ours finished in 20 minutes with no further volsnap errors.
Step 3: Stop it happening again
- Give shadow copy storage more room. From an elevated command prompt, check the current limit with
vssadmin list shadowstorage, then raise it, for examplevssadmin resize shadowstorage /for=C: /on=C: /maxsize=15%. Veeam suggests 15 to 20% of the source volume for busy volumes. - Move shadow storage on busy servers. Veeam recommends placing shadow storage on a different, less busy volume with plenty of free space, using
vssadmin add shadowstorage /for=X: /on=Z: /maxsize=##GB. - Free up disk space if the volume is nearly full.
- Keep heavy tasks out of the backup window, such as defragmentation, large copies and database maintenance.
Important Notes
- It is rarely a broken backup agent. Reinstalling Veeam won’t help if Windows is deleting the snapshot. Check the event log before changing anything.
- A machine that has been off creates a backup gap. Our computer had no backup for 15 days because it was switched off, not because the job was broken. If a device is a spare or being retired, remove its backup job rather than letting alarms pile up.
- Retries often succeed. One failure after a long power-off is common. Repeated failures every night point to storage or I/O load that needs fixing.
Related Resources
- Veeam KB4573: Backup fails with “The system cannot find the file specified” or “The device is not ready”
- Microsoft Learn: vssadmin resize shadowstorage
- Fix: Veeam “There Is Not Enough Space on the Disk” Backup Failure
- Fix: Veeam “Failed to Create Production Checkpoint” (Hyper-V)
- Business Continuity and Disaster Recovery at GCIT
Want someone watching your backups every day? Get in touch with our team and we’ll make sure every failure is caught and fixed.