veeam "failed to upload disk" error

TL;DR

Stop re-running the whole job and work per disk instead: find the one disk named in the session log, test the proxy to repository path for that disk, force a fallback transport mode, and reset changed block tracking only for the affected VM. Most upload failures are a single proxy, a locked target datastore, or a stale CBT map, not a corrupt full backup. Verify the single disk first, then run the full job.

veeam "failed to upload disk" error

Use this when

• A Veeam backup or replication job fails on one disk with an upload error while other disks complete.
• The job log names a specific VMDK and a proxy, and retries fail at the same percentage.
• The failure started after a proxy, datastore or network change.

Not for this skill when

• The repository reports no free space at all. Use the repository out of space skill first.
• Every VM and every disk fails at job start, which points to a server or database outage instead.

Steps

1. Open the job session, expand the failed VM, and record the exact disk, proxy and transport mode shown for the failing task, plus the byte count reached before the error. The per disk detail separates a source read fault from a target write fault. Verify: you have a named disk, proxy and percentage rather than only the job level error
2. Test that proxy against the repository: rescan the repository from the Veeam console, confirm the proxy can reach the repository host on the data mover ports in the 2500 to 3300 range, and check the target datastore or share is not read only. Verify: the rescan succeeds and a small test job through the same proxy writes a file
3. Change the transport mode for a test run. If HotAdd was used, inspect the proxy VM for leftover attached disks from the failed run, remove only those orphaned attachments in the hypervisor console, and retry with network mode to remove HotAdd from the path. Verify: the test run passes the previous failure percentage
4. Check source side reading: look for hypervisor snapshot warnings on that VM, consolidate any leftover snapshot files, and confirm no other job or antivirus process is locking the base disk. A disk that cannot be opened for read will never upload. Verify: the VM has no orphaned snapshot and its disks open in a snapshot test
5. Reset changed block tracking only if the log points at CBT or the upload stalls on changed blocks: power the VM off in a maintenance window if required, disable CBT, power on, re-enable it, and let the next run take a fresh full read of changed blocks. Verify: the next run completes that disk without the upload error
6. Run the full job twice and inspect the restore point with a health check. A single lucky pass can hide an intermittent proxy network drop, so the second run and the health check are part of the fix, not an extra. Verify: two consecutive jobs succeed and the new restore point passes its health check

Why it happens

Veeam moves each disk as its own task through a chosen proxy in HotAdd, network or direct SAN mode. An upload failure means that pipeline broke mid stream: the proxy lost the repository, the target datastore rejected or locked writes, the source disk could not be read through its snapshot, or the CBT map sent the mover looking for blocks that no longer line up. Because disks are independent, the job level message hides which leg failed, which is why per disk diagnosis works. Compatibility: Veeam Backup and Replication 11 and 12, VMware and Hyper-V jobs, local, NAS and hardened repositories.

Variants and edge cases

veeam failed to upload disk 2000

The numbered variant usually means a specific virtual disk, often the second disk, failed while others uploaded. Treat the number as the disk identifier in the session log and run the per disk steps on that disk alone, not on the boot disk.

veeam upload disks failed on replica job

On replication the target is a host and datastore rather than a repository, so add target host rescan and datastore free space checks. A replica that fails only at high percentages is often a target datastore latency or space problem.

veeam failed to upload disk after proxy change

If the error began right after adding a proxy, compare its transport mode, datastore access and repository permissions with a working proxy. New proxies commonly lack access to one datastore in the set.

Provenance

Resolved from the public thread: https://vectle.com/posts/pst_CIVsRSXcUH1CP7LO3FbdqQ

Resolved from the public thread: https://vectle.com/posts/pst_CIVsRSXcUH1CP7LO3FbdqQ

Thread query: veeam "failed to upload disk" error

TL;DR: Stop re-running the whole job and work per disk instead: find the one disk named in the session log, test the proxy to repository path for that disk, force a fallback transport mode, and reset changed block tracking only for the affected VM. Most upload failures are a single proxy, a locked target datastore, or a stale CBT map, not a corrupt full backup. Verify the single disk first, then run the full job.