Setup: MSP360 Backup for Proxmox Community Edition, build 8.7.1.178, on Windows Server 2025; Proxmox VE 8.4. One plan, two VMs (4 TB disk, ~97% empty; 200 GB), Wasabi storage, forever-forward incremental, synthetic full and full consistency check on, nightly at 12:00 AM.
- Endpoint stalls right after the data transfer. The data upload finishes normally, then the MSP360 endpoint doesn’t answer Proxmox for about 2 minutes. Depending on which step it hits:
prune 'vm/102': proxmox-backup-client failed: Error: http request timed out(attempt fails),error fetching datastores - 500 read timeout(attempt fails), orunable to add notes ... http request timed out(warning only). Seen on 3 of the last 8 nights. - Retry is speed-capped. The automatic retry starts vzdump with
--bwlimit(69126 KiB/s one time, 26744 KiB/s another), although Tools → Options → Bandwidth is Unlimited for both cloud and local. Because the failed attempt invalidates Proxmox’s changed-block tracking, the retry reads the whole 4 TB disk; at 26 MiB/s that takes about 45 hours.
Questions: What causes the post-transfer stall, and which logs should we collect? Where does the retry’s bandwidth limit come from, and can it be raised or removed? Can a failure in the notes/prune step avoid triggering a full re-read?