we have a Windows Server 2016 and BE20.3, new and fully patched The server-storage is shared in two Volumes for Backup2Disk( 63TB) and for Dedup (23TB). There are about 200 Backup-Jobs, some Windows-Backups, some Linux-Backups. It works fine so far, but every 2 to 5 days we get a hardware-error 58053, which means hardware-error in changer. But we don´t use a changer. The consequence is, the dedup-volume is getting offline and every dedup-Job get stuck. All services stayed executed, only a reboot of the server bring the dedup-Volume online.
In the background a deduplication storage folder acts as a virtual changer (hence why you might sometimes see the odd misleading message saying a problem with the changer.)
With regards your problem as it appears to be indicative of some kind of issue within the deduplication storage I would suggest that a formal support case may be needed.