SUMMARY
Multiple subsystem backups were failing, due to not being able to register/reserve space on a system.
ISSUE
Multiple subsystems are failing and not able to register / reserve space on the system.
RESOLUTION
Update to version 9.1.0 or later of the Unitrends software.
CAUSE
This condition is due to an excessive number of SMGR registrations that are stale (no longer being used).
Failures during open and close could result in a reservation being left on the system. There is a limit of 512 reservations so eventually this can cause the subsystems to fail when enough of these occur. This particular failure will have 10MB reservations that are stale and smgr_display will show an entry similar to the following, where the 4th field is the size of the registration:
2 1929379984 D2DBackups/1 10485760 0 Mon Aug 15 10:39:04 2016
The specific failure detected is where a backup begins and the client stops responding. The bpserverd-backup error log will have an entry:
Aug 15 11:12:51 : [LOG3] archive.c:5052: All data has been received from the client, finishing up ... Aug 15 11:12:51 : [LOG0] archive.c:5063: Expected EOF undetected, failing backup (total_data sent 133660672).