Occasionally on one server we run out of disk space when running our Hyperbac backup job (a SQL 2005 Maintenance Plan backup task running via SQL Server Agent) . When this happens the job "hangs" and all other backup tasks on the server (to disk or tape) are also stalled.
Stopping the Agent job still leaves the spid behind it running. If we kill the spid it sits in the Killed/Rollback state and it is still not possible to run any backups.
Trying to stop the Hyperbac service results in it going permanently into the Stopping status.
The only way we have to resolve this is to use procexp to kill the Hyperbac process. This allows the spid to clear and backups to be taken as normal. We can then restart Hyperbac.
The version of Hyperbac that we are running is 220.127.116.11. Prior to this we were running v3 which did not have this problem i.e. when a disk filled up the Agent job failed in the same way that would happen for a native SQL backup.
Do you have a fix for this?