If the core was processing a job that requires mounting a recovery point such as Rollup or a Recovery Point Check and the process was interrupted the EPOCH / Recovery Point file that was mounted may not be purged by the system automatically.
These files will remain in the repository directory for the specific agent it was processing and these files can take up repository space which could then cause other processes in the future to fail.
Using the following steps you can first verify whether this behavior has taken place as well as correct the issue and recoup repository space
On the Core
1. Open the console
2. Click on the agent in question so that the agent summary screen is displayed
3. Locate the repository path being used for the agent that is failing jobs
4. Using Windows Explorer navigate to the repository target path for the agent
5. Search the agents repository directory for file names that include DeleteOnClose
EXAMPLE / F:\tevorepository\AGENT NAME\87253ae1-c6e5-11e1-833f-00155d089b00DeleteOnClose_20289_20962.ReplayEPS
If you are seeing recovery point file names in the agents repository directory that include DeleteOnClose after a job fails then there could be some garbage / junk files that need to be purged to recoup repository space so proceed to next step
6. Close the Replay Admin Console
7. Stop the Replay Core Service and Agent Service if also running
8. Once the core / agent service has stopped open the Task Manager and End Task the following orphan processes if running
Replay Admin Console
9. Start the Replay Core Service and Open the console and Windows Explorer and check the repository space being reported and verify free space has been recouped.
***Repeat this process for any agents that are reporting snapshot failures, rollup failures, recovery point check failures***