Work unit captured by black hole

What is the tidiest way to close things for a backup without losing any work - just exit BOINC?

I tried suspending the two projects I have running (BBC CCE and Einstein@Home), but suspending the BBC CCE resulted in an attempt to download a second Einstein@Home work unit, which then has got orphaned since I exited BOINC during the download - it seems the system thinks it was downloaded okay but it wasn't.

2006-07-22 03:31:33 [---] Rescheduling CPU: application exited
2006-07-22 03:31:33 [Einstein@Home] Computation for task h1_1490.5_S5R1__4427_S5R1a_1 finished
2006-07-22 03:31:33 [Einstein@Home] Starting task h1_1490.5_S5R1__4426_S5R1a_1 using einstein_S5R1 version 402
2006-07-22 03:31:35 [Einstein@Home] Started upload of file h1_1490.5_S5R1__4427_S5R1a_1_0
2006-07-22 03:31:42 [Einstein@Home] Finished upload of file h1_1490.5_S5R1__4427_S5R1a_1_0
2006-07-22 03:31:42 [Einstein@Home] Throughput 24192 bytes/sec
2006-07-22 05:55:45 [Einstein@Home] Sending scheduler request to http://einstein.phys.uwm.edu/EinsteinAtHome_cgi/cgi
2006-07-22 05:55:45 [Einstein@Home] Reason: To report completed tasks
2006-07-22 05:55:45 [Einstein@Home] Reporting 1 tasks
2006-07-22 05:55:50 [Einstein@Home] Scheduler request succeeded
2006-07-22 12:19:04 [---] Rescheduling CPU: project suspended by user
2006-07-22 12:19:04 [BBC Climate Change Experiment] Pausing task hadcm3ln_7slv_00602841_1 (left in memory)
2006-07-22 12:19:06 [Einstein@Home] Sending scheduler request to http://einstein.phys.uwm.edu/EinsteinAtHome_cgi/cgi
2006-07-22 12:19:06 [Einstein@Home] Reason: To fetch work
2006-07-22 12:19:06 [Einstein@Home] Requesting 17280 seconds of new work
2006-07-22 12:19:06 [---] Rescheduling CPU: project suspended by user
2006-07-22 12:19:07 [Einstein@Home] Pausing task h1_1490.5_S5R1__4426_S5R1a_1 (left in memory)
2006-07-22 12:19:11 [---] Exit requested by user
2006-07-22 12:19:12 [---] Rescheduling CPU: exit_tasks

Thanks,
Graham.