I try to setup my BOINC to use the Enigma@home as a backup project when the SETI@home has no work. I set Enigma@home to reasonably small resourse share (0.005) and expect it to not download any WU while SETI@home WUs are available.

But instead of this, Enigma@home asks for WU every time it finish the previous one. The result is that one CPU core is always filled with Enigma@home work. That is not what i wish to see.

There is a part of the BOINC log:
06/04/2009 14:21:58 [wfd] ------- start work fetch state -------
06/04/2009 14:21:58 [wfd] target work buffer: 691200.00 sec
06/04/2009 14:21:58 [wfd] CPU: shortfall 0.00 nidle 0.00 est. delay 1024620.76 RS fetchable 100.00 runnable 100.00
06/04/2009 14:21:58 SETI@home [wfd] CPU: fetch share 1.00 debt 0.00 backoff dt 0.00 int 0.00
06/04/2009 14:21:58 Enigma@Home [wfd] CPU: fetch share 0.00 debt -96923.81 backoff dt 0.00 int 0.00 (comm deferred)
06/04/2009 14:21:58 [wfd] CUDA: shortfall 0.00 nidle 0.00 est. delay 864213.63 RS fetchable 100.00 runnable 100.00
06/04/2009 14:21:58 SETI@home [wfd] CUDA: fetch share 1.00 debt 0.00 backoff dt 0.00 int 0.00
06/04/2009 14:21:58 Enigma@Home [wfd] CUDA: fetch share 0.00 debt 0.00 backoff dt 22331.83 int 86400.00 (comm deferred)
06/04/2009 14:21:58 SETI@home [wfd] overall_debt 0
06/04/2009 14:21:58 Enigma@Home [wfd] overall_debt -96924
06/04/2009 14:21:58 [wfd] ------- end work fetch state -------
06/04/2009 14:21:58 [wfd] No project chosen for work fetch
06/04/2009 14:22:20 SETI@home [task_debug] result 20ja09af.10630.145208.11.8.144_2 checkpointed
06/04/2009 14:22:37 SETI@home [task_debug] result 20ja09af.26275.144390.10.8.52_0 checkpointed
06/04/2009 14:22:54 [work_fetch_debug] Request work fetch: Project backoff ended
06/04/2009 14:22:58 Enigma@Home chosen: CPU starved
06/04/2009 14:22:58 [wfd] ------- start work fetch state -------
06/04/2009 14:22:58 [wfd] target work buffer: 691200.00 sec
06/04/2009 14:22:58 [wfd] CPU: shortfall 0.00 nidle 0.00 est. delay 0.00 RS fetchable 100.01 runnable 100.00
06/04/2009 14:22:58 SETI@home [wfd] CPU: fetch share 1.00 debt 0.00 backoff dt 0.00 int 0.00
06/04/2009 14:22:58 Enigma@Home [wfd] CPU: fetch share 0.00 debt -96923.81 backoff dt 0.00 int 0.00
06/04/2009 14:22:58 [wfd] CUDA: shortfall 0.00 nidle 0.00 est. delay 0.00 RS fetchable 100.00 runnable 100.00
06/04/2009 14:22:58 SETI@home [wfd] CUDA: fetch share 1.00 debt 0.00 backoff dt 0.00 int 0.00
06/04/2009 14:22:58 Enigma@Home [wfd] CUDA: fetch share 0.00 debt 0.00 backoff dt 22271.42 int 86400.00
06/04/2009 14:22:58 SETI@home [wfd] overall_debt 0
06/04/2009 14:22:58 Enigma@Home [wfd] overall_debt -96924
06/04/2009 14:22:58 [wfd] ------- end work fetch state -------
06/04/2009 14:22:58 Enigma@Home [wfd] request: CPU (34.56 sec, 0) CUDA (0.00 sec, 0)
06/04/2009 14:22:58 Enigma@Home [sched_op_debug] Starting scheduler request
06/04/2009 14:22:58 Enigma@Home Sending scheduler request: To fetch work.
06/04/2009 14:22:58 Enigma@Home Requesting new tasks
06/04/2009 14:22:58 Enigma@Home [sched_op_debug] CPU work request: 34.56 seconds; 0 idle CPUs
06/04/2009 14:22:58 Enigma@Home [sched_op_debug] CUDA work request: 0.00 seconds; 0 idle GPUs
06/04/2009 14:23:01 SETI@home [task_debug] result 08ja09ac.385.481.14.8.202_2 checkpointed
06/04/2009 14:23:04 Enigma@Home Scheduler request completed: got 1 new tasks
06/04/2009 14:23:04 Enigma@Home [sched_op_debug] Server version 603
06/04/2009 14:23:04 Enigma@Home Project requested delay of 242 seconds
06/04/2009 14:23:04 Enigma@Home [task_debug] result state=NEW for hceyz72_0_4987548_r0_0 from handle_scheduler_reply
06/04/2009 14:23:04 Enigma@Home [sched_op_debug] estimated total CPU job duration: 4468 seconds
06/04/2009 14:23:04 Enigma@Home [sched_op_debug] estimated total CUDA job duration: 0 seconds
06/04/2009 14:23:04 Enigma@Home [sched_op_debug] Deferring communication for 4 min 2 sec
06/04/2009 14:23:04 Enigma@Home [sched_op_debug] Reason: requested by project
06/04/2009 14:23:04 [work_fetch_debug] Request work fetch: RPC complete
06/04/2009 14:23:05 Enigma@Home [task_debug] result state=FILES_DOWNLOADING for hceyz72_0_4987548_r0_0 from CS::update_results
06/04/2009 14:23:06 Enigma@Home Started download of 00hc_wu_4987548_r0
06/04/2009 14:23:07 Enigma@Home Finished download of 00hc_wu_4987548_r0
06/04/2009 14:23:08 Enigma@Home [task_debug] result state=FILES_DOWNLOADED for hceyz72_0_4987548_r0_0 from CS::update_results
06/04/2009 14:23:08 SETI@home [cpu_sched] Preempting 08ja09ac.385.481.14.8.202_2 (left in memory)
06/04/2009 14:23:08 SETI@home [task_debug] task_state=SUSPENDED for 08ja09ac.385.481.14.8.202_2 from suspend
06/04/2009 14:23:08 Enigma@Home Starting hceyz72_0_4987548_r0_0
06/04/2009 14:23:08 Enigma@Home [cpu_sched] Starting hceyz72_0_4987548_r0_0 (initial)
06/04/2009 14:23:08 Enigma@Home [task_debug] task_state=EXECUTING for hceyz72_0_4987548_r0_0 from start
06/04/2009 14:23:08 Enigma@Home Starting task hceyz72_0_4987548_r0_0 using enigma_m4_2 version 522

Is there any suggestions?

You want the backup project to only download work when the first project is down? BOINC can't do that. As soon as you attach it to two project, no matter the amount of resource share you give them, BOINC will always try to get at least one task per project in. And with a 14 day deadline, this means Enigma will just run.

If you truly want to do this, set Enigma to No New tasks and put it back on Allow New Tasks when Seti is down.

---

You want the backup project to only download work when the first project is down? BOINC can't do that.

So the article at http://www.boinc-wiki.info/index.php?title=Backup_Project has amission in part "Expected Behavior"?

The BOINC Powered Backup Project will download and complete one or a few Work Units initially. After that the BOINC Powered Backup Project will generally only get Work Units when the primary Project is unable to supply Work Units, and very rarely at other times...

Or the page is "outdated" because the BOINC 6.6.20 can't do something that the BOINC 4.4.5 can?

If you truly want to do this, set Enigma to No New tasks and put it back on Allow New Tasks when Seti is down.

That means i need babysit around the BOINC client and manually switch project(s) on and off? I'm a science enthusiast, not a maniac. Leave BOINC with the SETI@home alone and don't care if it have work or not is much simply for me. After reading the Wiki article i think, i can setup projects to switch them automaticaly due to the range of my "primary" and "auxiliary" science interests. If you say that i can't do that, the only question is: why so useful ability is not officially supported by BOINC developers?

If you truly want to do this, set Enigma to No New tasks and put it back on Allow New Tasks when Seti is down.

It requires manual intervention in BOINC behavior while all idea of having backup project is to free yourself from monitoring BOINC and projects state.
Why resource shares exist at all if BOINC will not obey them?
Should it be reported as bug or "new" feature requested?

---