Message boards :
Number crunching :
The/A Problem with BOINC 4.41
Message board moderation
Author | Message |
---|---|
MikeSW17 Send message Joined: 3 Apr 99 Posts: 1603 Credit: 2,700,523 RAC: 0 |
I hope the following log can Either: help the Devs fix 4.41... Or, let the Devs explain why there is nothing wrong. Nothing complicated, the system was re-booted and a Result was due to finish. I only run SETI, with a 3 day cache and nothing is anywhere near deadline. This simple log seems to demonstrate most of the problems people are reporting. 15/05/2005 22:22:10||Starting BOINC client version 4.41 for windows_intelx86 15/05/2005 22:22:10||Data directory: C:Program FilesBOINC 15/05/2005 22:22:10|SETI@home|Found app_info.xml; using anonymous platform 15/05/2005 22:22:10|SETI@home|Computer ID: 862543; location: home; project prefs: default 15/05/2005 22:22:10||General prefs: from SETI@home (last modified 2005-05-14 01:50:14) 15/05/2005 22:22:10||General prefs: no separate prefs for home; using your defaults 15/05/2005 22:22:10||Remote control allowed 15/05/2005 22:22:11|SETI@home|Resuming computation for result 19dc04aa.17564.27009.167318.111_2 using setiathome version 4.11 15/05/2005 22:22:14||schedule_cpus: must schedule 15/05/2005 23:01:32||request_reschedule_cpus: process exited 15/05/2005 23:01:32|SETI@home|Computation for result 19dc04aa.17564.27009.167318.111_2 finished 15/05/2005 23:01:32||schedule_cpus: must schedule 15/05/2005 23:01:32|SETI@home|Starting result 30dc04ab.22830.112.853420.231_2 using setiathome version 4.11 So far, so Good. 15/05/2005 23:01:34|SETI@home|Requesting 19320.42 seconds of work Why request work - In 4 seconds we're about to decide I'm over-commited? 15/05/2005 23:01:34|SETI@home|Sending scheduler request to http://setiboinc.ssl.berkeley.edu/sah_cgi/cgi 15/05/2005 23:01:35|SETI@home|Started upload of 19dc04aa.17564.27009.167318.111_2_0 15/05/2005 23:01:37|SETI@home|Finished upload of 19dc04aa.17564.27009.167318.111_2_0 15/05/2005 23:01:37|SETI@home|Throughput 581954 bytes/sec 15/05/2005 23:01:38|SETI@home|Scheduler request to http://setiboinc.ssl.berkeley.edu/sah_cgi/cgi failed 15/05/2005 23:01:38|SETI@home|No schedulers responded So, it issued a Scheduler Request while it was still uploading? 15/05/2005 23:01:38|SETI@home|Deferring communication with project for 1 minutes and 0 seconds 15/05/2005 23:01:45||Computer is overcommitted 15/05/2005 23:01:45||Nearly overcommitted. 15/05/2005 23:01:45||New work fetch policy: no work fetch allowed. 15/05/2005 23:01:45||New CPU scheduler policy: earliest deadline first. But we just requested 19320 secondes of work above (but didn't get it) - No way am I overcommitted either. ---- The 1 minute defer above did nothing so I hit update at this point... 15/05/2005 23:05:15||request_reschedule_cpus: project op 15/05/2005 23:05:15||schedule_cpus: must schedule 15/05/2005 23:05:15||New work fetch policy: work fetch allowed. 15/05/2005 23:05:15||New CPU scheduler policy: highest debt first. 15/05/2005 23:05:16|SETI@home|Requesting 18616.00 seconds of work So, Now I only want 18616 seconds work - 4 minutes ago it was 19320. I've done about 200 sec work since then, but need _less_ work now? Ah! Hitting Update did something then... 15/05/2005 23:05:16|SETI@home|Sending scheduler request to http://setiboinc.ssl.berkeley.edu/sah_cgi/cgi 15/05/2005 23:05:18|SETI@home|Scheduler request to http://setiboinc.ssl.berkeley.edu/sah_cgi/cgi succeeded 15/05/2005 23:05:19|SETI@home|Started download of 30dc04ab.18554.21456.53410.62 15/05/2005 23:05:26|SETI@home|Finished download of 30dc04ab.18554.21456.53410.62 15/05/2005 23:05:26|SETI@home|Throughput 59547 bytes/sec 5 Minutes ago I was over committed, No Work policy was set. What changed, why download now? 15/05/2005 23:05:26||request_reschedule_cpus: files downloaded 15/05/2005 23:05:26||schedule_cpus: must schedule So a New WU has been obtained, and Uploaded Result as been reported - just what it should do. But didn't we go round the houses to get there? I get the impression that the code is doing all the various upload/download and scheduling bits in the wrong order or they're not proberly synchronized with each other. It's a nightmare for the user to understand what is happening right and what is wrong. |
©2024 University of California
SETI@home and Astropulse are funded by grants from the National Science Foundation, NASA, and donations from SETI@home volunteers. AstroPulse is funded in part by the NSF through grant AST-0307956.