| Index | Recent Threads | Unanswered Threads | Who's Active | Guidelines | Search |
 |
World Community Grid Forums
Category: Support
Forum: BOINC Agent Support
Thread: WCG does not get new tasks |
| No member browsing this thread |
|
Thread Status: Active Total posts in this thread: 24
|
|
| Author |
|
|
markhl
Cruncher Joined: Sep 24, 2006 Post Count: 12 Status: Offline |
I run WCG under the BOINC 6.10.58 client on three devices. On two it works well. On the third, it often runs out of units, 24-36 hours after starting Windows. This happens one or two times per week so I am missing some opportunities to do WCG work. When it happens there are no error messages in the log. There are no scheduler messages. It just does not even try to contact the server!
Rebooting or clicking Projects -> WCG -> Update works. It contacts the server, reports completed tasks, downloads new tasks, starts work, and all is well. This is Win7 64-bit Home Premium. ASUS CM5675 desktop. 8 GB RAM, Intel i5 3.20 GHz CPU. Why is it doing this? How can I stop it running out of tasks, without having to check WCG task loading every few hours? Thanks, Mark |
||
|
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
Hello markhl,
Check your selected profile settings. First, make sure that you are using the website preferences. Reading the Messages tab of BOINC Manager after a reboot on that computer will tell you if it is using local preferences or global (website) preferences. Then select My Grid - Device Manager - (Selected Profile) on the WCG website. This will show you the Workunit Cache Settings. I use 'Connect to network about every 0.1 days' which draws a new work unit when the current work unit is estimated to complete in 2.4 hours. If that does not show an obvious problem, then please post the Messages tab after a reboot so we can see what sort of system you have. Lawrence |
||
|
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
Don't know and never encountered this. Could speculate, but lets first collect some data by adding a few debug flags to the cc_config.xml file per this manual: http://boinc.berkeley.edu/wiki/Cc_config.xml
----------------------------------------The flags to add into the <log_flags> section are <work_fetch_debug>1</work_fetch_debug> <sched_op_debug>1</sched_op_debug> When inserted and saved, do a read config in the BOINC Advanced menu. Then let it run till the next time it goes idle again / not fetching work. When it does, copy the whole message log from the top to bottom in a post/reply so we can read what BOINC is deciding. Normally a backfill would be constant according the connect and additional buffer settings. Also, can you confirm that the activity menu settings are Run Always and Network Always available. Finally, the screen where there is the Update button has a properties button. Select WCG and hit that button. Post a screenshot or copy-type the lines in the scheduling section to a next reply. Presuming that only WCG is the attached project and not any other. --//-- edit: added the <sched_op_debug> flag for more decision detail. [Edit 3 times, last edit by Former Member at Jan 14, 2012 7:36:11 AM] |
||
|
|
TPCBF
Master Cruncher USA Joined: Jan 2, 2011 Post Count: 2173 Status: Offline Project Badges: 
|
Don't know and never encountered this. I did though. Irregular though with no indication as to why...Also, can you confirm that the activity menu settings are Run Always and Network Always available. In my case(s), yes, it's the default way it is installed. And all setups I have seen this (can't tell about a couple of laptops of friends that I don't get to see that often), are running 24/7/365 (with the occasional reboot/shutdown due to power outage or Windows/etc updates), on "always on" Internet connections...As mentioned, it happens randomly, on various hosts, and once I update manually, it might not happen again for weeks or months...  Ralf |
||
|
|
JmBoullier
Former Community Advisor Normandy - France Joined: Jan 26, 2007 Post Count: 3716 Status: Offline Project Badges: 
|
Finally, the screen where there is the Update button has a properties button. Select WCG and hit that button. At least, do it before pressing the Update button every time you need to do a manual update. And check if there is anything in the Status part of WCG's line in the Projects tab.What you describe looks like your client is under a long backing off request from the server when it happens. Also please tell us the cache settings for these machines. Until we find an explanation, increasing the cache size might help avoid any shortage situation. |
||
|
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
 We ought to have a football pileup icon.  Lawrence |
||
|
|
TPCBF
Master Cruncher USA Joined: Jan 2, 2011 Post Count: 2173 Status: Offline Project Badges:
|
Happened again on a remote machine today. After the last of three previously downloaded WUs was finished, it just sat there idle and downloaded new WUs only after I hit the update button. Nothing about any server backoff or anything like that, just sitting there doing nothing for the better part of the day. Here's a snippet from the message log:
1/17/2012 10:09:04 PM World Community Grid Computation for task faah29126_ZINC32865215_1_x3NF6b_00_0 finishedAt 9:26am, the last WU finished processing and uploading, then just silence until I hit the update button at 8:42pm Ralf |
|||
|
||||
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
Plz go to my first post in this thread and provide the detail message log info resulting from that. Also a copy of a section in the client_state.xml file located in the BOINC data directory [path printed at top when client is started]. The section to copy / paste in a reply is per the sample below:
<time_stats> <on_frac>0.946925</on_frac> <connected_frac>0.853568</connected_frac> <active_frac>0.999422</active_frac> <gpu_active_frac>0.973669</gpu_active_frac> <last_update>1326960191.115511</last_update> </time_stats> <net_stats> <bwup>24066.018049</bwup> <avg_up>26531682.472974</avg_up> <avg_time_up>1326960194.820723</avg_time_up> <bwdown>554843.719437</bwdown> <avg_down>166303756.939236</avg_down> <avg_time_down>1326943997.588347</avg_time_down> </net_stats> Also we like to know the Duration Correction Factor value that you can see by going to the BOINC Manager Projects tab, selecting the WCG line and hitting Properties button on left. What are your cache settings, connect interval, additional buffer, on-time per day. What client version are you running? Please include that in your reply. --//-- |
||
|
|
TPCBF
Master Cruncher USA Joined: Jan 2, 2011 Post Count: 2173 Status: Offline Project Badges:
|
What are your cache settings, connect interval, additional buffer, on-time per day. What client version are you running? Please include that in your reply. I thought I had posted that before already, main problem right now is that this is a remote machine and that it is in use right now, so I can not get any of that info again until later today or after office hours.This is an absolute stock install, as downloaded from the WCG site, BOINC v6.10.58. As a rule, I do not play with any of the settings, never had a need to. And more than likely, this same workstation will work just fine for a few weeks before it might show this problem again, but then another one of the roughly two dozen that I have running with WCG will randomly show the same problem... Ralf |
||
|
|
TPCBF
Master Cruncher USA Joined: Jan 2, 2011 Post Count: 2173 Status: Offline Project Badges:
|
Ok, had this now happen on another PC during the day, not the one where I had the problem last time (different location, OS as well).
Don't know and never encountered this. Could speculate, but lets first collect some data by adding a few debug flags to the cc_config.xml file per this manual: Ok, that whole part doesn't make much sense right now, as it might be weeks until this might happen again. <snip> Normally a backfill would be constant according the connect and additional buffer settings. Also, can you confirm that the activity menu settings are Run Always and Network Always available. Yes.Finally, the screen where there is the Update button has a properties button. Select WCG and hit that button. Post a screenshot or copy-type the lines in the scheduling section to a next reply. Presuming that only WCG is the attached project and not any other. SchedulingCPU scheduling priority -1593.06 CPU work fetch priority 0.0 CPU work fetch deferred for --- CPU work fetch deferral interval --- Duration correction factor 0.9359 Also a copy of a section in the client_state.xml file located in the BOINC data directory <time_stats>And here's the part of the message logs from today: 1/19/2012 7:54:25 AM World Community Grid update requested by user 1/19/2012 7:54:29 AM World Community Grid Sending scheduler request: Requested by user. 1/19/2012 7:54:29 AM World Community Grid Reporting 2 completed tasks, not requesting new tasks 1/19/2012 7:54:31 AM World Community Grid Scheduler request completed 1/19/2012 1:29:45 PM World Community Grid Computation for task GFAM_x1kmvHumanDHFRdry_0002921_0980_1 finished 1/19/2012 1:29:48 PM World Community Grid Started upload of GFAM_x1kmvHumanDHFRdry_0002921_0980_1_0 1/19/2012 1:29:51 PM World Community Grid Finished upload of GFAM_x1kmvHumanDHFRdry_0002921_0980_1_0 1/19/2012 4:11:25 PM World Community Grid Computation for task DSFL_00000104_0000029_0165_1 finished 1/19/2012 4:11:27 PM World Community Grid Started upload of DSFL_00000104_0000029_0165_1_0 1/19/2012 4:11:32 PM World Community Grid Finished upload of DSFL_00000104_0000029_0165_1_0 Ralf |
||
|
|
|