| Index | Recent Threads | Unanswered Threads | Who's Active | Guidelines | Search |
 |
World Community Grid Forums
Category: Support
Forum: BOINC Agent Support
Thread: Cannot Upload--COmmunication Deferred |
| No member browsing this thread |
|
Thread Status: Active Total posts in this thread: 95
|
|
| Author |
|
|
SekeRob
Master Cruncher Joined: Jan 7, 2013 Post Count: 2741 Status: Offline |
Bump, did an edit to previous post.
|
||
|
|
Mumak
Senior Cruncher Joined: Dec 7, 2012 Post Count: 477 Status: Offline Project Badges: 
|
Disabling requesting new tasks cleared the "Ready to report" tasks.
----------------------------------------However even if I set the Store N days of work to 0.01 I still get the error. Going to try a project reset.  |
||
|
|
Mumak
Senior Cruncher Joined: Dec 7, 2012 Post Count: 477 Status: Offline Project Badges:
|
Project reset didn't help.
----------------------------------------Not even setting the "Store at least ..." to a very low numbers or 0 helps... Upgraded client to v7.6.33 - no help. Now I'm out of work... [Edit 3 times, last edit by Mumak at Jul 9, 2016 12:40:43 PM] |
||
|
|
SekeRob
Master Cruncher Joined: Jan 7, 2013 Post Count: 2741 Status: Offline |
Maybe this July 16, next weekend ”network update ".http://www.worldcommunitygrid.org/about_us/viewNewsArticle.do?articleId=485 is going to address the various issues MD5\RSA\Reporting\Fetching, but, it's always 'need to know' basis, dont set any expectation😇
----------------------------------------[Edit 1 times, last edit by SekeRob* at Jul 9, 2016 12:34:19 PM] |
||
|
|
BobCat13
Senior Cruncher Joined: Oct 29, 2005 Post Count: 295 Status: Offline Project Badges: 
|
This problem can usually be traced to a timed-out work request, meaning the client requests work, the server issues work, but for some reason the communication between the two does not complete and reaches the 5 minute time limit, showing something in the Event log like Failure to received data from server; time-out reached.
This causes the Lost results feature to kick in and then when the lost results are sent, the client will get the No close tag error. I have not had it happen on WCG, but have seen it on a few other projects. The only way I could resolve it was to set the client up to get 1 resend at a time until all resends have been received. You can try setting the client to New new tasks on WCG and then enable the <fetch_on_update>1</fetch_on_update> option in cc_config.xml and press Update for WCG in the Projects tab of BOINC Manager. If the client still tries to receive more than 1 resend at a time using this procedure and gets the No close tag error, then try also enabling <fetch_minimal_work>1</fetch_minimal_work> which will only get 1 task per device. |
||
|
|
Mumak
Senior Cruncher Joined: Dec 7, 2012 Post Count: 477 Status: Offline Project Badges:
|
Bob,
----------------------------------------I tried using those options, but the client doesn't request any tasks: update requested by user Sending scheduler request: Requested by user. Not requesting tasks Scheduler request completed |
||
|
|
SekeRob
Master Cruncher Joined: Jan 7, 2013 Post Count: 2741 Status: Offline |
<fetch_minimal_work>1</fetch_minimal_work> will only work if there is -no- work in the buffer and will just fetch 1 task, for 1 core and 1 core only... which facilitates sizing the scheduler file down to the minimum, the object of the exercise [to not incur truncation/timeout]. More roads to Rome [Tiburtinas], the same is achieved with an empty zero days buffer setting and only allow 1 core to crunch... all to break out of the vicious circle [as Tony succeeded to do]
----------------------------------------[Edit 1 times, last edit by SekeRob* at Jul 9, 2016 5:31:16 PM] |
||
|
|
SekeRob
Master Cruncher Joined: Jan 7, 2013 Post Count: 2741 Status: Offline |
BTW, there's various http timeout settings available for cc_config... played them a bit, but since I don't have the issue [because probably I only report and fetch piecemeal], not been able to identify the effect, and I'm not going to break what's whole ;D
Makes me think, if I set something for http in the config file, does it then also apply to https, which WCG favors? |
||
|
|
KWSN-A Shrubbery
Senior Cruncher Joined: Jan 8, 2006 Post Count: 476 Status: Offline Project Badges: 
|
Same results as Mumak, project reset didn't work, re-install didn't work.
----------------------------------------All the cc_config options haven't cleared things up. I sit at zero work, unable to request more. At this point, I'm waiting for the "lost" results to go past the deadline and see if it starts getting more work. This is a disaster for anyone who isn't tech saavy or doesn't read the forums. It simply means an increasing number of hosts being excluded from BOINC as the issue arrises.  |
||
|
|
SekeRob
Master Cruncher Joined: Jan 7, 2013 Post Count: 2741 Status: Offline |
A reset does not clear everything [may vary per client version], a detach, re-add WCG does [not sure if you lose the app_config.xml if you do, so make a copy before this blunt action]
|
||
|
|
|