| Index | Recent Threads | Unanswered Threads | Who's Active | Guidelines | Search |
 |
World Community Grid Forums
Category: Support
Forum: Website Support
Thread: Unable to report tasks |
| No member browsing this thread |
|
Thread Status: Active Total posts in this thread: 4
|
|
| Author |
|
|
keithhenry
Ace Cruncher Senile old farts of the world ....uh.....uh..... nevermind Joined: Nov 18, 2004 Post Count: 18665 Status: Offline Project Badges: 
|
Here's one of those pesky intermittent problems. On C2D running XP and BOINC 6.10.58, can report tasks fine. On quad running Vista and BOINC 6.10.58, reporting tasks gives these messages:
----------------------------------------1/7/2013 12:36:44 AM World Community Grid update requested by user 1/7/2013 12:36:46 AM World Community Grid Sending scheduler request: Requested by user. 1/7/2013 12:36:46 AM World Community Grid Reporting 6 completed tasks, not requesting new tasks 1/7/2013 12:36:52 AM Project communication failed: attempting access to reference site 1/7/2013 12:36:52 AM World Community Grid Scheduler request failed: Unrecognized HTTP Content-Encoding 1/7/2013 12:36:53 AM Internet access OK - project servers may be temporarily down. Eventually, as it has this time, it "clears up" and works fine. Restart of BOINC and reboot of machine have no impact. It has happened before at other times for varying lengths of time. Is this related to some cron job running on one of the schedulers? Oh, and FWIW, it looks like the tasks do actually get reported even though that isn't reflected in BOINC Manager. The six tasks from the above message snippet show being reported in Results Status a good half hour earlier and were validated by the time the sheduler request succeeded according to BOINC Manager. ---------------------------------------- [Edit 1 times, last edit by keithhenry at Jan 7, 2013 6:09:20 AM] |
||
|
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
Reporting to the server and confirming back the OK to the client are 2 steps in the cycle. The first can succeed and the second can fail. The server will continue to try and confirm back to the client, else the RtR entry wont clear in the BM view. Hence, the server could be showing a result reported and the client not yet. When setting the extra <sched_ops> log flag, you can actually see the acknowledgement being recorded. Sample:
----------------------------------------2182 World Community Grid 1/7/2013 8:28:13 AM [sched_op] Starting scheduler request 2183 World Community Grid 1/7/2013 8:28:13 AM Sending scheduler request: To report completed tasks. 2184 World Community Grid 1/7/2013 8:28:13 AM Reporting 1 completed tasks 2185 World Community Grid 1/7/2013 8:28:13 AM Not requesting tasks: don't need 2186 World Community Grid 1/7/2013 8:28:13 AM [sched_op] CPU work request: 0.00 seconds; 0.00 devices 2187 World Community Grid 1/7/2013 8:28:17 AM Scheduler request completed 2188 World Community Grid 1/7/2013 8:28:17 AM [sched_op] Server version 701 2189 World Community Grid 1/7/2013 8:28:17 AM Project requested delay of 11 seconds 2190 World Community Grid 1/7/2013 8:28:17 AM [sched_op] handle_scheduler_reply(): got ack for task c4cw_target07_000547572_0 Some of these messages you will not get though on the 6.10.58 client, but the <sched_ops> entries are same across v6 and v7 FWIW ;>) edit: Actually, think I made a mistake advertising this option. The next panic-master will start posting that her/his client is not receiving the acks within 4 seconds. ;P [Edit 1 times, last edit by Former Member at Jan 7, 2013 9:33:38 AM] |
||
|
|
keithhenry
Ace Cruncher Senile old farts of the world ....uh.....uh..... nevermind Joined: Nov 18, 2004 Post Count: 18665 Status: Offline Project Badges:
|
Thanks. As usual, your answer makes a lot of sense. Hopefully, I'll be able to remember this info the next time I see this error.
---------------------------------------- Rather than add that option and get all that detail info (with all the related overhead that adds to BM and thus, reduced computing capacity - fair warning to your panic-masters  ), I'll just take a quick glance at Results Status and confirm that the tasks BM tried to report did in fact make it there and know that the error is just with the acks and forget about it. Needless to say, subsequent attempts by BM to report the tasks must simply get a "already have those reported" response and those acks will clear them out of BM. |
||
|
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
Yes, there will be a "result already reported, ignoring" (or something to that effect) when the original ack got lost on the Intertubes. I've yet to see a single result get lost [except during that first hour of server 700 go live, March last year. Crunched a dozen tasks twice
 ] , so I can't be bothered. Eventually the RtR's will clear. Cow dung happens BTW, think that visiting the RS pages takes more cycles than logging [look at the CPU time in Task Manager your webbrowser takes in a day versus what the BM needs... 1-2 minutes a day]. Actually, I never use the BM but to temp shut down a host. I'm a BOINCTasks addict.... used according the TM 4:29 minutes in past 48 hours monitoring 3 hosts, and logging the completed result history, to include the acks times [when white they're uploaded, when green they were confirmed back]. |
||
|
|
|