Index | Recent Threads | Unanswered Threads | Who's Active | Guidelines | Search |
![]() |
World Community Grid Forums
![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() |
No member browsing this thread |
Thread Status: Active Total posts in this thread: 65
|
![]() |
Author |
|
alanb1951
Veteran Cruncher Joined: Jan 20, 2006 Post Count: 972 Status: Offline Project Badges: ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() |
I receive 1 or 2 a day ARPs at the moment. This one is strange: As it happens, the example you linked to now seems to have resolved (two valid, one invalid), but for a general insight into why some tasks seem to stick at Pending Validation for days see my answer to MJH333's query on the first page of this thread...https://www.worldcommunitygrid.org/contribution/workunit/155115138 I received it early this day and sent it back in the evening - still "Pending Validation". The other two were sent on 25.8. - still "Pending Verification" after completion. What is the difference? Thanks for an explanation. Cheers - Al. |
||
|
robertmiles
Senior Cruncher US Joined: Apr 16, 2008 Post Count: 443 Status: Offline Project Badges: ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() |
3 tasks in 1 workunit all pending validation, with one computer being a little late. https://www.worldcommunitygrid.org/contribution/workunit/155141517 Looks like ARP1 resend is either very slow or not working with "Waiting to be sent". https://www.worldcommunitygrid.org/contribution/workunit/155390774 My BOINC event log shows: Scheduler request completed: got 0 new tasks No tasks are available for Africa Rainfall Project I will guess ARP1 may be slowed down or stopped at this moment. Yes, all the recent ones I have had have been resneds. Only picked up one yesterday. and that was its fifth attempt. Completed fine though. Interestingly, two others had completed the task. I had always thought only two successful completions were needed? Two successful completions are enough IF THE VALIDATOR DECIDES THAT THEY AGREE ENOUGH. Otherwise, another task will be issued. Repeat this until two tasks agree enough. The problem appears to be that the deadlines are based on when the list of input files is sent. Basing it on when all of the input files have been downloaded would be a better choice. The deadline can be changed on the server, but not also on the computer doing the task, after it is downloaded. This would have been a better way to handle the slow downloads, if the deadline changes continue only as long as the client keeps trying to download an input file. The slow downloads problem appears to be fixed, but now there appears to be a new problem of not enough tasks available to keep up with requests. |
||
|
alanb1951
Veteran Cruncher Joined: Jan 20, 2006 Post Count: 972 Status: Offline Project Badges: ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() |
@robertmiles - for information - if you know all this already, my apologies...
Regarding deadlines being based on request time -- once the list of work and associated files has been sent from the server to the client the central BOINC system has no idea as to what happened to that work until the client reports the success (or otherwise) of each task. The download and upload servers are effectively independent of the rest of the core system, the only interactions being the addition and removal of files, so the BOINC scheduler has no [simple] way of knowing that a client is struggling to get (or return) files. The only way to know when the downloads have completed would be for the client to make an intermediate report... If the downloads fail, a well-behaved client will let the server know reasonably quickly; this mechanism works fine except when there are severe download problems, and changing it to cope with a situation that should not happen when the BOINC server systems are running properly would involve adding unnecessary complexity! Cheers - Al. |
||
|
zdnko
Senior Cruncher Joined: Dec 1, 2005 Post Count: 225 Status: Offline Project Badges: ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() |
The slow downloads problem appears to be fixed, but now there appears to be a new problem of not enough tasks available to keep up with requests. Today I got a lot of "transient HTTP error" again! ![]() |
||
|
catchercradle
Advanced Cruncher England Joined: Jan 16, 2009 Post Count: 130 Status: Offline Project Badges: ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() |
Seven of my tasks that had been validation pending are now valid so things seem to be moving. Got three tasks currently running and a fourth which is trying to download but a few retries seems to get them to finish.
----------------------------------------Can anyone point me to where I set the maximum number of this task type to have at once? I used to know where it was but can't find it any more. Thanks. Edit: that last just to check. and I now have five tasks running, four has been the most I have had at once since things started working again prior to now and often only one. [Edit 1 times, last edit by catchercradle at Sep 5, 2022 10:08:20 AM] |
||
|
Bryn Mawr
Senior Cruncher Joined: Dec 26, 2018 Post Count: 345 Status: Offline Project Badges: ![]() ![]() ![]() ![]() ![]() ![]() ![]() |
Seven of my tasks that had been validation pending are now valid so things seem to be moving. Got three tasks currently running and a fourth which is trying to download but a few retries seems to get them to finish. Can anyone point me to where I set the maximum number of this task type to have at once? I used to know where it was but can't find it any more. Thanks. Edit: that last just to check. and I now have five tasks running, four has been the most I have had at once since things started working again prior to now and often only one. From memory it’s in the device manager. |
||
|
alanb1951
Veteran Cruncher Joined: Jan 20, 2006 Post Count: 972 Status: Offline Project Badges: ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() |
Can anyone point me to where I set the maximum number of this task type to have at once? I used to know where it was but can't find it any more. Thanks. [Edit; I see Bryn Mawr posted something a lot shorter whilst I was writing this :-) -- I'll leave this as is in case it helps someone...]If you want to control the maximum number of tasks WCG will send for each individual project, you need to adjust the Device Profile associated with the particular machine(s). If you want to control the actual mix of work running on your system at any given time that's a different matter (search the forums for [recent] posts about app_config.xml.) Trying to find the relevant page can be a bit of an adventure, so here's the relevant URL... https://www.worldcommunitygrid.org/ms/device/viewProfiles.do Once you have selected the Profile Name that goes with your device, scroll down to near the foot of the page where you will find a section called Project Liimits where you can set things to suit your requirements. By the way, if you are running ARP1 tasks and don't want to risk returning work too late (or not at all!) don't set the count for ARP1 higher than the number of copies you are willing to run at once. Quite a lot of the retries we see are because folks didn't return tasks within the deadline :-( Cheers - Al. [Edit 1 times, last edit by alanb1951 at Sep 5, 2022 12:56:36 PM] |
||
|
catchercradle
Advanced Cruncher England Joined: Jan 16, 2009 Post Count: 130 Status: Offline Project Badges: ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() |
Thank you . that helps. I only run ARP at the moment so that simplifies things a bit :)
|
||
|
Bryn Mawr
Senior Cruncher Joined: Dec 26, 2018 Post Count: 345 Status: Offline Project Badges: ![]() ![]() ![]() ![]() ![]() ![]() ![]() |
Can anyone point me to where I set the maximum number of this task type to have at once? I used to know where it was but can't find it any more. Thanks. [Edit; I see Bryn Mawr posted something a lot shorter whilst I was writing this :-) -- I'll leave this as is in case it helps someone...]If you want to control the maximum number of tasks WCG will send for each individual project, you need to adjust the Device Profile associated with the particular machine(s). If you want to control the actual mix of work running on your system at any given time that's a different matter (search the forums for [recent] posts about app_config.xml.) Trying to find the relevant page can be a bit of an adventure, so here's the relevant URL... https://www.worldcommunitygrid.org/ms/device/viewProfiles.do Once you have selected the Profile Name that goes with your device, scroll down to near the foot of the page where you will find a section called Project Liimits where you can set things to suit your requirements. By the way, if you are running ARP1 tasks and don't want to risk returning work too late (or not at all!) don't set the count for ARP1 higher than the number of copies you are willing to run at once. Quite a lot of the retries we see are because folks didn't return tasks within the deadline :-( Cheers - Al. Thanks, I was being very lazy :-) |
||
|
catchercradle
Advanced Cruncher England Joined: Jan 16, 2009 Post Count: 130 Status: Offline Project Badges: ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() |
Thanks, found it and upped it to 8 the number of real cores I have on my machine.
|
||
|
|
![]() |