Why have I stopped receiving work?
Message boards :
Number crunching :
Why have I stopped receiving work?
Message board moderation
Author | Message |
---|---|
![]() Send message Joined: 29 Mar 20 Posts: 19 Credit: 9,076,779 RAC: 8,626 |
I stopped receiving CPU work about 10 days ago. I use Boinc Manager, and currently have 2 active projects enabled for CPU work. I have all CPU resources allocated to Asteroids, zero percent to Milkyway. I continue to receive work from Milkyway, but none from Asteroids. I have tried updating, nothing. |
![]() ![]() Send message Joined: 16 Nov 22 Posts: 120 Credit: 108,634,293 RAC: 438,103 |
Check to be sure that you are modifying the host venue location that the host is actually attached to. Also set some logging options to have Boinc tell you WHY you aren't getting any work here. Use the Manager Options >> Logging Options function to add sched_ops_debug and work_fetch_debug to the standard logging options. Use the sched_ops_debug option first to see how much work you are requesting on every scheduler connection. If that does not show anything useful, then go and toggle on the work_fetch_debug option. That produces a lot of output and is hard to decipher is the host is attached to a lot of projects. But it sounds like you are only connected to Asteroids and Milkyway so the output should be pretty clear to decipher. ![]() ![]() A proud member of the OFA (Old Farts Association) |
![]() Send message Joined: 29 Mar 20 Posts: 19 Credit: 9,076,779 RAC: 8,626 |
|
![]() ![]() Send message Joined: 16 Nov 22 Posts: 120 Credit: 108,634,293 RAC: 438,103 |
How many seconds of cpu work do you see in the scheduler request when you set the sched_op_debug flag? Even though you are only doing cpu tasks on two projects, setting the work_fetch_debug flag is VERY talkative and will show output for all tasks and projects. It will be a lot of output to wade through. But if you find the the output for Asteroids in the work fetch output in the Event Log, you can highlight that line and then select the "Show only this project" in the Event Log's interface via the button on the bottom of the window. That will make is easier to search through the logging output. Work_fetch_debug output will explain Boinc's reasoning how much cpu work it requested and then the reason WHY it was unable to get any. Normal reason is the project weighting disallows it or deadlines have to be met for other project work. Each scheduler connection, Boinc runs through a round-robin simulation of the hosts resources and the amount of work it has in queue already to determine how much work to request. Boinc does an incredible amount of calculation behind the scenes to determine just what and how much to run for all the projects on the host. It tries to equilibrate the amount of credit spread across all attached projects. It does this on every scheduler connection. But this is averaged over a very long period of ten days via a stock setting of <rec_half_life_days>10.000000</rec_half_life_days> in the cc_config.xml file in the Boinc directory. The REC variable stands for Recent Estimated Credit. Its this function that attempts to constantly equilibrate the credit across all projects. Whenever there is an imbalance between projects because one project that has recently done a lot of high credit work causes an imbalance to the other projects, that project gets put on the 'back burner' so to speak to let other projects catch up and payoff the credit debt to that project. You can speed up this equilibration by changing the default value of REC from ten days ot only 1 day. I also recommend you try changing this value in your cc_config.xml file. ![]() ![]() A proud member of the OFA (Old Farts Association) |
![]() Send message Joined: 29 Mar 20 Posts: 19 Credit: 9,076,779 RAC: 8,626 |
What I see is "Can't fetch cpu; Account Manager prefs." I'm wondering if this will clear eventually, or if I need to adjust a preference. I'm willing to wait but it has been more than 2 weeks. It was switching between the projects just fine for several months, and then it seems like something changed. The project has sometimes not had work available, but that doesn't seem to be the case now. |
![]() Send message Joined: 29 Mar 20 Posts: 19 Credit: 9,076,779 RAC: 8,626 |
|
![]() Send message Joined: 23 Apr 21 Posts: 83 Credit: 100,100,141 RAC: 251,466 |
Under Boinc Manager project properties for Asteroids it says "Don't request tasks for cpu", Account Manager preference. I never set any parameter to not request cpu tasks. I don't know how this changed. How do I reset it to request cpu tasks? Thanks- remove your account manager and manage the host by yourself. ![]() ![]() |
![]() Send message Joined: 29 Mar 20 Posts: 19 Credit: 9,076,779 RAC: 8,626 |
Thanks, I need to figure out how to do that. Detaching the project from BM seems fairly straightforward except it still appears on the project list, and BM gave a warning that removing it may cause a CPID issue if the project retires. From there I need to figure out how to request work for it myself. I would prefer to keep it attached and fix the issue if possible, it has always worked fine that way up until a couple weeks ago. |
![]() ![]() Send message Joined: 16 Nov 22 Posts: 120 Credit: 108,634,293 RAC: 438,103 |
|
![]() Send message Joined: 29 Mar 20 Posts: 19 Credit: 9,076,779 RAC: 8,626 |
|
![]() ![]() Send message Joined: 16 Nov 22 Posts: 120 Credit: 108,634,293 RAC: 438,103 |
First go to the the BoincStats site and then to your BAM account. Now detach all your projects from your "My Projects" menu in the BAM account manager. Then go to your Boinc Manager in the Advanced View and to the Tools Menu. Then in the Tools menu, select "Use an Account Manager" and select "None" Save the settings and exit the Manager. Open the Manager again and now use the Tools menu to "Add a Project" and join all the projects again you removed from the BAM account. Use the same email address you used previously to join projects so you keep the same CPID. ![]() ![]() A proud member of the OFA (Old Farts Association) |
![]() Send message Joined: 29 Mar 20 Posts: 19 Credit: 9,076,779 RAC: 8,626 |
|
![]() Send message Joined: 1 Jan 23 Posts: 6 Credit: 10,405,014 RAC: 34,799 |
|
![]() Send message Joined: 23 Apr 21 Posts: 83 Credit: 100,100,141 RAC: 251,466 |
NO tasks for over 24 hours read here: https://asteroidsathome.net/boinc/forum_thread.php?id=911&postid=8078 ![]() ![]() |
Message boards :
Number crunching :
Why have I stopped receiving work?