What's up with this boinc-versions? Running almost empty
16 Jun 2009 21:47:15 UTC
Topic 194399
(moderation:
I joined 4 projects.
resource settings are 100 for Einstein@Home
60 for Orbit at home, but this project has almost never any work
2 for Rectilinear Crossing Number
1 for SETI@Home
This is my box.
http://einsteinathome.org/host/1967526
Boinc is 6.6.31
Now, at exactly this time I have:
7 E@H-WUs which are uploaded, but not reported
4 Rectilinear WUs which are stopped by boinc, remaining time for each is 4 Minutes
8 E@H WUs which are in progress, between 1% and 85%
5 E@H WUs in the queue
8 Rectilinear WUs in the queue
Please note that most Rectilinear WUs take less than a minute, the maximum time seems to be 4:05 hours.
So I do not even have work for the next 10 hours.
In the settigs I have a connection of 2.4 days, I expect that my queue should be filled enough for these 2.4 days.
Now, what ist wrong here?
Quote:
$ egrep "Reporting |got .* new tasks" stdoutdae.txt| tail -28
15-Jun-2009 15:29:47 [SETI@home] Scheduler request completed: got 0 new tasks
15-Jun-2009 15:46:11 [Rectilinear Crossing Number] Reporting 6 completed tasks, requesting new tasks
15-Jun-2009 15:46:16 [Rectilinear Crossing Number] Scheduler request completed: got 0 new tasks
15-Jun-2009 18:40:33 [orbit@home] Scheduler request completed: got 0 new tasks
15-Jun-2009 23:24:54 [Einstein@Home] Reporting 8 completed tasks, requesting new tasks <-- this was the only manual click I did
15-Jun-2009 23:24:59 [Einstein@Home] Scheduler request completed: got 0 new tasks
15-Jun-2009 23:26:04 [Einstein@Home] Scheduler request completed: got 0 new tasks
15-Jun-2009 23:28:15 [Einstein@Home] Scheduler request completed: got 0 new tasks
15-Jun-2009 23:30:22 [Einstein@Home] Scheduler request completed: got 0 new tasks
15-Jun-2009 23:34:31 [Einstein@Home] Scheduler request completed: got 0 new tasks
15-Jun-2009 23:34:41 [Rectilinear Crossing Number] Scheduler request completed: got 0 new tasks
15-Jun-2009 23:41:44 [Einstein@Home] Scheduler request completed: got 0 new tasks
15-Jun-2009 23:45:51 [SETI@home] Scheduler request completed: got 0 new tasks
15-Jun-2009 23:57:51 [orbit@home] Scheduler request completed: got 0 new tasks
16-Jun-2009 00:08:27 [Einstein@Home] Reporting 2 completed tasks, requesting new tasks
16-Jun-2009 00:08:32 [Einstein@Home] Scheduler request completed: got 0 new tasks
16-Jun-2009 00:34:50 [Rectilinear Crossing Number] Scheduler request completed: got 0 new tasks
16-Jun-2009 01:01:15 [Einstein@Home] Scheduler request completed: got 0 new tasks
16-Jun-2009 01:38:04 [Einstein@Home] Reporting 1 completed tasks, requesting new tasks
16-Jun-2009 01:38:09 [Einstein@Home] Scheduler request completed: got 0 new tasks
16-Jun-2009 04:20:41 [SETI@home] Scheduler request completed: got 0 new tasks
16-Jun-2009 04:36:00 [Einstein@Home] Reporting 5 completed tasks, requesting new tasks
16-Jun-2009 04:36:05 [Einstein@Home] Scheduler request completed: got 0 new tasks
16-Jun-2009 05:17:36 [orbit@home] Scheduler request completed: got 0 new tasks
16-Jun-2009 06:34:10 [SETI@home] Scheduler request completed: got 0 new tasks
16-Jun-2009 08:13:24 [Einstein@Home] Reporting 5 completed tasks, requesting new tasks
16-Jun-2009 08:13:29 [Einstein@Home] Scheduler request completed: got 0 new tasks
16-Jun-2009 08:36:43 [SETI@home] Scheduler request completed: got 0 new tasks
16-Jun-2009 15:11:09 [orbit@home] Scheduler request completed: got 0 new tasks
16-Jun-2009 16:58:11 [orbit@home] Scheduler request completed: got 0 new tasks
16-Jun-2009 17:30:45 [SETI@home] Scheduler request completed: got 0 new tasks
16-Jun-2009 19:31:47 [Einstein@Home] Reporting 13 completed tasks, requesting new tasks
16-Jun-2009 19:31:57 [Einstein@Home] Scheduler request completed: got 0 new tasks
16-Jun-2009 19:49:30 [SETI@home] Scheduler request completed: got 0 new tasks
16-Jun-2009 20:49:39 [SETI@home] Scheduler request completed: got 0 new tasks
16-Jun-2009 21:49:44 [SETI@home] Scheduler request completed: got 0 new tasks
16-Jun-2009 22:47:34 [Rectilinear Crossing Number] Scheduler request completed: got 0 new tasks
16-Jun-2009 22:49:50 [SETI@home] Scheduler request completed: got 0 new tasks
If my counting is correct, I have uploaded 40 WUs.
During that time neither Einstein@Home, nor Rectilinear and not even SETI@Home gave me any new work.
I think my box is fast enough, the uptime is close to 100% (ignoring a few reboots needed to set up some new version of the operating system), so whats wrong here.
I do not want to click around in the boinc manager, I do not want to hack any xml-files, it should simple work without a nurse.
Or with other words: Which version of boinc is stable enough to install and forget.
What's up with this boinc-versions? Running almost empty
)
Seems it was not running out of jobs, it rather entered a "collect WUs" stage:
Yesterday I had 5 E@H-WUs in the queue, now I have 39 (plus 8 which are stopped by boinc).
On the other hand, boinc seems to panic, 8 Rectilinear WUs are "Running, high priority".
Well, I have seen this "Running, high priority" before. It will suddenly vanish and those Rectilinear WUs will be stopped a few minutes before they are finished. This behaviour is very similar to a manic depression.
RE: Seems it was not
)
Boinc is not a very good scheduler but it is what we have to use. Over time Boinc will even out the units to pretty much what you set it for but in the short term it is running into deadline differences at all the different projects you are running. That is why alot of us became 'gardeners' and 'farmers' and 'ranchers', which are designations based on how many computers each person has. This gives us a way to avoid Boincs scheduling issues and run mostly one project per machine, I currently am a 'rancher' with 17 pc's running here at home. No I don't run 17 projects but could without all the issues you are having. Unfortunately there is no real solution other than to become a 'gardener' in your case, which is the lowest number of computers but would still be enough to satisfy your need to crunch for many projects. Basically your problem is ...too many different project deadlines for Boinc to try and keep track of, each project is wildly different from the others. This leads to a problem you haven't mentioned yet, Long Term Debt(LTD) issues. This is when one project gets no work for an extended amount of time, a week or more, and then all of a sudden all other projects are put on hold while it gets all the crunching power for a while.
One thing you might consider is to crunch for project A for a few months with project B being the backup with a lower percentage. Then switch to projects C and D for the next few months, then back to A and B, etc. Or join the ranks of us multiple pc owners.
RE: Seems it was not
)
Boinc is not a very good scheduler but it is what we have to use. Over time Boinc will even out the units to pretty much what you set it for but in the short term it is running into deadline differences at all the different projects you are running. That is why alot of us became 'gardeners' and 'farmers' and 'ranchers', which are designations based on how many computers each person has. This gives us a way to avoid Boincs scheduling issues and run mostly one project per machine, I currently am a 'rancher' with 17 pc's running here at home. No I don't run 17 projects but could without all the issues you are having. Unfortunately there is no real solution other than to become a 'gardener' in your case, which is the lowest number of computers but would still be enough to satisfy your need to crunch for many projects. Basically your problem is ...too many different project deadlines for Boinc to try and keep track of, each project is wildly different from the others. This leads to a problem you haven't mentioned yet, Long Term Debt(LTD) issues. This is when one project gets no work for an extended amount of time, a week or more, and then all of a sudden all other projects are put on hold while it gets all the crunching power for a while.
One thing you might consider is to crunch for project A for a few months with project B being the backup with a lower percentage. Then switch to projects C and D for the next few months, then back to A and B, etc. Or join the ranks of us multiple pc owners.
I hate farming, I hate
)
I hate farming, I hate gardens.
Maybe because when I was young my parents had a house with a garden and my job was to mow the lawn.
Now, seriously. It is not me who runs in problems, I am just watching boinc for a few days, and then I will forget it. If it runs okay, that is fine for those projects, if it fails don't blame me, rather blame the boinc crew.
I bought that big machine to reduce my farm and to be able to run multiple OS on a single machine at the same time, maybe you have seen that "xen" in the version of the kernel? Finally I will have three machines running. One Macbook Pro (which may crunch again, if I fix those fans somewhen), one Dual-Opteron acting as a server and that new big thing. The rest will have only one job: collecting dust.
RE: I hate farming, I hate
)
Yea I too would love to downgrade the total number of machines but somehow they just keep multiplying! I would also love to see Boinc be able to have one project run exclusively on a single core of a machine and a different project run on a different core, without any input from me other than the initial setup! I know this is not a priority for Dr. A and crew but I wish it would be!!! I think this would solve alot of problem, including yours and probably increase the hope of me getting rid of some pc's. It is summer and it is hot where my pc's are!!