i am running boinc 6.6.36 on win7 rc1 x64

i have 2 projects: seti and milkyway

some time ago i ran out of work for the milkway project, ever since i dont receive new work anymore
telling me this:

08/07/2009 21:49:59 Milkyway@home Message from server: No work sent

08/07/2009 21:49:59 Milkyway@home Message from server: (won't finish in time) BOINC runs 71.3% of time, computation enabled 98.2% of that

when i suspended seti and i ordered an update of milkyway i received work again....

i assume i configured something wrong, but is there a possibility this is a bug?

No bug.
BOINC only works the averages. Over a period of a few months, BOINC will see to it that your computer gets the ratio of work from your attached projects that you've specified.
If you keep "micro managing" things yourself, this will muck up BOINC's plans, and it will have to start again with when it will allocate time to different projects.

Unfortunately, micro-managing the scheduling oneself is the only way to make 6.6.36 work. I have CPDN running and with reporting due in May 2010. I have given equal preference to Rosetta (it has 3 hour WUs) on two dual core computers and BOINC gives way too much priority to CPDN. If I left it to BOINC, I would get no Rosetta units - the server thinks the WU will not finish in time. This is an issue with scheduling. It is well documented. There are many posts on this forum and on project fora saying the same thing. A common response is that it is known about but that there is no will to fix it.

---

Unfortunately 6.6.36 doesn't work well with Milkyway. I, and several others at MW had the same problem. Each person solved the problem by going back to an earlier Boinc version (6.6.28 in my case). And I might add, it DID finish work in time, and all the other work, too.

So, somehow 6.6.36 and MW have uncovered a bug which doesn't seem to exist unless both are involved. But for now, go back to an earlier version and you'll find work downloading again.

Rod

Unfortunately, micro-managing the scheduling oneself is the only way to make 6.6.36 work. I have CPDN running and with reporting due in May 2010. I have given equal preference to Rosetta (it has 3 hour WUs) on two dual core computers and BOINC gives way too much priority to CPDN. If I left it to BOINC, I would get no Rosetta units - the server thinks the WU will not finish in time. This is an issue with scheduling. It is well documented. There are many posts on this forum and on project fora saying the same thing. A common response is that it is known about but that there is no will to fix it.

no will to fix... sounds a little irrisponsible towards the projects participating and being the victim of this bug...(i know the boinc dev's are busy people so no offense meant but...)


i am running boinc on 20 computers and adding. I specially installed the latest boinc to have all pc's the same,i dont feel like walking by them again and reinstall an older version.

tho maybe it is a good idea anyways since i also experience uploading problems, leaving me with maybe 200 or 300 completed units (seti/milkyway) not getting uploaded while downloading works perfectly

no will to fix...

I'm not sure what sources rtX is using for his information, but what he says isn't true. I would take his information with large grains of salt, from now on. Don't just believe everything others write, seek some confirmation. At best read that it's his own opinion.

David is looking into it, but there needs to be a logging flag on, on the project's server, to get a communications log of affected machines. This flag isn't on by default as it takes up quite a bit of space when all machines keep making these logs.

I have asked one project to enable the flag, but they ran out of work and are introducing a new application, so it can take a while before I get the necessary log from them and I can send it on to David.

At this moment a workaround is to reset the project, although that's a heavy measurement and should only be taken if you have no work for the project and nothing left to crunch, upload or report, as a reset will delete all that.

A reset may not immediately fix it, as in the next communication with the project you may still get the message once. A new reset will then fix that. (In my experience at least).

The upload problems at Seti aren't due to BOINC, but due to the upload server at Seti. It was off for a couple of days and now that it has turned back on, all computers out there are trying to upload their data at the same time. It's only a 100Mbit line that they have. So it can take some time, or as I did for calculations on the Seti forums:

Now, when uploading on a 100Mbit line that's completely clogged up, you may reach the scheduler, but by the time you actually start uploading the result file, the whole pipe is clogged. Although you seem to be uploading, the data isn't getting through, it'll even push backwards.

Compare it to draining water from the sink, where once there's not enough air in the pipe, the water will splosh back up, suck in some air and go down again.

100Mbit divided by 8 bits equals 12.5Mbytes. That's the maximum that can be uploaded per second. 12.5Mbytes times 1024 equals 12,800 kilobytes. Now divide 12,800 by an average of 25 kilobytes per result that's uploading, that's 512 results. Not necessarily 512 hosts...

But OK, so that's 512 results that can be uploaded per second. If all result files are 25 KB, which they aren't. Some are bigger, some are smaller.

Now, with 180,000 hosts out there, with about 4,911,372 results to be uploaded (at a guess, that number is the total amount of Seti Enhanced tasks out in the field, not necessarily all crunched and waiting to upload), you can see that they won't all fit through the 100Mbit pipe at the same time.

In a perfect world, where everyone had the same speed internet connection, where everyone could upload their results in one second, it would take 4,911,372 divided by 512 equals 9592 seconds for all that work to come back.

But since we don't live in a perfect world and not everyone can upload their work in one second, add a whole lot more!

---

sorry ageles, i was frustrated at the moment of reading from a lot of other stuff and having my results run up to a couple hundred still waiting to send in....
i just took it for granted....

ill keep waiting for the upload to work propperly again...

about the problem with not receiving: for now i will make use of the suspending seti and let milky way get results in and resume seti until there is a permanent fix in a next release?
its too much of a hasle to lookup the password for every computer, then make the projects stop fetching work and wait until the crunching is done then wait for all to be send ...

David has added some code to the source code for people to test. I will try to build a BOINC client this weekend that uses this code.

If you want to test that client and see if it works, PM me with your email address (do not post it in a post here in the forums, spam bots read these forums as well!) and when I manage to build a version without errors, I'll email it to you.

---

i send you a PM including the setup from 2 different systems i can test that client on

thank you

I've given this one out to Wim already, as well as Walt.

I've built a client with the latest source code that fixes the "won't finish in time" problem. It is 32bit and Windows only.

Its .RAR archive can be gotten from my Skydrive. It is called BOINC 6.6.76, to make sure I am outside the range of Berkeley products and comes with my own BOINC Manager which shows both the CPU time and Wall Time columns in the Tasks tab.

Steps to follow wrote:

Before you do anything, stop BOINC. Then make sure to backup your original boinc.exe, boinc.dll and boincmgr.exe (sqlite3.dll comes only with 6.6.28 and above and does not necessarily need to be backed up, it's the same file all over the place)

Now unpack the archive into your BOINC directory (default: C:\Program Files\BOINC\). The sqlite3.dll is used by BOINC Manager, so if you do not have it in your BOINC directory, it's at least included in this package.

When that's done, just restart BOINC. It'll start doing benchmarks, since I named my version 6.6.76 (to be outside the possible range Berkeley upholds). And that's it.

You should now have no trouble getting work from projects.

Spread the word. ;-)

---

It is not that rare of a problem... I am experiencing a similar lack of short term work.

Mon Jul 20 17:53:41 2009 Milkyway@home Sending scheduler request: To fetch work.
Mon Jul 20 17:53:41 2009 Milkyway@home Requesting new tasks
Mon Jul 20 17:53:46 2009 Milkyway@home Scheduler request completed: got 0 new tasks
Mon Jul 20 17:53:46 2009 Milkyway@home Message from server: No work sent
Mon Jul 20 17:53:46 2009 Milkyway@home Message from server: (won't finish in time) BOINC runs 99.9% of time, computation enabled 97.1% of that
Mon Jul 20 17:53:51 2009 orbit@home Sending scheduler request: To fetch work.
Mon Jul 20 17:53:51 2009 orbit@home Requesting new tasks
Mon Jul 20 17:53:56 2009 orbit@home Scheduler request completed: got 0 new tasks
Mon Jul 20 17:53:56 2009 orbit@home Message from server: (Project has no jobs available)
Mon Jul 20 17:54:01 2009 SETI@home Sending scheduler request: To fetch work.
Mon Jul 20 17:54:01 2009 SETI@home Requesting new tasks
Mon Jul 20 17:54:26 2009 SETI@home Scheduler request completed: got 0 new tasks
Mon Jul 20 17:54:26 2009 SETI@home Message from server: No work sent
Mon Jul 20 17:54:26 2009 SETI@home Message from server: No work is available for Astropulse v5
Mon Jul 20 17:54:26 2009 SETI@home Message from server: (won't finish in time) BOINC runs 99.9% of time, computation enabled 97.1% of that
Mon Jul 20 17:54:31 2009 Einstein@Home Sending scheduler request: To fetch work.
Mon Jul 20 17:54:31 2009 Einstein@Home Requesting new tasks
Mon Jul 20 17:54:41 2009 Einstein@Home Scheduler request completed: got 0 new tasks
Mon Jul 20 17:54:41 2009 Einstein@Home Message from server: No work sent
Mon Jul 20 17:54:41 2009 Einstein@Home Message from server: (won't finish in time) BOINC runs 99.9% of time, computation enabled 97.1% of that

Unfortunately my power went out when I was typing this and I thought it was lost... What I didnt get included was that I am getting this on a Mac Pro (2x 2.66Ghx Dual Core Xeon) running v 6.6.36. Plus Seti is one of the projects not getting work -- it doesnt have any work. So the fix for windows doesnt help me.....

The post I did after reboot also includes the additional information (as I was searching the threads trying to find my post)... I am running 2 long term projects (Aqua and CPDN). It now appears I am getting a long term debt buildup that is causing BOINC to ignore the cache settings preventing any short term work from being downloaded until the cache empties of these long duration workunits....

Thanks anyway

Yes, well, sorry, I can't build this client on the Mac. You will either have to build it yourself, or wait for someone else to do so, or wait for the new client to be released by Berkeley.

See this page if you are in for the challenge to build the new client from source.

---

Everyone can hold off on downloading my client. I just saw that Alpha 6.6.38 is about to be released on all platforms.

---

i just level all my pc's to .37 today... my timing sucks this week...

Ok did some experimenting with my BOINC setup last night and reset projects that were not getting work, that didnt help, tried suspending different projects including SETI didnt help me until I suspended CPDN (4 wu's due dates all in June 2010). I got work for all of my projects except Milky Way. Hopefully what ever is jammed up in v6.6.36 scheduling gets corrected soon because I don't want to have to pay attention to how much work I have downloaded and suspend CPDN to get more....

I have the same problem, involving Chess960. I have to suspend everything in order to get any Chess960 WUs. And then the completed WUs never seem to get reported. Basically, I have to manually control BOINC every 15 mins. :(

And then the completed WUs never seem to get reported.

never being what period of time? AT LEAST 24 hours?

And which version of BOINC? The latest versions work a bit differently in this matter to the earlier versions.

And then the completed WUs never seem to get reported.

never being what period of time? AT LEAST 24 hours?

And which version of BOINC? The latest versions work a bit differently in this matter to the earlier versions.

No, not 24hrs, but at least 8hrs, which is much greater than other projects. I have my preferences set to connect every 0.1 days.

6.6.36 Win32
6.2.18 Linux x86

Ageless has posted a list of "Report" triggers here. I mentioned 24 hours because of item 4, and the text just before the list.