Message boards :
Number crunching :
about to run dry of work
Message board moderation
Author | Message |
---|---|
David@home Send message Joined: 16 Jan 03 Posts: 755 Credit: 5,040,916 RAC: 28 |
Hi, I have upgraded to Boinc Mgr v 4.43. This has been Ok for 2-3 days but today it has stopped downloading new WUs. It is now on the last WU and even if I run update on the project manually no new work is down loaded. My cache size is set to 2 days. This is all that is displayed in the message tab: 26/05/2005 17:01:10||request_reschedule_cpus: project op 26/05/2005 17:01:10||schedule_cpus: must schedule 26/05/2005 17:01:11|SETI@home|Sending scheduler request to http://setiboinc.ssl.berkeley.edu/sah_cgi/cgi 26/05/2005 17:01:12|SETI@home|Scheduler request to http://setiboinc.ssl.berkeley.edu/sah_cgi/cgi succeeded The server status seems to be OK, any ideas on how to get some WUs downloaded? Thanks |
Pooh Bear 27 Send message Joined: 14 Jul 03 Posts: 3224 Credit: 4,603,826 RAC: 0 |
Everyone will be running dry of work. The splitters are offline and not making new work, as the team is going to be moving stuff to new machines. This info is from the front page: May 24, 2005 The splitters and assimilator are offline for a day or so while we relocate the backend science database. My movie https://vimeo.com/manage/videos/502242 |
David@home Send message Joined: 16 Jan 03 Posts: 755 Credit: 5,040,916 RAC: 28 |
Everyone will be running dry of work. The splitters are offline and not making new work, as the team is going to be moving stuff to new machines. I saw that news item, I did not think it was related as 1) my other PC is downloading WUs OK, it got one very recently. 2) the status page is still showing plenty of WUs ready for downloading 3) I do not see any messages to the effect that no work is available The PC in question has not downloaded anything for a day and is just emptying the cache. Is the status page out of date in terms of available WUs or has the scheduler on this specific PC gone wrong? |
Pooh Bear 27 Send message Joined: 14 Jul 03 Posts: 3224 Credit: 4,603,826 RAC: 0 |
My suspicion is the scheduler in 4.43 was not fully ready for release, and has some issues. I see there is over 100K as of about 20 minutes ago. This goes down a few thousand an hour, but with the weekend coming, I know people will be trying to get extra WUs for the weekend, and it will run dry quickly. I hope they can make the move when the WU cache is gone (which I expect to be sometime around 00:00 - 03:00 UTC. I do other projects, so I am not so worried about being out of work, except that 2 of the other project also are having some issues. So I might actually run out of work this weekend. No biggie to me, cause I am here to help the science, and not worried if there is no work. Life goes on. My movie https://vimeo.com/manage/videos/502242 |
David@home Send message Joined: 16 Jan 03 Posts: 755 Credit: 5,040,916 RAC: 28 |
My suspicion is the scheduler in 4.43 was not fully ready for release, and has some issues. I see there is over 100K as of about 20 minutes ago. This goes down a few thousand an hour, but with the weekend coming, I know people will be trying to get extra WUs for the weekend, and it will run dry quickly. I think your correct about the scheduler. I also have einstein@home but that project is suspended and set not to download work. I think the scheduler must still be thinking of giving the suspended project an allocation of CPU time. I just detached from einstein@home, I did not want to do that as I would have lost my machine id number as an early user I kinda liked the low machine number. As soon as I did the detach it tried to download some seti@home work: 26/05/2005 17:36:09|SETI@home|Message from server: Not sending work - last RPC too recent: 2 sec Still no work but at least it tried. Looks like a bug in the scheduler when you have a suspended project. How do you report potential bugs to the developers? |
keputnam Send message Joined: 2 Jul 99 Posts: 242 Credit: 2,736,564 RAC: 3 |
In one of the other threads on the subject, JM7 suggested resetting any projects that have a very large positive Long Term Debt. A project like Pirates or LHC that has not sent out work recently probably has a huge positive LTD. If you use BoincView, from the projects tab, look all the way to the right to see LTD This "tweak" worked for me (I have both Pirates and LHC on all my machines). Before I even finished multiple resets on both of these, the CC was already downloading another Seti WU. |
rsisto Send message Joined: 30 Jul 03 Posts: 135 Credit: 729,936 RAC: 0 |
It is not necesary to reset all projects, just to suspend them. This will make long term debt 0. |
David@home Send message Joined: 16 Jan 03 Posts: 755 Credit: 5,040,916 RAC: 28 |
It is not necesary to reset all projects, just to suspend them. This will make long term debt 0. In my case the second project (einstein@home) was suspended and it was also set not to download work. I have now detached from the project. I wish I had tried a reset but it is too late now :-( My best guess is that the scheduler was building up a debt for einstein@home even though it was suspended. I will have a look at boincview, may that should go in the wish list for Boinc mgr to show these debt values. If I had seen those maybe I could have worked out what was going on. |
David@home Send message Joined: 16 Jan 03 Posts: 755 Credit: 5,040,916 RAC: 28 |
Just a quick update. Firstly a big thanks to whoever came in over the w/e to sort out the server issues. Much appreciated as I am sure you had better things planned for the weekend. As I ran out of SETI work I re-signed up for Einstein@home. I am now using version 4.44 of boinc.exe so was able to test the long-term debt stuff with this release. When seti@home came back SETI got some WUs OK so I suspended einstein@home and set einstein not to download more work. Alas the issue is still present in 4.44 in that SETI@home worked though its cache and would not download due to it building up a negative debt. i.e. seti@home was the only active project but it still would not download, it almost ran out of work. I tried resetting the einstein@home project but this had no effect. I had to detach from einstein@home, as soon as I did that seti@home downloaded more work. I think if a project is suspended then the long-term debt should not be applied. |
John McLeod VII Send message Joined: 15 Jul 99 Posts: 24806 Credit: 790,712 RAC: 0 |
Just a quick update. The LT debt of projects that are suspended, are marked as no download and have no work (graceful suspended) or have communications deferred and have no work on the host (not supplying work) should not have their LT debt move. There is a bug that may have made this not work quite right in 4.44. I believe it to be fixed in 4.45. In any case if a CPU is idle the client is supposed to get work from any project (even one with a negative LT debt). However, this does not work quite right in 4.44 either, and I believe that it is fixed in 4.45. BOINC WIKI |
David@home Send message Joined: 16 Jan 03 Posts: 755 Credit: 5,040,916 RAC: 28 |
Just a quick update. Many thanks for the update John |
©2024 University of California
SETI@home and Astropulse are funded by grants from the National Science Foundation, NASA, and donations from SETI@home volunteers. AstroPulse is funded in part by the NSF through grant AST-0307956.