Tasks stuck in 'Ready to Report' status


Advanced search

Message boards : Number crunching : Tasks stuck in 'Ready to Report' status

AuthorMessage
Profile John C MacAlister
Send message
Joined: Sep 12 10
Posts: 28
Credit: 229,803
RAC: 0
Message 15241 - Posted 3 Feb 2011 12:27:06 UTC

    Last modified: 3 Feb 2011 12:27:42 UTC

    Hi!

    My tasks are stuck. Some days they report automatically, others, like today, they do not. Am I missing a preference setting?

    Thanks,

    John
    ____________

    hardy
    Volunteer moderator
    Project administrator
    Project developer
    Avatar
    Send message
    Joined: Feb 18 09
    Posts: 198
    Credit: 127,936
    RAC: 19
    Message 15242 - Posted 3 Feb 2011 17:01:04 UTC - in response to Message 15241.

      There can be long delays between when the client reports jobs (this is to reduce the number of communications the server needs to handle). Is this what you mean? If you wait several hours (while connected to the internet), do they report? (Check log messages. Use the advanced->"do network communication" option if you need to test.)

      Profile John C MacAlister
      Send message
      Joined: Sep 12 10
      Posts: 28
      Credit: 229,803
      RAC: 0
      Message 15243 - Posted 3 Feb 2011 17:26:53 UTC - in response to Message 15242.

        Thank you, I will wait and review the messages later. My three PCs are on line at least 15 hours a day
        ____________

        Profile mikey
        Avatar
        Send message
        Joined: Mar 23 07
        Posts: 6611
        Credit: 5,978,056
        RAC: 2,637
        Message 15246 - Posted 4 Feb 2011 10:08:59 UTC - in response to Message 15241.

          Hi!

          My tasks are stuck. Some days they report automatically, others, like today, they do not. Am I missing a preference setting?

          Thanks,

          John


          Units will always report no later than their deadline or whenever you need new work, whichever comes first. If your cache is large it could be several days between contacts, as hardy said this helps the Project keep the need for faster and more bandwidth capabilities to a minimum. As you can imagine with a few hundred thousand of us all trying to get and return workunits it can be quite busy on the Servers end.

          There is a way to make it happen manually though, Open the Boinc Manager, down by the clock, by right clicking on it and clicking Open, then click on the Projects tab and then, after highlighting the project you want, click update.

          Profile John C MacAlister
          Send message
          Joined: Sep 12 10
          Posts: 28
          Credit: 229,803
          RAC: 0
          Message 15248 - Posted 4 Feb 2011 11:04:59 UTC - in response to Message 15246.

            Thank you for your note. I have been doing the manual updating and will leave my computers running unattended for the next while. Yesterday, I looked at progress on my computers ID: 182854 (Linux - 4 processors) and ID: 184768 (Linux - 2 processors). During the night, ID: 182854 changed to running tasks on three processors and ID: 184768 changed to running tasks on one processor, only. I changed my computing preferences “On multiprocessors, use at most 8 processors” (was 4) and “On multiprocessors, use at most Enforced by version 6.1+” to 100% of the processors (was 90%).

            ID: 182854 immediately returned to processing four tasks while ID: 184768 refuses to process more than one concurrent task. The Linux performance monitor shows both processors sharing one task – one going from low to full utilization as the other processor goes from full to low utilization. I think I may have a BIOS issue with my computer ID: 184768 and until I retired in 2008 I had never “looked under the hood” and only recently started using Linux. I am now way out of my depth…. I need to ask my system builder to help.

            ____________

            Profile mikey
            Avatar
            Send message
            Joined: Mar 23 07
            Posts: 6611
            Credit: 5,978,056
            RAC: 2,637
            Message 15251 - Posted 5 Feb 2011 10:29:07 UTC - in response to Message 15248.

              Thank you for your note. I have been doing the manual updating and will leave my computers running unattended for the next while. Yesterday, I looked at progress on my computers ID: 182854 (Linux - 4 processors) and ID: 184768 (Linux - 2 processors). During the night, ID: 182854 changed to running tasks on three processors and ID: 184768 changed to running tasks on one processor, only. I changed my computing preferences “On multiprocessors, use at most 8 processors” (was 4) and “On multiprocessors, use at most Enforced by version 6.1+” to 100% of the processors (was 90%).

              ID: 182854 immediately returned to processing four tasks while ID: 184768 refuses to process more than one concurrent task. The Linux performance monitor shows both processors sharing one task – one going from low to full utilization as the other processor goes from full to low utilization. I think I may have a BIOS issue with my computer ID: 184768 and until I retired in 2008 I had never “looked under the hood” and only recently started using Linux. I am now way out of my depth…. I need to ask my system builder to help.


              Go into the Boinc Manager and click on Advanced, preferences, processor usage tab and look at what the line says that says "when processor usage is less than [] percent (0 means no restriction)" Change that to a zero and click okay at the bottom of the page and your other cpu's should pick up and start crunching. This setting tells Boinc to stop crunching when it is busy doing something else, normally you would want it crunching 24/7 but for some reason the default of 25 is in there. If you use your pc for other things play with this number and figure out what works best for you. On the next line down you can set Boinc so it only crunches during certain hours, which works well for those with a home business or who telework for example.

              Profile John C MacAlister
              Send message
              Joined: Sep 12 10
              Posts: 28
              Credit: 229,803
              RAC: 0
              Message 15283 - Posted 11 Feb 2011 13:48:38 UTC - in response to Message 15251.

                Thank you for your note. I have been doing the manual updating and will leave my computers running unattended for the next while. Yesterday, I looked at progress on my computers ID: 182854 (Linux - 4 processors) and ID: 184768 (Linux - 2 processors). During the night, ID: 182854 changed to running tasks on three processors and ID: 184768 changed to running tasks on one processor, only. I changed my computing preferences “On multiprocessors, use at most 8 processors” (was 4) and “On multiprocessors, use at most Enforced by version 6.1+” to 100% of the processors (was 90%).

                ID: 182854 immediately returned to processing four tasks while ID: 184768 refuses to process more than one concurrent task. The Linux performance monitor shows both processors sharing one task – one going from low to full utilization as the other processor goes from full to low utilization. I think I may have a BIOS issue with my computer ID: 184768 and until I retired in 2008 I had never “looked under the hood” and only recently started using Linux. I am now way out of my depth…. I need to ask my system builder to help.


                Go into the Boinc Manager and click on Advanced, preferences, processor usage tab and look at what the line says that says "when processor usage is less than [] percent (0 means no restriction)" Change that to a zero and click okay at the bottom of the page and your other cpu's should pick up and start crunching. This setting tells Boinc to stop crunching when it is busy doing something else, normally you would want it crunching 24/7 but for some reason the default of 25 is in there. If you use your pc for other things play with this number and figure out what works best for you. On the next line down you can set Boinc so it only crunches during certain hours, which works well for those with a home business or who telework for example.


                Many thanks for your help, Mikey. Everything is now running fine.

                ____________

                Profile mikey
                Avatar
                Send message
                Joined: Mar 23 07
                Posts: 6611
                Credit: 5,978,056
                RAC: 2,637
                Message 15287 - Posted 12 Feb 2011 9:50:05 UTC - in response to Message 15283.

                  Thank you for your note. I have been doing the manual updating and will leave my computers running unattended for the next while. Yesterday, I looked at progress on my computers ID: 182854 (Linux - 4 processors) and ID: 184768 (Linux - 2 processors). During the night, ID: 182854 changed to running tasks on three processors and ID: 184768 changed to running tasks on one processor, only. I changed my computing preferences “On multiprocessors, use at most 8 processors” (was 4) and “On multiprocessors, use at most Enforced by version 6.1+” to 100% of the processors (was 90%).

                  ID: 182854 immediately returned to processing four tasks while ID: 184768 refuses to process more than one concurrent task. The Linux performance monitor shows both processors sharing one task – one going from low to full utilization as the other processor goes from full to low utilization. I think I may have a BIOS issue with my computer ID: 184768 and until I retired in 2008 I had never “looked under the hood” and only recently started using Linux. I am now way out of my depth…. I need to ask my system builder to help.


                  Go into the Boinc Manager and click on Advanced, preferences, processor usage tab and look at what the line says that says "when processor usage is less than [] percent (0 means no restriction)" Change that to a zero and click okay at the bottom of the page and your other cpu's should pick up and start crunching. This setting tells Boinc to stop crunching when it is busy doing something else, normally you would want it crunching 24/7 but for some reason the default of 25 is in there. If you use your pc for other things play with this number and figure out what works best for you. On the next line down you can set Boinc so it only crunches during certain hours, which works well for those with a home business or who telework for example.


                  Many thanks for your help, Mikey. Everything is now running fine.


                  I am glad to hear it, crunch on!

                  robertmiles
                  Send message
                  Joined: May 21 08
                  Posts: 56
                  Credit: 690,094
                  RAC: 606
                  Message 15330 - Posted 21 Feb 2011 3:46:36 UTC

                    Do you know just what's happened to the UCT Malaria BOINC project? After their website became unreachable, it was about two months before the last few workunits I had from them were able to upload the output files (or until BOINC decided they were too old to keep, possibly). Since then, those workunits have been stuck in 'Ready to Report' status for more than 4 months more.

                    Profile mikey
                    Avatar
                    Send message
                    Joined: Mar 23 07
                    Posts: 6611
                    Credit: 5,978,056
                    RAC: 2,637
                    Message 15331 - Posted 21 Feb 2011 9:49:49 UTC - in response to Message 15330.

                      Last modified: 21 Feb 2011 9:53:18 UTC

                      Do you know just what's happened to the UCT Malaria BOINC project? After their website became unreachable, it was about two months before the last few workunits I had from them were able to upload the output files (or until BOINC decided they were too old to keep, possibly). Since then, those workunits have been stuck in 'Ready to Report' status for more than 4 months more.


                      I believe the UCT website closed down. There was a thread back in June about this
                      http://www.malariacontrol.net/forum_thread.php?id=1007

                      robertmiles
                      Send message
                      Joined: May 21 08
                      Posts: 56
                      Credit: 690,094
                      RAC: 606
                      Message 15337 - Posted 21 Feb 2011 20:11:15 UTC - in response to Message 15331.

                        Do you know just what's happened to the UCT Malaria BOINC project? After their website became unreachable, it was about two months before the last few workunits I had from them were able to upload the output files (or until BOINC decided they were too old to keep, possibly). Since then, those workunits have been stuck in 'Ready to Report' status for more than 4 months more.


                        I believe the UCT website closed down. There was a thread back in June about this
                        http://www.malariacontrol.net/forum_thread.php?id=1007


                        I've now detached from that project, even though there was no way to report the last batch of workunits first.

                        Larry Paulson
                        Send message
                        Joined: Feb 25 11
                        Posts: 1
                        Credit: 101,234
                        RAC: 212
                        Message 15494 - Posted 18 Mar 2011 16:33:22 UTC

                          I had seven or eight tasks that had been stuck in “Ready to Report" status for what seemed to be a considerable time (over an hour?) And shortly before I would have to shut down for the weekend (which would have made those tasks overdue).

                          (No, I don't leave my machine on unless I'm actually using it.)

                          I decided to abort all downloaded tasks that hadn't commenced execution. This seems to have given Boinc a kick, because the stuck tasks uploaded immediately.

                          Snagletooth
                          Send message
                          Joined: Dec 24 09
                          Posts: 10
                          Credit: 110,741
                          RAC: 46
                          Message 15500 - Posted 20 Mar 2011 11:44:03 UTC - in response to Message 15494.

                            I had seven or eight tasks that had been stuck in “Ready to Report" status for what seemed to be a considerable time (over an hour?) And shortly before I would have to shut down for the weekend (which would have made those tasks overdue).
                            I decided to abort all downloaded tasks that hadn't commenced execution. This seems to have given Boinc a kick, because the stuck tasks uploaded immediately.



                            They weren't stuck. BOINC tries to minimize the load on the project servers by grouping together contacts that require searching the database. It will hold tasks up to 24 hours before reporting. From the BOINC FAQ:

                            When you're done with a task in BOINC, it uploads the data that is has collected to a directory on the project's server. Then a notice has to be sent to the database that you have finished this task with whatever outcome you had, be it that you finished it correctly, had an error or aborted it.

                            This notice is the "ready to report" part.
                            BOINC will do this automatically in 24 hours, but if you can't wait that long, open BOINC Manager, Advanced view, Projects tab, select the project the task is from and press the Update button. That reports the tasks that are ready to report immediately.


                            When you aborted tasks BOINC would have tried to replace them with new ones to refill your queue. The reporting of uploaded tasks would have been combined with the request for new tasks. It may appear that you gave BOINC a kick but really it was just behaving as designed.

                            Best,
                            Snags

                            Profile mikey
                            Avatar
                            Send message
                            Joined: Mar 23 07
                            Posts: 6611
                            Credit: 5,978,056
                            RAC: 2,637
                            Message 15501 - Posted 20 Mar 2011 12:21:15 UTC - in response to Message 15500.

                              Last modified: 20 Mar 2011 12:21:45 UTC

                              I had seven or eight tasks that had been stuck in “Ready to Report" status for what seemed to be a considerable time (over an hour?) And shortly before I would have to shut down for the weekend (which would have made those tasks overdue).
                              I decided to abort all downloaded tasks that hadn't commenced execution. This seems to have given Boinc a kick, because the stuck tasks uploaded immediately.



                              They weren't stuck. BOINC tries to minimize the load on the project servers by grouping together contacts that require searching the database. It will hold tasks up to 24 hours before reporting. From the BOINC FAQ:

                              When you're done with a task in BOINC, it uploads the data that is has collected to a directory on the project's server. Then a notice has to be sent to the database that you have finished this task with whatever outcome you had, be it that you finished it correctly, had an error or aborted it.


                              To expand a little bit Larry, each connection to a projects Server takes approximately 0.3 seconds to open a connection, a couple of seconds to actually transfer the data depending on the connection speed, and then some time spent by the Server itself trying to figure what to do with what you just sent in. If you multiply this by all of the people that are attached to a project, it can take all day long and multiple connections and multiple servers to make this happen smoothly. So the projects try to group things so that at each connection multiple things are done, the 0.3 seconds to make a connection is a fixed amount and therefore is there whether you send in one unit or hundreds of them.

                              Malaria itself has 49,068 Users, which is +17 from yesterday and 110,274 computers attached to it. That is ALOT of 0.3 second connections for each and every computer to make in a day!
                              ____________

                              Profile John C MacAlister
                              Send message
                              Joined: Sep 12 10
                              Posts: 28
                              Credit: 229,803
                              RAC: 0
                              Message 15509 - Posted 21 Mar 2011 13:04:23 UTC

                                Thanks for the clear explanation, Mikey

                                John

                                Post to thread

                                Message boards : Number crunching : Tasks stuck in 'Ready to Report' status


                                Return to malariacontrol.net main page


                                Copyright © 2014 africa@home