error: Output is empty


Advanced search

Message boards : SZTAKI Desktop Grid : error: Output is empty

AuthorMessage
LiborA
Send message
Joined: Mar 17 06
Posts: 25
Credit: 12,902
RAC: 0
Message 3666 - Posted 27 Jul 2006 6:10:56 UTC

    Now I report 3 WU with error message \"Output is empty\"
    http://szdg.lpds.sztaki.hu/szdg/result.php?resultid=207
    http://szdg.lpds.sztaki.hu/szdg/result.php?resultid=155
    http://szdg.lpds.sztaki.hu/szdg/result.php?resultid=152

    Whats the problem?

    Profile kadam
    Project administrator
    Avatar
    Send message
    Joined: May 25 05
    Posts: 589
    Credit: 38,614
    RAC: 0
    Message 3689 - Posted 28 Jul 2006 13:42:53 UTC - in response to Message 3666.

      If the process starts correctly (I mean, there is no linux lib problems and etc.) than there is a very little chance that the process won\'t produce an output. I don\'t know how, but it can happen, although it shouldn\'t. I\'ll explore around this next week.
      ____________
      If you like BOINC, you may also find CaretCursor to be appealing.

      larry1186
      Send message
      Joined: Sep 25 06
      Posts: 37
      Credit: 18,502
      RAC: 0
      Message 4877 - Posted 29 Nov 2006 18:02:00 UTC - in response to Message 3689.

        If the process starts correctly (I mean, there is no linux lib problems and etc.) than there is a very little chance that the process won\'t produce an output. I don\'t know how, but it can happen, although it shouldn\'t. I\'ll explore around this next week.


        @Adam: Any update on this issue?

        The \"empty output\" is mentioned several times recently (How Long Is Short, New application L2.04/W2.02 Feedback, New application L2.03/W2.01/M2.00 - Feedback, Why so many \"client errors\" recently?) and from what I have read and experienced, I will say that the empty output shows up when a WU is interrupted, but not every time it is interrupted. Some of my WU have been interrupted and resumed just fine, finished, and were granted credit. By interrupted I mean BOINC is exited or the task is removed from memory.

        Some people have suggested that, to get a WU going again after seemingly being stuck for a long time, one just needs to exit and restart BOINC. This will allow the WU to magically jump ahead and resume, but I am afraid it also causes the empty output.

        As far as I can tell, it does not matter which application version is being used:
        W2.04, W2.02, W2.01, W2.00 but I don\'t know if it is platform specific. Perhaps BOINC 5.4.11 is to blame? It seems that the one-line WUs are more likely to have empty results since if other lines are crunched successfully, then the output of the WU would not be empty (but the output of the line that was interrupted would be).

        A couple of points I am curious about:
        1) Has anybody allowed a WU to finish completely uninterrupted and still get the empty result?
        2) Has anybody interrupted the WU to get it going again with the \"magical jump\" and not get the empty result? on a one-line WU?
        3) Has anybody using MacOS or Linux received an empty result?

        If anybody has some input or additional information, that would be great. The WUs should be able to be interrupted and resume without any problems (i.e. - no empty result).
        ____________
        Don't get distracted by shiny objects.

        Stick
        Send message
        Joined: Jun 12 06
        Posts: 193
        Credit: 66,244
        RAC: 0
        Message 4880 - Posted 29 Nov 2006 22:02:01 UTC - in response to Message 3689.

          Last modified: 29 Nov 2006 22:36:27 UTC

          If the process starts correctly (I mean, there is no linux lib problems and etc.) than there is a very little chance that the process won\'t produce an output. I don\'t know how, but it can happen, although it shouldn\'t. I\'ll explore around this next week.


          Adam,

          I have also posted several times on this issue - most recently here.

          And (a while back) here and here and here.

          Hopefully, these observations will help, as you \"look into it\".

          Stick

          ____________

          NJMHoffmann
          Send message
          Joined: May 19 06
          Posts: 18
          Credit: 10,221
          RAC: 0
          Message 4898 - Posted 1 Dec 2006 6:59:13 UTC - in response to Message 4877.

            Some people have suggested that, to get a WU going again after seemingly being stuck for a long time, one just needs to exit and restart BOINC. This will allow the WU to magically jump ahead and resume, but I am afraid it also causes the empty output.

            Even if it doesn\'t produce an empty output line, I think you can be quite sure that those \"jumping\" WUs will not validate. There must be something wrong in the restart logic.

            Norbert
            ____________

            Stick
            Send message
            Joined: Jun 12 06
            Posts: 193
            Credit: 66,244
            RAC: 0
            Message 5094 - Posted 16 Dec 2006 17:40:05 UTC - in response to Message 4880.

              @ Adam,

              I have added a few posts to this thread regarding the \"Output is empty\" problem. And I am hoping that you might be able to give us an update relating to the \"exlporing\" you did into the issue.

              Stick

              If the process starts correctly (I mean, there is no linux lib problems and etc.) than there is a very little chance that the process won\'t produce an output. I don\'t know how, but it can happen, although it shouldn\'t. I\'ll explore around this next week.


              Adam,

              I have also posted several times on this issue - most recently here.

              And (a while back) here and here and here.

              Hopefully, these observations will help, as you \"look into it\".

              Stick


              ____________

              miketoth1001
              Avatar
              Send message
              Joined: Apr 3 06
              Posts: 20
              Credit: 27,967
              RAC: 0
              Message 5545 - Posted 24 Jan 2007 18:49:55 UTC - in response to Message 4877.

                1) Has anybody allowed a WU to finish completely uninterrupted and still get the empty result?
                2) Has anybody interrupted the WU to get it going again with the \"magical jump\" and not get the empty result? on a one-line WU?
                3) Has anybody using MacOS or Linux received an empty result?

                If anybody has some input or additional information, that would be great. The WUs should be able to be interrupted and resume without any problems (i.e. - no empty result).


                1) I just put online 2 Linux hosts, 6 CPUs in total, and they\'ll crunch 24/7 just for this project. We\'ll see what happens, as they\'ll not be interrupted for 30 days.

                2) I\'ve never noticed.

                3) Yes. On Linux, I\'ve gotten the \"empty result\".

                I wanted to put these hosts, and a copuple more to come in the next day or two, to see if they\'ll actually produce good results. If they still show the empty error, or don\'t help reaching \"concensus\", then we can be pretty sure that there\'s a different problem that\'s running parallel to the checkpointing problem.

                gwg
                Avatar
                Send message
                Joined: Aug 1 06
                Posts: 58
                Credit: 213,886
                RAC: 0
                Message 5546 - Posted 24 Jan 2007 20:41:51 UTC - in response to Message 5545.

                  1) Has anybody allowed a WU to finish completely uninterrupted and still get the empty result?
                  2) Has anybody interrupted the WU to get it going again with the \"magical jump\" and not get the empty result? on a one-line WU?
                  3) Has anybody using MacOS or Linux received an empty result?

                  If anybody has some input or additional information, that would be great. The WUs should be able to be interrupted and resume without any problems (i.e. - no empty result).


                  1) I just put online 2 Linux hosts, 6 CPUs in total, and they\'ll crunch 24/7 just for this project. We\'ll see what happens, as they\'ll not be interrupted for 30 days.

                  2) I\'ve never noticed.

                  3) Yes. On Linux, I\'ve gotten the \"empty result\".

                  I wanted to put these hosts, and a copuple more to come in the next day or two, to see if they\'ll actually produce good results. If they still show the empty error, or don\'t help reaching \"concensus\", then we can be pretty sure that there\'s a different problem that\'s running parallel to the checkpointing problem.


                  I have had many empty results on a G3 Dual processor Mac running OS X 10.4, which runs 24/7, but SZTAKI gets interrupted for other BOINC processes, and although I have set up for the processes to remain in memory, I keep getting BOINC timeouts at process switch time.

                  If a process doesn\'t get that 30 second heartbeat from BOINC, it kills itself. I\'m not sure why I get these timeouts, but it is always associated with I/O, usually with BOINCstats BAM! Bandwidth isn\'t the problem, as I have 512/128 ADSL broadband.

                  I suspect that BOINC is programmed to wait on I/O and doesn\'t send out the heartbeat while waiting.

                  My conclusion is that SZTAKI never returns a zero result unless it has been killed, either due to machine shutdown or heartbeat timeout.

                  George
                  ------

                  ____________
                  Dr George W Gerrity
                  4 Coral Place
                  Campbell, ACT 2612
                  AUSTRALIA

                  Ph: +61 2 6156 0286
                  Time: +10 hours (ref GMT)
                  PGP RSA Public Key Fingerprint:
                  73EF 318A DFF5 EB8A 6810 49AC 0763 AF07

                  miketoth1001
                  Avatar
                  Send message
                  Joined: Apr 3 06
                  Posts: 20
                  Credit: 27,967
                  RAC: 0
                  Message 5549 - Posted 25 Jan 2007 1:40:08 UTC - in response to Message 5546.

                    Last modified: 25 Jan 2007 1:42:13 UTC

                    I have had many empty results on a G3 Dual processor Mac running OS X 10.4, which runs 24/7, but SZTAKI gets interrupted for other BOINC processes, and although I have set up for the processes to remain in memory, I keep getting BOINC timeouts at process switch time.

                    If a process doesn\'t get that 30 second heartbeat from BOINC, it kills itself. I\'m not sure why I get these timeouts, but it is always associated with I/O, usually with BOINCstats BAM! Bandwidth isn\'t the problem, as I have 512/128 ADSL broadband.

                    I suspect that BOINC is programmed to wait on I/O and doesn\'t send out the heartbeat while waiting.

                    My conclusion is that SZTAKI never returns a zero result unless it has been killed, either due to machine shutdown or heartbeat timeout.


                    I\'m not sure about that. On one project I normally crunch for, I\'ve had the out file showing no heartbeat for 82K CPU seconds. That means that it should allow past 30 seconds.

                    Also, not every box is setup to run just one project 24/7. I have one box, with 8 CPUs, crunching for about 10 projects. Only project I know that has problems like here is uFluids. They say to suspend to memory, but it still generates errors about 15 to 20 percent of the time.

                    Anyway. So far, neither box has gotten the infamous series of WUs. They\'ve all been the 50 line ones, and crunched them in far less time than the 1 liner WUs. But, they\'re just crunching for this project. We\'ll see what it\'s doing in a week or two.

                    BTW. I should mention that these 2 boxes are new Debian installs. So far, no lib errors. (Although, it seems that whenever one box contacts the server, a new host number gets generated.)

                    gwg
                    Avatar
                    Send message
                    Joined: Aug 1 06
                    Posts: 58
                    Credit: 213,886
                    RAC: 0
                    Message 5562 - Posted 26 Jan 2007 0:06:22 UTC - in response to Message 5549.

                      I have had many empty results on a G3 Dual processor Mac running OS X 10.4, which runs 24/7, but SZTAKI gets interrupted for other BOINC processes, and although I have set up for the processes to remain in memory, I keep getting BOINC timeouts at process switch time.

                      If a process doesn\'t get that 30 second heartbeat from BOINC, it kills itself. I\'m not sure why I get these timeouts, but it is always associated with I/O, usually with BOINCstats BAM! Bandwidth isn\'t the problem, as I have 512/128 ADSL broadband.

                      I suspect that BOINC is programmed to wait on I/O and doesn\'t send out the heartbeat while waiting.

                      My conclusion is that SZTAKI never returns a zero result unless it has been killed, either due to machine shutdown or heartbeat timeout.


                      I\'m not sure about that. On one project I normally crunch for, I\'ve had the out file showing no heartbeat for 82K CPU seconds. That means that it should allow past 30 seconds.

                      Also, not every box is setup to run just one project 24/7. I have one box, with 8 CPUs, crunching for about 10 projects. Only project I know that has problems like here is uFluids. They say to suspend to memory, but it still generates errors about 15 to 20 percent of the time.

                      Anyway. So far, neither box has gotten the infamous series of WUs. They\'ve all been the 50 line ones, and crunched them in far less time than the 1 liner WUs. But, they\'re just crunching for this project. We\'ll see what it\'s doing in a week or two.

                      BTW. I should mention that these 2 boxes are new Debian installs. So far, no lib errors. (Although, it seems that whenever one box contacts the server, a new host number gets generated.)


                      The heartbeat misses are only on my big Mac. The small ones only run one process — not STAKI — also 24/7. I also run a dual 2.7GHz Athlon CPU under Windows XP Pro. I get no heartbeat misses, but Windows rarely runs more than 24 hours without some sort of crash. Thus, I get problems with SZTAKI because I have to keep rebooting Windows.

                      George
                      ------
                      ____________
                      Dr George W Gerrity
                      4 Coral Place
                      Campbell, ACT 2612
                      AUSTRALIA

                      Ph: +61 2 6156 0286
                      Time: +10 hours (ref GMT)
                      PGP RSA Public Key Fingerprint:
                      73EF 318A DFF5 EB8A 6810 49AC 0763 AF07

                      robert.mouris
                      Send message
                      Joined: Nov 3 05
                      Posts: 129
                      Credit: 4,124,189
                      RAC: 0
                      Message 5596 - Posted 30 Jan 2007 16:57:41 UTC - in response to Message 5562.

                        Last modified: 30 Jan 2007 16:59:48 UTC

                        Thus, I get problems with SZTAKI because I have to keep rebooting Windows.

                        I had several system problems these last days and I am also forced to reboot the system for other reasons from time to time, so I send back unusable results that may trigger the \"too many results\" error and penalize fellow crunchers and the project itself. I\'m giving up hope to see the 5/8/5 limit being raised. But I wonder if there isn\'t a way to wipe the slate clean by fiddling with the files in the project folder, so that I can restart my WU from scratch. Any ideas?

                        Robert
                        ____________

                        Odysseus
                        Avatar
                        Send message
                        Joined: Feb 27 06
                        Posts: 212
                        Credit: 221,397
                        RAC: 0
                        Message 5598 - Posted 30 Jan 2007 20:14:31 UTC - in response to Message 5596.

                          But I wonder if there isn\'t a way to wipe the slate clean by fiddling with the files in the project folder, so that I can restart my WU from scratch. Any ideas?

                          I believe that if you back up all your BOINC data (including not only the project folders but client_state.xml and so on) immediately after receiving new work, later on you can start those tasks over at the beginning by stopping BOINC, replacing the data files and relaunching. (I wouldn’t recommend this if you’re running multiple projects.)
                          ____________

                          robert.mouris
                          Send message
                          Joined: Nov 3 05
                          Posts: 129
                          Credit: 4,124,189
                          RAC: 0
                          Message 5637 - Posted 3 Feb 2007 14:57:56 UTC - in response to Message 5598.

                            Last modified: 3 Feb 2007 15:12:00 UTC

                            I believe that if you back up all your BOINC data (including not only the project folders but client_state.xml and so on) immediately after receiving new work, later on you can start those tasks over at the beginning by stopping BOINC, replacing the data files and relaunching. (I wouldn’t recommend this if you’re running multiple projects.)

                            Thank you for the hint. This may work, but is (too) hard to do all the time. I hoped to find an easier way.

                            Anyhow, things are getting out of hand. Several of my WUs are issuing their 8th result and as there are many \"empty\" results, it is useless to crunch them, they will invariably exceed the 5/8/5 limit because of the \"too many results\" error, like this one. Others will fail because of the \"too many success results\" error, like this one. The computer with the 209732 sec (more than 58 hours) is mine. All the other results are \"empty\". This WU fill fail and a new WU will be created, starting from scratch.

                            Under these conditions I think that it is useless for me to continue. So I decided to put my computers on \"no new work\". I will finish the 158 WUs that I still have (unless they have a canonical result or have failed before they start). I will resume my participation as soon as one the following conditions is met:

                            • the project gets a working checkpoint system
                            • the 5/8/5 limit is again raised to 10/20/10
                            • the new application announced by Ádám on January 8th arrives (I don\'t know if the BinSys application will be finished in a few weeks or a few years... no information is available)
                            • someone succeeds in convincing me to resume immediately.


                            Until that day, bye!

                            Robert
                            ____________

                            Philip Martin Kryder
                            Send message
                            Joined: Apr 3 06
                            Posts: 73
                            Credit: 10,801
                            RAC: 0
                            Message 5641 - Posted 4 Feb 2007 2:43:27 UTC - in response to Message 5637.

                              I believe that if you back up all your BOINC data (including not only the project folders but client_state.xml and so on) immediately after receiving new work, later on you can start those tasks over at the beginning by stopping BOINC, replacing the data files and relaunching. (I wouldn’t recommend this if you’re running multiple projects.)

                              Thank you for the hint. This may work, but is (too) hard to do all the time. I hoped to find an easier way.

                              Anyhow, things are getting out of hand. Several of my WUs are issuing their 8th result and as there are many \"empty\" results, it is useless to crunch them, they will invariably exceed the 5/8/5 limit because of the \"too many results\" error, like this one. Others will fail because of the \"too many success results\" error, like this one. The computer with the 209732 sec (more than 58 hours) is mine. All the other results are \"empty\". This WU fill fail and a new WU will be created, starting from scratch.

                              Under these conditions I think that it is useless for me to continue. So I decided to put my computers on \"no new work\". I will finish the 158 WUs that I still have (unless they have a canonical result or have failed before they start). I will resume my participation as soon as one the following conditions is met:

                              • the project gets a working checkpoint system
                              • the 5/8/5 limit is again raised to 10/20/10
                              • the new application announced by Ádám on January 8th arrives (I don\'t know if the BinSys application will be finished in a few weeks or a few years... no information is available)
                              • someone succeeds in convincing me to resume immediately.


                              Until that day, bye!

                              Robert




                              yeah - me too - watching from the sidelines and hoping for news of \"fixes\" and a return to the 10/20/10.

                              I\'ve asked several times how 5/8/5 is better -
                              Has anyone seen an answer?





                              ____________

                              robert.mouris
                              Send message
                              Joined: Nov 3 05
                              Posts: 129
                              Credit: 4,124,189
                              RAC: 0
                              Message 5644 - Posted 4 Feb 2007 9:47:30 UTC

                                As far as I know, the 10/20/10 limit was temporary and was just supposed to make the WUs from mid-September come to a success. The real solution is of course a good checkpoint system, then we don\'t need the higher limit.
                                ____________

                                Profile Nightbird
                                Forum moderator
                                Avatar
                                Send message
                                Joined: Jul 12 05
                                Posts: 920
                                Credit: 114,924
                                RAC: 0
                                Message 5645 - Posted 4 Feb 2007 13:04:24 UTC - in response to Message 5637.

                                  Last modified: 8 Feb 2007 20:18:33 UTC

                                  I believe that if you back up all your BOINC data (including not only the project folders but client_state.xml and so on) immediately after receiving new work, later on you can start those tasks over at the beginning by stopping BOINC, replacing the data files and relaunching. (I wouldn’t recommend this if you’re running multiple projects.)

                                  Thank you for the hint. This may work, but is (too) hard to do all the time. I hoped to find an easier way.

                                  Anyhow, things are getting out of hand. Several of my WUs are issuing their 8th result and as there are many \"empty\" results, it is useless to crunch them, they will invariably exceed the 5/8/5 limit because of the \"too many results\" error, like this one. Others will fail because of the \"too many success results\" error, like this one. The computer with the 209732 sec (more than 58 hours) is mine. All the other results are \"empty\". This WU fill fail and a new WU will be created, starting from scratch.

                                  Under these conditions I think that it is useless for me to continue. So I decided to put my computers on \"no new work\". I will finish the 158 WUs that I still have (unless they have a canonical result or have failed before they start). I will resume my participation as soon as one the following conditions is met:

                                  • the project gets a working checkpoint system
                                  • the 5/8/5 limit is again raised to 10/20/10
                                  • the new application announced by Ádám on January 8th arrives (I don\'t know if the BinSys application will be finished in a few weeks or a few years... no information is available)
                                  • someone succeeds in convincing me to resume immediately.


                                  Until that day, bye!
                                  Robert


                                  [mod. off]
                                  This project needs a deep cleaning (application, db, server, feedback ...)
                                  Obviously the admins don\'t read the forum and i\'m not sure that they read the emails.
                                  So \"no new work\" since the beginning of the year
                                  [/mod. off]

                                  ____________

                                  Stick
                                  Send message
                                  Joined: Jun 12 06
                                  Posts: 193
                                  Credit: 66,244
                                  RAC: 0
                                  Message 5647 - Posted 4 Feb 2007 22:39:34 UTC - in response to Message 5645.

                                    Obviously the admins don\'t read the forum and I\'m not sure that they read the emails.


                                    Sadly, a very telling observation.

                                    The bugs don\'t bother me that much. (I actually enjoy the process of identifying problems and finding workarounds.) But, when our attempts at helping out are ignored, it is very frustrating. It says: \"The project just needs your computer\'s extra time; don\'t bother us with your observations or ideas.

                                    Therefore, I am joining this unofficial little \"boycott\".
                                    ____________

                                    Profile [B^S] Gamma^Ray
                                    Send message
                                    Joined: Jan 3 07
                                    Posts: 18
                                    Credit: 5,977
                                    RAC: 0
                                    Message 5648 - Posted 4 Feb 2007 23:30:48 UTC

                                      Last modified: 4 Feb 2007 23:43:37 UTC

                                      Ok, Here\'s my situation. I just finished running a 35+ Hour WU, I never paused the client, Never ran any other project against it, It had full control over my clent. After 35 hours of this, It ends with the dreaded Output is empty. Being that I was the 6th person to crunch this particular wu, I felt it was good to go forth with it when I saw it was one of those long ones, Well, Now that I\'ve just turned it in, I now see that the other five users who turned theirs in before me all were credited their credits asked for (Which all of those seemed to have ended with Output is empty also), But for whatever reason, I asked for 342.49 credits, And received 0 ! Here is the wu 58705 So what am I missing ? Did I just crunch 35 hours just so the other users can get their credits ? I now see the wu results shows \"errors Too many success results\" Anyway the adms at least can give me the credit before the db is purged as I think that is also an issue also ?

                                      G^R

                                      ____________

                                      Philip Martin Kryder
                                      Send message
                                      Joined: Apr 3 06
                                      Posts: 73
                                      Credit: 10,801
                                      RAC: 0
                                      Message 5650 - Posted 5 Feb 2007 4:24:22 UTC - in response to Message 5644.

                                        As far as I know, the 10/20/10 limit was temporary and was just supposed to make the WUs from mid-September come to a success. The real solution is of course a good checkpoint system, then we don\'t need the higher limit.


                                        Temporary or not - I\'ve never seen an explanation of why 5/8/5 is better.
                                        As I noted -
                                        If 3 results are \"successful\" and then validate then no more need to be sent.

                                        If on the other hand there are \"problems\" for any reason (client, CPU failure, bored participant or whatever, more work units can be sent until validation is successful...




                                        ____________

                                        Philip Martin Kryder
                                        Send message
                                        Joined: Apr 3 06
                                        Posts: 73
                                        Credit: 10,801
                                        RAC: 0
                                        Message 5651 - Posted 5 Feb 2007 4:28:26 UTC - in response to Message 5645.


                                          [mod. off]
                                          This project needs a deep clean up (application, db, server, feedback ...)
                                          Obviously the admins don\'t read the forum and i\'m not sure that they read the emails.
                                          So \"no new work\" since the beginning of the year
                                          [/mod. off]


                                          You mention \"admins\"

                                          Are there others besides \"Adam?\"

                                          I always keep wondering if
                                          rather than a math project,
                                          this is really a psychological experiment
                                          designed to see what the reaction of a volunteer community is to various stimuli.

                                          ____________

                                          robert.mouris
                                          Send message
                                          Joined: Nov 3 05
                                          Posts: 129
                                          Credit: 4,124,189
                                          RAC: 0
                                          Message 5653 - Posted 5 Feb 2007 9:24:50 UTC - in response to Message 5648.

                                            Last modified: 5 Feb 2007 9:26:49 UTC

                                            Did I just crunch 35 hours just so the other users can get their credits?

                                            You\'re not alone. This is the last result I sent in today. Even the computers which worked for less than 3 minutes got credit, but my 81 hour job got nothing. True, I reported too late, but I don\'t have the choice when, due to complete misestimation of processing time, the computer downloads more than 100 WUs and many of them need a few days\' crunching time. I\'m interested in scientific progress and not in credits, but my result is discarded and the whole WU must be reprocessed from scratch.

                                            ____________

                                            robert.mouris
                                            Send message
                                            Joined: Nov 3 05
                                            Posts: 129
                                            Credit: 4,124,189
                                            RAC: 0
                                            Message 5654 - Posted 5 Feb 2007 9:35:13 UTC - in response to Message 5651.

                                              Last modified: 5 Feb 2007 9:35:36 UTC

                                              Are there others besides \"Adam?\"

                                              A very long time ago, there was also an administrator called Attila. But now Ádám is alone. He said on 12 January that he will be preparing his examinations, which a sensible student is of course supposed to do.

                                              During the (European) nights, Nightbird acts as a forum moderator, but otherwise the project is unattended.

                                              ____________

                                              Profile [B^S] Gamma^Ray
                                              Send message
                                              Joined: Jan 3 07
                                              Posts: 18
                                              Credit: 5,977
                                              RAC: 0
                                              Message 5657 - Posted 5 Feb 2007 16:51:55 UTC

                                                Well, I guess everyone has their own priorities, But personally if his is now based on his own studies and not on this project, Then perhaps he should stop sending out more work units and resolve the now pending one\'s that are just sitting there ? As although it is his project to do what he wants with, I hope he would also understand that the wasted cpu cycle\'s that many of us seem to be throwing away here, Could have and would have been used for other projects where they would really be needed, Such as WCG, Rosetta, Malaria, Or to other projects of ones persoanal interests. Its not so much the fact that credits aren\'t awarded that bugs me, As much as it is the wasted cpu hours of many of us here that surely would have rathered spent them elsewhere. But, It is what it is I guess, And all of this is just my opinion anyways. :)

                                                Regards,
                                                G^R
                                                ____________

                                                miketoth1001
                                                Avatar
                                                Send message
                                                Joined: Apr 3 06
                                                Posts: 20
                                                Credit: 27,967
                                                RAC: 0
                                                Message 5688 - Posted 10 Feb 2007 12:19:10 UTC - in response to Message 4877.

                                                  3) Has anybody using MacOS or Linux received an empty result?


                                                  Now I can say that a Linux box CAN get an empty result. Came whe the UPS got turned off. ::sigh::

                                                  Profile [B^S] Gamma^Ray
                                                  Send message
                                                  Joined: Jan 3 07
                                                  Posts: 18
                                                  Credit: 5,977
                                                  RAC: 0
                                                  Message 5706 - Posted 12 Feb 2007 20:20:40 UTC - in response to Message 5648.

                                                    Last modified: 12 Feb 2007 20:23:25 UTC

                                                    Ok, Here\'s my situation. I just finished running a 35+ Hour WU, I never paused the client, Never ran any other project against it, It had full control over my clent. After 35 hours of this, It ends with the dreaded Output is empty. Being that I was the 6th person to crunch this particular wu, I felt it was good to go forth with it when I saw it was one of those long ones, Well, Now that I\'ve just turned it in, I now see that the other five users who turned theirs in before me all were credited their credits asked for (Which all of those seemed to have ended with Output is empty also), But for whatever reason, I asked for 342.49 credits, And received 0 ! Here is the wu 58705 So what am I missing ? Did I just crunch 35 hours just so the other users can get their credits ? I now see the wu results shows \"errors Too many success results\" Anyway the adms at least can give me the credit before the db is purged as I think that is also an issue also ?

                                                    G^R


                                                    Ok, It\'s now been one week, The wu is now \"Purged\", And no response from the Adm (And of course no credit for it either). What else does a user have to do to get reply from you ? I noticed Kornafeld Ádám, you will answer a new post that was posted today concerning closing an account, But after a week, your no response here indicates to me, 35 hours of waisted cpu time is acceptable ?
                                                    ____________

                                                    Profile [B^S] thierry@home
                                                    Avatar
                                                    Send message
                                                    Joined: Jul 12 05
                                                    Posts: 98
                                                    Credit: 255,646
                                                    RAC: 0
                                                    Message 5712 - Posted 14 Feb 2007 21:15:48 UTC

                                                      I tried to restart crunching three times for Sztaki in the last months.
                                                      Result: between 3000 & 4000 credits vanished, dissaperead, pfuit ....
                                                      Gone.

                                                      The worst to me is that I still don\'t know (maybe it\'s explained somewhere) what\'s happen.
                                                      I liked this project very mmuch and I was one of the first crunchers but I\'m very sad about what happens here.

                                                      miketoth1001
                                                      Avatar
                                                      Send message
                                                      Joined: Apr 3 06
                                                      Posts: 20
                                                      Credit: 27,967
                                                      RAC: 0
                                                      Message 5750 - Posted 23 Feb 2007 22:07:10 UTC

                                                        Well... I\'ve been running those Linux boxes here, and have uploaded many valid results. (Almost all of them were valid.) Problem is, is that there are still many results that were invalid, and even more piling on after I crunch them.

                                                        It doesn\'t bother me, as I wasn\'t going to use those boxes for BOINC, anyway. They\'re old < 1 GHz CPUs, and really don\'t help with other things. So, I\'ll let them run for a bit more.
                                                        ____________

                                                        miketoth1001
                                                        Avatar
                                                        Send message
                                                        Joined: Apr 3 06
                                                        Posts: 20
                                                        Credit: 27,967
                                                        RAC: 0
                                                        Message 5976 - Posted 24 Mar 2007 17:32:29 UTC

                                                          UPDATE: I gave up. I keep running clean results, and still get hardly any credit. 2+ months of running JUST Sztaki, and nothing to really show for it. At one point, I had over 4K of pending credits, and now I have around 800. Problem is, is that my total for the project went up only a couple hundred. The rest of those credits? POOF!!! Gone.

                                                          And, today I see that a machine I haven\'t had online for almost a full year somehow downloaded 97 WUs, and had client error on them all. It did this over the period of \"2\" days. (Before and after midnight UTC.) The machine isn\'t even plugged in. Just wild. Maybe a false CPID, but I deleted that machine LONG ago from the database.

                                                          Profile [B^S] Gamma^Ray
                                                          Send message
                                                          Joined: Jan 3 07
                                                          Posts: 18
                                                          Credit: 5,977
                                                          RAC: 0
                                                          Message 5981 - Posted 25 Mar 2007 20:00:47 UTC

                                                            Sorry to hear that Mike (The loss of time/cpu cycles). But it doesn\'t appear that things will be changing here anytime soon. :(


                                                            Regards
                                                            G^R
                                                            ____________

                                                            Post to thread

                                                            Message boards : SZTAKI Desktop Grid : error: Output is empty


                                                            Home | My Account | Message Boards


                                                            Copyright © 2017 SZTAKI Desktop Grid