Posts by MichaelB

1) Questions and Answers : Unix/Linux : Maybe interrupted upload - lost result? (Message 14150) Posted 6 Jul 2005 by MichaelB Post: That doesn't seem to be the answer: one more trickle, but no further update on my lost unit. Will the work have been lost? The final trickle from that unit was within an hour or so of the work-unit completing. (I'm not really interested in credits: I don't want to see six weeks of computing wasted :-( )
2) Questions and Answers : Unix/Linux : Maybe interrupted upload - lost result? (Message 14143) Posted 6 Jul 2005 by MichaelB Post: Short story: I\'m worried that I have inadvertantly interrupted the upload of the results of a work-unit. I believe I have the results archived safely, and want to know how to be sure that the CP project has actually received my results. Long story: Due to a muddled introduction to CPDN, I have had two units running one associated with \"www.climateprediction.net\" and another with \"climateprediction.net\". This seems messy, and I\'ve been eagerly waiting for the work-units to complete so that I could detach from one URL and continue with one work-unit at a time. I set my preferences to allow only 0.03GB disk usage to prevent another work-unit from being down-loaded. Last night, \"www.climateprediction.net\" (Result ID 843684) completed and appeared to finish uploading. Six hours or so later, \"climateprediction.net\" (Result ID 844519) completed. Results did not upload, and error output showed that contact was deferred for about 23 hours. I tarred up the directory for Result 843684 and moved it to a safe place, then ran \"boin... -detach_project www.climateprediction.net\" which completely removed that project directory. I still have the results safely archived, however. Impatient with the deferral, I stopped boinc, edited client_state.xml to reset the next contact time to now+60 seconds, and restarted boinc. The second work-unit appeared to upload successfully, and a new work-unit was down-loaded (and is now happily crunching away.) My account page, however, now lists only ResultID 844519 as having completed. ResultID 843684 is listed as still in progress, and I suspect I have interrupted a process that was not complete after all. How can I complete the delivery of the work-unit output to the project?