Questions and Answers :
Windows :
Recovery of corrupted cp files after a bad crash?
Message board moderation
Author | Message |
---|---|
Send message Joined: 10 Mar 05 Posts: 3 Credit: 4,962 RAC: 0 |
I\'m running a cp model under BOINC 4.19, on a laptop with a Celeron 2GHz CPU and WinXP SP2. I also have SETI@home and Einstein@home running. The laptop seems to have a heat-triggered breaker that shuts down the system instantly when it reaches a certain temperature--a definite problem given the summer heat here. The most recent accident of the sort caused some BOINC/cp files to be corrupted; I\'m still hunting for the scandisk log to be able to tell which ones, but I know the computer has recovered four file fragments. The effect is that BOINC no longer knows where it is in the model. It seems to have restarted from scratch, only now the time to completion is almost 19 million hours (no kidding). SETI and Einstein are fine and have lost nothing. What I\'d like to know is can I somehow salvage the data from the recovered files? Or should I reset the project? This would be a major bummer, as I\'d completed the first phase of the project, a matter of over 300 hours of CPU time... Even more frustrating is that I backed up my files three days ago, but forgot to include the data in the Program Files. Is there a way to relocate the BOINC data into My Documents so that it gets included next time? |
Send message Joined: 7 Aug 04 Posts: 2184 Credit: 64,822,615 RAC: 5,275 |
Re: cooling...See <a href="http://www.climateprediction.net/board/viewtopic.php?t=1954&start=0">this thread</a> for suggestions on keeping the laptop cool, if you haven't already tried them. Re: can you save this model?...likely not if something has been corrupted and it started back at the beginning. If it started at the beginning of the same model, which it sounds like, you won't get any credits until you get past the point you were before. If that doesn't bother you, I would continue on with that model, recognizing that the estimated time to completion will be way off. If you are concerned about the credits, a project reset will get you going again on a new model. Re: backing up...you should be able to copy the BOINC folder out of Program Files, to a location that will be backed up, uninstall BOINC, then reinstall to the new directory you copied BOINC to. This should work for continuing on with BOINC. However, it gets a little tricky since you run several projects, i.e. if CPDN has a problem and you want to recover from a backup, you can't just copy over the whole folder and start up again...or at least there will also be redundant work done on the other projects as well. There may be a way just to copy back the CPDN associated files, but since some of the info resides in the xml files in the BOINC directory, it can be problematic. Perhaps someone else can come along and tell you what you might need to do in a multiproject environment. |
Send message Joined: 10 Mar 05 Posts: 3 Credit: 4,962 RAC: 0 |
> Re: cooling...See <a href="http://www.climateprediction.net/board/viewtopic.php?t=1954&start=0">this thread</a> Thanks for the reference, I hadn't seen that thread. I've been propping up my laptop on books and the like to allow air circulation, but this sounds like a better solution. :) > Re: can you save this model?...likely not I was afraid of that. > Re: backing up...you should be able to copy the BOINC folder out of Program > Files, to a location that will be backed up, uninstall BOINC, then reinstall > to the new directory you copied BOINC to. Again, thank you. I assume I can take that opportunity, if I'm going to be uninstalling BOINC anyway, to upgrade to BOINC 4.25? |
Send message Joined: 7 Aug 04 Posts: 2184 Credit: 64,822,615 RAC: 5,275 |
> Again, thank you. I assume I can take that opportunity, if I'm going to be > uninstalling BOINC anyway, to upgrade to BOINC 4.25? I think in that instance, you would uninstall boinc, copy the directory with the remaining files to the new location, then install 4.25 into that directory. It should work. |
©2024 cpdn.org