climateprediction.net (CPDN) home page
Thread 'Model Totally Wacked'

Thread 'Model Totally Wacked'

Message boards : Number crunching : Model Totally Wacked
Message board moderation

To post messages, you must log in.

AuthorMessage
Profileold_user87778

Send message
Joined: 14 Jul 05
Posts: 8
Credit: 27,173
RAC: 0
Message 16075 - Posted: 16 Sep 2005, 21:19:49 UTC

My Slab model (20y7_000115506_0) went fromover 169000 time steps to just over 1000 and back to Phase #1. I was already up to July 2060 in Phase #3 when this happened. We lost power (As has happened before without problems)and when the model came back up, it showed only 1% complete at Phase #1.

Should I dump (Detatch/Reattatch) and get a new experiment?
ID: 16075 · Report as offensive     Reply Quote
ProfileAndrew Hingston
Volunteer moderator

Send message
Joined: 17 Aug 04
Posts: 753
Credit: 9,804,700
RAC: 0
Message 16076 - Posted: 16 Sep 2005, 21:47:09 UTC - in response to Message 16075.  
Last modified: 16 Sep 2005, 21:47:48 UTC

Should I dump (Detatch/Reattatch) and get a new experiment?


It\'s bad luck - there have been a few reports of this happening, though normally it should only rewind up to a year. It is always wise to take a periodic backup of BOINC because there is always a risk that a computer crash will screw up the data files, which is probably what happened here.

In theory, you should be OK now, but unfortunately you won\'t get any credits for the reworking, until you get past the highest reported trickle point. That may not matter a jot, but if it does there is no real harm in dumping the WU at this stage. If you are using BOINC Manager (BOINC versions above 4.19) there is no need to detach, though - you can abort a WU in the work tab. You can even suspend it and get a new one if you haven\'t quite made up your mind whether to throw it away :-) It would be quite a good moment to upgrade BOINC if you are not using v4.45, before you start a new WU.

ID: 16076 · Report as offensive     Reply Quote
Profileold_user87778

Send message
Joined: 14 Jul 05
Posts: 8
Credit: 27,173
RAC: 0
Message 16077 - Posted: 16 Sep 2005, 22:00:33 UTC

Thank You.

I wonder if Honza would agree. I am using BOINC 4.45 which is what supised me about this. I hate to loose Phase three but I don\'t know much about how data files are stored. If they were stored in SHTML files then one might find what got corrupted/reset and fix it.

Ok I guess I\'ll abort it after all I did get 2/3 of the phase #3 trickles in.

Again Thanks
ID: 16077 · Report as offensive     Reply Quote
ProfileAndrew Hingston
Volunteer moderator

Send message
Joined: 17 Aug 04
Posts: 753
Credit: 9,804,700
RAC: 0
Message 16079 - Posted: 16 Sep 2005, 23:35:05 UTC - in response to Message 16077.  

Thank You.

I wonder if Honza would agree.


You could ask him. Suspending the WU will automatically download a new WU but always give you the option of going back to it later if you change your mind.

I don\'t know much about how data files are stored. If they were stored in SHTML files then one might find what got corrupted/reset and fix it.


It would be nice, but I don\'t think anyone has ever done it without a backup.

ID: 16079 · Report as offensive     Reply Quote

Message boards : Number crunching : Model Totally Wacked

©2024 cpdn.org