climateprediction.net home page
Can I restart incomplete WU\'s?

Can I restart incomplete WU\'s?

Questions and Answers : Windows : Can I restart incomplete WU\'s?
Message board moderation

To post messages, you must log in.

AuthorMessage
old_user99076

Send message
Joined: 20 Sep 05
Posts: 5
Credit: 14,499
RAC: 0
Message 17163 - Posted: 13 Nov 2005, 8:51:40 UTC

I have 3 workunits that are incomplete, one was running phase 2 the other two are barely into phase one before they stopped. For some reason these work units did not restart after my PC was rebooted and a new work unit was requested. Is there any way of resuming the work, or is it gone forever? It seems like a waste of computing resources if they can\'t be resumed
ID: 17163 · Report as offensive     Reply Quote
Les Bayliss
Volunteer moderator

Send message
Joined: 5 Sep 04
Posts: 7629
Credit: 24,240,330
RAC: 0
Message 17164 - Posted: 13 Nov 2005, 9:00:10 UTC

The only way to restart a crashed model is from a backup made before the crash.

ID: 17164 · Report as offensive     Reply Quote
old_user99076

Send message
Joined: 20 Sep 05
Posts: 5
Credit: 14,499
RAC: 0
Message 17165 - Posted: 13 Nov 2005, 9:26:55 UTC - in response to Message 17164.  

The only way to restart a crashed model is from a backup made before the crash.



Thanks Les, are you referring to a backup I have mase using something like BackupExec or the Windows backup? Or are you talking about a BOINC backup? If you are talking about a BOINC backup how do you restore? Thanks
ID: 17165 · Report as offensive     Reply Quote
Les Bayliss
Volunteer moderator

Send message
Joined: 5 Sep 04
Posts: 7629
Credit: 24,240,330
RAC: 0
Message 17166 - Posted: 13 Nov 2005, 9:40:19 UTC

A copy of the entire BOINC folder and sub folders, made after suspending BOINC. This is necessary because of the many files involved, (and the time it takes to make the backup), some of which may get written to after others have been saved, thus getting things out of sync.
There is also a lock file which otherwise won\'t get copied.
I\'m not sure what will happen if BOINC wasn\'t suspended.

To restore, delete the original BOINC folder and sub folders, and copy the backup into it\'s place.
For people who run multiple projects, this can get messy. If you\'re in that situation, say so, and I\'ll post a how-to-do-it.

ID: 17166 · Report as offensive     Reply Quote
Profile old_user44382

Send message
Joined: 28 Jan 05
Posts: 1
Credit: 11,314
RAC: 0
Message 17176 - Posted: 14 Nov 2005, 14:30:02 UTC - in response to Message 17164.  
Last modified: 14 Nov 2005, 14:33:32 UTC

The only way to restart a crashed model is from a backup made before the crash.


I make a Winzip archive of the complete BOINC folder every day or so. When the model crashes, which it does frequently on my machine, I delete the old BOINC folder and unzip the most recent archive in its place. It is best to keep more than one backup in case the backup should fail, although the archive program usually warns if there is an error. With this in mind, it is important to suspend and shut down the model before archiving it, otherwise it will not backup properly.
ID: 17176 · Report as offensive     Reply Quote
staffann

Send message
Joined: 23 Oct 05
Posts: 22
Credit: 526,746
RAC: 0
Message 17184 - Posted: 14 Nov 2005, 21:12:40 UTC - in response to Message 17166.  

For people who run multiple projects, this can get messy. If you\'re in that situation, say so, and I\'ll post a how-to-do-it.


It would be very interesting to know if I can restore cp without messing up work done in other projects. I had to restore a backup today, fortunatly with no major consequence for other projects. It would be good to know how to do it next time.

ID: 17184 · Report as offensive     Reply Quote
Les Bayliss
Volunteer moderator

Send message
Joined: 5 Sep 04
Posts: 7629
Credit: 24,240,330
RAC: 0
Message 17185 - Posted: 14 Nov 2005, 22:06:02 UTC

I originally posted a possible solution <a href=\"http://climateapps2.oucs.ox.ac.uk/cpdnboinc/forum_thread.php?id=3268\"> here.</a>

The relevant part was:
Perhaps the simple way, (after a crash), is to suspend cpdn, finish the other projects, suspend them, get back the saved BOINC, suspend all the other projects, (which have been finished), and restart cpdn.
This will probably require BOINC 4.45, which has, I think, an option to stop the download of more wus from a project.

Since then, the 5.* series of BOINC have been released, with different options for suspending, etc.

As you can see, the person in question tried it, and documented the proceedure <a href=\"http://www.adrianxw.dk/personalsite/boinc/cpdn_bu.html\"> here.</a>
Backup has now also been added to the BOINC Wiki, <a href=\"http://boinc-doc.net/boinc-wiki/index.php?title=Backup_BOINC\"> here.</a>

ID: 17185 · Report as offensive     Reply Quote
crandles
Volunteer moderator

Send message
Joined: 16 Oct 04
Posts: 692
Credit: 277,679
RAC: 0
Message 17186 - Posted: 14 Nov 2005, 22:23:04 UTC
Last modified: 14 Nov 2005, 22:23:55 UTC

<a href=\"http://boinc-doc.net/boinc-wiki/index.php?title=Backup_BOINC\">WIKI Backup_BOINC</a> has some suggestions. Further discussion on <a href=\"http://climateapps2.oucs.ox.ac.uk/cpdnboinc/forum_thread.php?id=3268\">this thread</a>

Edit: Oops Les beat me to it.
Visit BOINC WIKI for help

And join BOINC Synergy for all the news in one place.
ID: 17186 · Report as offensive     Reply Quote

Questions and Answers : Windows : Can I restart incomplete WU\'s?

©2024 climateprediction.net