| Index | Recent Threads | Unanswered Threads | Who's Active | Guidelines | Search |
 |
World Community Grid Forums
Category: Support
Forum: Website Support
Thread: Hello, is there anyone here? |
| No member browsing this thread |
|
Thread Status: Active Total posts in this thread: 35
|
|
| Author |
|
|
BobbyB
Veteran Cruncher Canada Joined: Apr 25, 2020 Post Count: 638 Status: Offline Project Badges: 
|
I also use backup projects and the catch is that when WCG comes back with work you can get caught with other projects WUs stuck in your queues. Now I have to force them through.
I allowed one machine to pull Einstein, but it wanted to tie-up every CPU with huge WU's lasting 1-2 days apiece, so stayed with the others. I too had the same problem with MilkyWay (deleted it)... and Sidock WUs are all 1+day in length.Backups have a resource share of 0 so luckily they only download as many WUs as the number of cores you have. Now WCG kicks back in I still have 5 WUs each rated 1+days in my queue. I have to remember to force them through otherwise they will expire and thus waste all that computing done so far. No new tasks for the backup project then repeat suspend and resume WCG until they are gone. I raised "switch between tasks" to 90 minutes to lower the babysitting down a bit. Has anyone found a better method? |
||
|
|
Sgt.Joe
Ace Cruncher USA Joined: Jul 4, 2006 Post Count: 7846 Status: Offline Project Badges: 
|
Before enabling your backup project (when necessary) lower your cache settings to maybe 1 or 2 days. Once WCG kicks back in, your cache of existing units from your backup project should finish in plenty of time. Also, once WCg is again back in business, shut off the supply from your backup project.
----------------------------------------Cheers
Sgt. Joe
*Minnesota Crunchers* |
||
|
|
BobbyB
Veteran Cruncher Canada Joined: Apr 25, 2020 Post Count: 638 Status: Offline Project Badges:
|
Hmmm,... I don't even have those high cache limits.
The only machine with 1 day/.5 day is my 3900x. The rest have .5/.5 or lower. I always leave Denis and Tn-Grid on because their WUs are like WCG ~1-2 hour max and they will only kick in if WCG goes dry. The problem ones are the long WUs greater than the "switch between tasks" time left. Still have 3 left which will terminate in 3 more cycles of suspend/resume. So are you saying I should up my cache for the 3900x? Looked at FOLDing but it requires another client. The rest of the Boinc projects are not health or medical science related. I'm not really interested in prime numbers or the like and really do not want to look for ET. Yes I know they stopped. |
||
|
|
Sgt.Joe
Ace Cruncher USA Joined: Jul 4, 2006 Post Count: 7846 Status: Offline Project Badges:
|
Hmmm,... I don't even have those high cache limits. The only machine with 1 day/.5 day is my 3900x. The rest have .5/.5 or lower. I always leave Denis and Tn-Grid on because their WUs are like WCG ~1-2 hour max and they will only kick in if WCG goes dry. The problem ones are the long WUs greater than the "switch between tasks" time left. Still have 3 left which will terminate in 3 more cycles of suspend/resume. So are you saying I should up my cache for the 3900x? Looked at FOLDing but it requires another client. The rest of the Boinc projects are not health or medical science related. I'm not really interested in prime numbers or the like and really do not want to look for ET. Yes I know they stopped. If you are running caches that low, you should never get enough work units you are not going to complete in time. When I get both WCG and Denis units at the same time, I just let BOINC decide which order to run them in so they all get done. My setting for switch between tasks is currently 120 minutes and I run a 2 day cache. BOINC so far has not let me down with any tasks which were overdue. If WCG is running steady, I leave DENIS turned off, but if it looks like WCG is going to dry up, I enable DENIS always realizing the work units from DENIS are sporadic at best. So far this works for both machines I leave on 24/7 and those which I shut down at night or when I am away. Cheers
Sgt. Joe
*Minnesota Crunchers* |
||
|
|
BobbyB
Veteran Cruncher Canada Joined: Apr 25, 2020 Post Count: 638 Status: Offline Project Badges:
|
OK so 120 minutes it is for all machines. I know when longer work than that comes in since I manually let them in.
|
||
|
|
Unixchick
Veteran Cruncher Joined: Apr 16, 2020 Post Count: 1298 Status: Offline Project Badges: 
|
I always lower my available cpus (50%) before going to backup projects (resource share 0%), so that when I see that WCG is back, I can turn the backup project to No new tasks and flip the cpus back to 100%
|
||
|
|
Bryn Mawr
Senior Cruncher Joined: Dec 26, 2018 Post Count: 385 Status: Offline Project Badges: 
|
@BobbyB
I have to remember to force them through otherwise they will expire and thus waste all that computing done so far. That should never happen. As soon as one of the backup tasks approaches the point where it is in danger of missing its deadline Boinc will set it to high priority and force it through ignoring any RAC weighting you might have set. |
||
|
|
TPCBF
Master Cruncher USA Joined: Jan 2, 2011 Post Count: 2173 Status: Offline Project Badges:
|
Well, the "Overview" page, more precisely those obnoxious progress bars have started to act up, looks like we might be in for some more system downtime for this very fast approaching weekend.
I already noticed that the overall amount of WUs kept in the Results database has been increasing considerably during the week, quite possible that some database for the web site is running out of space yet again...  Ralf |
||
|
|
BobbyB
Veteran Cruncher Canada Joined: Apr 25, 2020 Post Count: 638 Status: Offline Project Badges:
|
@Bryn Mawr
As soon as one of the backup tasks approaches the point where it is in danger of missing its deadline Boinc will set it to high priority and force it through ignoring any RAC weighting you might have set. Ah! did not know that. Next time I will put it to the test with at least 1 WU.There will be a next time. |
||
|
|
nivrip
Senior Cruncher North Yorkshire Joined: Sep 13, 2007 Post Count: 285 Status: Offline Project Badges: 
|
I already noticed that the overall amount of WUs kept in the Results database has been increasing considerably during the week, quite possible that some database for the web site is running out of space yet again. Yup, me too. Four times what I'm used to.
ЮРКШИР КРУНЧЕР
|
||
|
|
|