| Index | Recent Threads | Unanswered Threads | Who's Active | Guidelines | Search |
 |
World Community Grid Forums
Category: Completed Research
Forum: Help Cure Muscular Dystrophy - Phase 2 Forum
Thread: won't get work from my requested projects because of monster WUs |
| No member browsing this thread |
|
Thread Status: Active Total posts in this thread: 8
|
|
| Author |
|
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
Whenever my client requests new work, I get the following:
Sun 07 Jun 2009 07:25:01 PM EDT|World Community Grid|Message from server: No work sent Sun 07 Jun 2009 07:25:01 PM EDT|World Community Grid|Message from server: No work available for the applications you have selected. Please check your settings on the website. Sun 07 Jun 2009 07:25:01 PM EDT|World Community Grid|Message from server: (won't finish in time) Computer on 100.0% of time, BOINC on 100.0% of that I had aborted a CMD2 job at 77 hours because it was only 25% done, another did finish at 71 hours but was invalid and I have another waiting after 59 hours (which I suspect will be denied), now all my jobs have a To Completion time above 100 hours and I get a message from BOINC that won't return my requested projects probably because of these erroneous jobs. Will my machine "fix" itself with reasonable completion times and will I get jobs from projects I want to work on? How do I get back to normal? Is this a linux thing because my windows boxes are fine (getting jobs from the same requested projects that I can't get to with my linux box). Why did these jobs run so long (just to get rejected)? I'm on Linux (Ubuntu 8.10) on intel (quad). Any help or pointers are welcome. |
||
|
|
Van Fanel
Cruncher Joined: Dec 27, 2006 Post Count: 42 Status: Offline Project Badges: 
|
Have you tried increasing the Work buffer? You can find it under the menu Advanced --> Preferences --> Network Usage tab.
|
||
|
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
Yes, I have tried it. Bumped it to 5 a few days ago.
|
||
|
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
Hi ripleys,
Sure, Linux will fix itself. It just needs to run some standard jobs in standard times. Just let it crunch and don't bother with a big cache. A long queue is the easiest way to get into deadline pressure if the estimated work lengths go wrong. Lawrence |
||
|
|
Sekerob
Ace Cruncher Joined: Jul 24, 2005 Post Count: 20043 Status: Offline |
rDCF mayhem. Yes there is an FAQ for that. The moment the BOINC developers introduce computing this value at science level and not project level (WCG is just 1 project to them), we'll be seeing massive improvement in buffer sizing and TTC projection.
----------------------------------------As I understood it, 6.6.31 [I'm testing] has a 'learn about a new science' feature to help operation planning, but right now, my 24/7 only WCG quad shows a rDCF of 4.25, 25 hours for FLU and 9:21 hours for HCMD2 (WCG project daily mean is 2.4 hours), so don't hope for any improvement from the latest 'recommended' at Berkeley in that department.
WCG
Please help to make the Forums an enjoyable experience for All! |
||
|
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
My rDCF is 62.424475 and slowly edging downwards. I so badly want to change the client_state.xml file manually but I won't as I saw your warning in another post.
|
||
|
|
Sekerob
Ace Cruncher Joined: Jul 24, 2005 Post Count: 20043 Status: Offline |
Set work fetch in the project tab No New Tasks (button changes to Allow new tasks), finish the jobs you have and then do a project reset. That will set the rDCF specific to WCG back to 1.000000, or just let it run. BOINC should be fetching new work just shortly before the current work finishes till things normalize.
----------------------------------------
WCG
Please help to make the Forums an enjoyable experience for All! |
||
|
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
My rDCF is 62.424475 and slowly edging downwards. I so badly want to change the client_state.xml file manually but I won't as I saw your warning in another post. Terminate boinc, change client_state.xml manually, restart boinc. It's quick and easy. If you know the appropriate sensible value for the machine (e.g. from a backup or from having noticed in the past), all the better. |
||
|
|
|