Message boards : Number crunching : "Stuck" WU?
Message board moderation

To post messages, you must log in.

AuthorMessage
Ensor
Avatar

Send message
Joined: 13 Oct 07
Posts: 10
Credit: 150,571
RAC: 0
Message 19623 - Posted: 15 May 2008, 18:02:14 UTC

Anyone else having problems with WU #2642346?

On my host at least, progress has reached 100% after just over 6hrs of processing, but it's status is showing as "Waiting to run".

Looking at the rewsults page for the WU I notice that noone else who's been issued this WU has yet returned a result yet either....

I'm getting close to just aborting it.


TTFN - Pete.

ID: 19623 · Report as offensive     Reply Quote
Invisible Man

Send message
Joined: 23 May 07
Posts: 18
Credit: 19,129
RAC: 0
Message 19625 - Posted: 15 May 2008, 22:55:51 UTC - in response to Message 19623.  
Last modified: 15 May 2008, 22:57:29 UTC

I'm getting close to just aborting it.


TTFN - Pete.


TTFN? Wow, showing your age now old son. Even I can remember the ITMA radio programme !!
ID: 19625 · Report as offensive     Reply Quote
Profile Paul D. Buck

Send message
Joined: 2 Sep 04
Posts: 545
Credit: 148,912
RAC: 0
Message 19626 - Posted: 16 May 2008, 6:09:56 UTC - in response to Message 19623.  

Anyone else having problems with WU #2642346?

On my host at least, progress has reached 100% after just over 6hrs of processing, but it's status is showing as "Waiting to run".

Looking at the rewsults page for the WU I notice that noone else who's been issued this WU has yet returned a result yet either....

I'm getting close to just aborting it.

BOINC has this annoying habit of halting processing, at times, just before it completes a task.

Because of the way that work is scheduled the task was run and JUST before it was able to complete and upload, it gets suspended. I have had tasks that have run for long times, with just a minute or so to complete, get suspended and thus cluttering up my work queue ... which is not as big an issue to me as the fact that the task is vulnerable to a computer crash ...

Anyway, my suggestion (years ago) that the work scheduler take a look at time to complete before halting work on a task was never allowed to come to fruition. Sadly. So, in a case like this, all you can do is wait until LHC is scheduled again and the task completes ... or to suspend other projects and/or tasks to force this to run. The bad news is that if you are not careful you can "unbalance" the CPU scheduler so that it DL more work for other projects.



ID: 19626 · Report as offensive     Reply Quote
Profile Ocean Archer
Avatar

Send message
Joined: 13 Jul 05
Posts: 143
Credit: 263,300
RAC: 0
Message 19628 - Posted: 16 May 2008, 17:33:20 UTC

As per usual, Paul has hit the nail directly on the head -- again.

Before I start, I wish to make it perfectly clear, that I am not a programmer. That said, it would seem that a change to the program so it would test for two or three specific points --- completion percentage of WU in question; the processing duration for the projects (how many minutes before it switches from one project to another); the extimated run time left on the WU in question ...

What I'm trying to say, is if the darn project is within a time cycle of finishing, let it finish, then switch to the next job in the line ...


If I've lived this long, I've gotta be that old
ID: 19628 · Report as offensive     Reply Quote
Profile ChertseyAl
Avatar

Send message
Joined: 28 Nov 05
Posts: 31
Credit: 115,957
RAC: 0
Message 19629 - Posted: 16 May 2008, 18:39:50 UTC - in response to Message 19628.  

Disclaimer: <i>This is rather OT for this thread, but WTH ...</i>

What I'm trying to say, is if the darn project is within a time cycle of finishing, let it finish, then switch to the next job in the line ...


This is something that I'd really like to see implemented - It's amazing how many times I find a WU with a few seconds left to run switched out and left to wait for several hours before being able to return a result.

But, some projects lie about their completion time, and often run on beyond 100%. Maybe some unscrupulous project could run it's WUs up to 99.9% initially and then hold the host for ransom until completed.

Also, frankly, most/all of the 'wrappered' projects don't work well with BOINC. Checkpointing seems to be a joke in some cases. But that's another rant for another forum :)

Al.

ID: 19629 · Report as offensive     Reply Quote
Ensor
Avatar

Send message
Joined: 13 Oct 07
Posts: 10
Credit: 150,571
RAC: 0
Message 19630 - Posted: 16 May 2008, 19:25:55 UTC - in response to Message 19625.  


TTFN? Wow, showing your age now old son. Even I can remember the ITMA radio programme !!

Hmm, actually I got that from the first computer system I used at school about 1978....the operating system (MAXIMOP) printed that when you logged out.

Ah, the hours wasted playing "Star Trek" on a TTY @ 110baud.... :-)


TTFN - Pete.

ID: 19630 · Report as offensive     Reply Quote
Ensor
Avatar

Send message
Joined: 13 Oct 07
Posts: 10
Credit: 150,571
RAC: 0
Message 19631 - Posted: 16 May 2008, 19:35:01 UTC - in response to Message 19626.  
Last modified: 16 May 2008, 20:05:28 UTC


Hi,

BOINC has this annoying habit of halting processing, at times, just before it completes a task....

Unfortunately, in this case that's not what's happened....BOINC is showing the WU as 100% complete, but it's status as "Waiting to run".

In fact, it's downloaded and completed several other LHC WU's in the meantime, but this one just refuses to budge. :-(


....I have had tasks that have run for long times, with just a minute or so to complete, get suspended and thus cluttering up my work queue....

I find that damn annoying too, which is why I've set the time between switching tasks to 4 hours on my host (which suits the projects I run). It tends to avoid that problem.

On the odd occasion I do spot such a WU, I suspend all others and allow it to finish.


Anyway, my suggestion (years ago) that the work scheduler take a look at time to complete before halting work on a task was never allowed to come to fruition....

Glad I'm not the only one with that opinion, pity the BOINC devs can't take it on board. It's such an obvious improvement they could make....


TTFN - Pete.

ID: 19631 · Report as offensive     Reply Quote
Ensor
Avatar

Send message
Joined: 13 Oct 07
Posts: 10
Credit: 150,571
RAC: 0
Message 19632 - Posted: 16 May 2008, 20:10:34 UTC


Oh, incidentally, I've aborted this "stuck" WU....nice waste of 6hrs processing time, sigh.... :-(


TTFN - Pete.

ID: 19632 · Report as offensive     Reply Quote
Invisible Man

Send message
Joined: 23 May 07
Posts: 18
Credit: 19,129
RAC: 0
Message 19635 - Posted: 17 May 2008, 22:00:12 UTC - in response to Message 19630.  


TTFN? Wow, showing your age now old son. Even I can remember the ITMA radio programme !!

Hmm, actually I got that from the first computer system I used at school about 1978....the operating system (MAXIMOP) printed that when you logged out.

Ah, the hours wasted playing "Star Trek" on a TTY @ 110baud.... :-)


TTFN - Pete.


Sorry Pete, you're about 30 years too young. If you are interested, Google UK sites for TTFN and ITMA... (for Mrs Mopp & Tommy Handley).

TTFN - Viv.

ID: 19635 · Report as offensive     Reply Quote

Message boards : Number crunching : "Stuck" WU?


©2024 CERN