| Index | Recent Threads | Unanswered Threads | Who's Active | Guidelines | Search |
 |
World Community Grid Forums
Category: Community
Forum: Chat Room
Thread: Project Status (First Post Updated) |
| Member(s) browsing this thread: Link64 |
|
Thread Status: Active Total posts in this thread: 661
|
|
| Author |
|
|
alanb1951
Veteran Cruncher Joined: Jan 20, 2006 Post Count: 1341 Status: Offline Project Badges: 
|
Good to see they're going to set an early warning system on the WAS thread pool management! Touchy software, isn't it
 And now we know why getting work signed off seemed to be going backwards. If they could find a way of snuffing out that that bad parsing of scheduler requests (presumably a BOINC issue rather than a WCG one) at source, the "fake" Linux jobs (and associated non-HR validation issues) would eventually stop being a problem! And if they can solve that (or BOINC fix whatever it is in the (Docker-aware?) client and/or the server code, validation would be simplified again -- however, no mention of that in the update... Recently I've been working at ways of digging out examples of stuck "2 valid, 1 pending" cases and was about to post some examples when this update appeared -- no need now, as that case appears on their list of issues and they almost certainly don't (and didn't) need examples from the end users anyway... Hopefully, this week will see some of the issues sorted -- I don't expect them to get the lot in one go, but sorting out one part will release some of the pressure elsewhere and might make it easier to track down and resolve those awkward validation edge cases! Cheers - Al. |
||
|
|
Unixchick
Veteran Cruncher Joined: Apr 16, 2020 Post Count: 1316 Status: Offline Project Badges: 
|
Thank you Al for spotting the update and posting it to the thread. I really like how they are letting us know the issues and what they are working on.
|
||
|
|
Grumpy Swede
Master Cruncher Svíþjóð Joined: Apr 10, 2020 Post Count: 2557 Status: Recently Active Project Badges: 
|
No wonder that it could be double(ish) credits for some, when the stats looks like this. The global history stats, can't possible be higher than the only project (MCM1) that we are running now. ARP hasn't been run since before the migration, and there's no new stats history for any other project than MCM1. And with the extremely low validation rate we have now, I doubt that any of the "Results returned" numbers below, is anywhere near the true values. They look much higher than the low validation rate suggests.
----------------------------------------Statistics history (Mapping Cancer Markers History) Statistics date Total run time Points generated Results returned [Edit 5 times, last edit by Grumpy Swede at Dec 16, 2025 10:51:49 AM] |
||
|
|
adriverhoef
Master Cruncher The Netherlands Joined: Apr 3, 2009 Post Count: 2361 Status: Offline Project Badges: 
|
Latest Operational Status update... December 15, 2025
They finally get it and are now posting in UTC, bravo!  Thank you, thank you, thank you.Adri [Edit 1 times, last edit by adriverhoef at Dec 16, 2025 11:20:47 AM] |
||
|
|
alanb1951
Veteran Cruncher Joined: Jan 20, 2006 Post Count: 1341 Status: Offline Project Badges:
|
Regarding the discrepancies between Global daily stats and MCM1 daily stats...
If you look at the last 30 days of both and consider the post-migration days only, you'll find a few more cases where Global > MCM1 and several where MCM1 = 2*Global! I suspect that replaying of validation datasets cause these confusions, and I'm not sure how reliable either data set is likely to be now, especially given that the Global>MCM1 examples all seem to solve out as 2*Global = 3*MCM, suggesting that the correct value in all cases could well be the difference between them! Ah, well, it'll either sort out eventually or there will be some people with lots of extra [unearned] credit in perpetuity Cheers - Al. |
||
|
|
Grumpy Swede
Master Cruncher Svíþjóð Joined: Apr 10, 2020 Post Count: 2557 Status: Recently Active Project Badges:
|
@alanb1951
Yeah, I have also seen the same discrepancy before. One thing is sure though, for now one can not rely on either the MCM1 stats, or the Global stats. We'll se what happens in the future, when the validations start working reliable. |
||
|
|
Unixchick
Veteran Cruncher Joined: Apr 16, 2020 Post Count: 1316 Status: Offline Project Badges:
|
This is an interesting match up. Mine is the error. I'm assuming this is a download error. https://www.worldcommunitygrid.org/contribution/workunit/792354903
|
||
|
|
TLD
Veteran Cruncher USA Joined: Jul 22, 2005 Post Count: 863 Status: Offline Project Badges: 
|
This is an interesting match up. Mine is the error. I'm assuming this is a download error. https://www.worldcommunitygrid.org/contribution/workunit/792354903 Does your computer have a intel CPU or Apple?  |
||
|
|
Unixchick
Veteran Cruncher Joined: Apr 16, 2020 Post Count: 1316 Status: Offline Project Badges:
|
Apple M4
----------------------------------------[Edit 1 times, last edit by Unixchick at Dec 16, 2025 8:34:20 PM] |
||
|
|
alanb1951
Veteran Cruncher Joined: Jan 20, 2006 Post Count: 1341 Status: Offline Project Badges:
|
Unixchick's "interesting match up" job did indeed get the the dreaded Permanent HTTP download error
---------------------------------------- -- at present, retries can go almost anywhere because of the Homogeneous Redundancy changes in place to avoid the "other platforms" issue.I, for one, will be glad when they can turn HR back on again, if only because users seem to end up with the much slower 32-bit retries at present. And yes, that seems to happen to Windows users as well as Linux users if the log files for successful retries that went to so-called "Alpine Linux" clients are anything to go by... Perhaps some users who have deliberately disabled alt_platform in their client could say whether they actually get any retries at present and, if so, has it respected the "64-bit only" constraint? There are certainly still enough retries available, with download errors and [too many] missed deadline tasks... Cheers - Al. P.S. That download error was yet another "it happens around 16:00 to 18:00 UTC" case. I wonder what's happening early afternoon local time over there... [Edit 1 times, last edit by alanb1951 at Dec 17, 2025 1:30:21 AM] |
||
|
|
|