| Index | Recent Threads | Unanswered Threads | Who's Active | Guidelines | Search |
 |
World Community Grid Forums
Category: Official Messages
Forum: News
Thread: Comprehensive Issue List & Report Thread (Feb. 24, 2023) |
| No member browsing this thread |
|
Thread Status: Active Thread Type: Sticky Thread Total posts in this thread: 427
|
|
| Author |
|
|
Dr_Mabuse
Cruncher Joined: Jan 23, 2006 Post Count: 21 Status: Offline Project Badges: 
|
Hello, if I try to downloaf work units i get the answer:
15.03.2023 03:17:25 | World Community Grid | Server error: feeder not running This was all the time during the server problems. Are you going to fix this or is the problem on my side ? thanks for help Jocen from Germany |
||
|
|
bfmorse
Senior Cruncher US Joined: Jul 26, 2009 Post Count: 303 Status: Offline Project Badges: 
|
Hello, if I try to downloaf work units i get the answer: 15.03.2023 03:17:25 | World Community Grid | Server error: feeder not running This was all the time during the server problems. Are you going to fix this or is the problem on my side ? thanks for help Jocen from Germany Issue is covered at this link: https://www.worldcommunitygrid.org/about_us/article.s?articleId=780 |
||
|
|
Warped@RSA
Senior Cruncher South Africa Joined: Jan 15, 2006 Post Count: 423 Status: Offline Project Badges: 
|
Thanks. Just wondering what the potential timeline on "soon" is ???? Soon is a safe word to use Cheers  How Much Time Is Soon   Ralf  Soon is relative. Relative to the age of the earth (about 4.5 billion years), a few hundred years would be considered "soon".
Dave
 |
||
|
|
Tomahawk4196
Advanced Cruncher USA Joined: Aug 16, 2007 Post Count: 93 Status: Offline Project Badges: 
|
All:
----------------------------------------Please tell everyone to review this link in the forums: https://www.worldcommunitygrid.org/forums/wcg/viewpostinthread?post=682966 WCG needs an infusion of greenery, for St. Patrick's day, to help stop these constant problems. Please donate whatever you can, $5 or $10 :::: if 1,000 crunchers can each donate at least $5, many of these problems get resolved PERMANENTLY. Thanks for anything you can do.  |
||
|
|
Wingardium Leviosa
Cruncher Joined: May 4, 2020 Post Count: 3 Status: Offline |
Are OPN workunits being properly validated?
Hi everyone, hope that I'm not repeating an issue already discussed. On my results page, OPN workunits (not OPNG) are mostly set to valid and BOINC credits are granted exactly the same as claimed. When I click into the job, there is no replication. I've observed a few old workunits received the "averaged" credit, but the most OPN workunits returned seem to be set immediately valid without replication. Interestingly there are OPN workunits in Pending validation, although there is no other replication waiting. Is OPN validator working?! The greatest concern of mine is whether the researchers get rigorously validated data. I hope that we are not wasting electricity and CPU time. In the meanwhile MCM seems to work as before, waiting for replications returned and claimed credits averaged. Any thought or explanation would be appreciated. Cheers, WL |
||
|
|
alanb1951
Veteran Cruncher Joined: Jan 20, 2006 Post Count: 983 Status: Recently Active Project Badges: 
|
@Wingardium Leviosa
You asked "Are OPN workunits being properly validated?"... What you are observing is a BOINC mechanism referred to as Adaptive Replication - some projects (such as OPN1/OPNG) are suited to its use whereas others (such as MCM1 and ARP1) are not! Adaptive replication considers whether a client is "reliable" based on the number of recent consecutive valid results returned. If the first client to request a task for a work unit is considered reliable the system will send it a task and wait for it to return; if the first requester is not [yet] "reliable" the system will arrange for a second task to be sent out more or less immediately. When the single task returns (assuming it isn't a failure return of some sort), there's another hoop to jump through - s simple random number generator is used to ensure that not every singleton gets validated without a wingman, providing some primitive monitoring! If it passes that test, the result will be passed to the validator which then checks whether the returned result meets basic validation criteria[1]... Adaptive replication is unsuitable for use on projects such as ARP1 because they require extreme precision in the results (I believe ARP1 validates using binary data matching!) or because they only have a single work unit looking at a particular subset of the data and require results to match within some limits that the validator can check. Other projects may have many work units looking at the same subset of data but with slightly different control parameters -- those are often more suited to adaptive replication... Hope this helps.. Cheers - Al. P.S. I am not a WCG staff member :-) [1] It is not unknown for a task to get Validation Error rather than Invalid -- sometimes result files are incomplete or have obviously invalid contents :-) |
||
|
|
TPCBF
Master Cruncher USA Joined: Jan 2, 2011 Post Count: 1957 Status: Offline Project Badges: 
|
The issues we are aware of but remain unresolved: Well, that might not be all of it...Devices missing from the Devices list in My Contributions (In progress) By our estimate, devices should appear within your Devices page within a week. If this does not happen, feel free to send us a support email (support@worldcommunitygrid.org) with your device’s name & host ID so we can add it manually. Sorry, but you can't be serious about this. This is an ongoing issue since last year, it just doesn't seem to be feasible to request people to send device IDs per email, when according to the feedback on the forum, people never get any response or see the issue for them resolved when they send a request via email. And you can't keep doing this for potentially hundreds if not thousands of users User progress as shown in the BOINC Manager is not reflected in your My Contribution page. (In progress) There is an issue where the queue that connects the BOINC database to the website database to update user contributions omits credits from devices that are not yet listed in the website database. Regardless of whether you can see it in your Overview page or not, all of your contributions have been processed and sent to their respective projects. Once resolved, the indeterminate period a user must wait to see a new device listed in Devices or My Contribution on the website after it is successfully added to BOINC will no longer be required. We are also working to add any stats pre-dating the time at which your device was first visible in My Contribution. This doesn't look like a serious approach to me neither.I didn't experience this problem myself in the past, but do now, after I added 3 higher end processing hosts in the last 10 days. Not only do those devices not show up in the device list, while they are showing up in the "Results" list, it doesn't seem like those credits accumulated by those devices are properly accounted for in any stats, internal or external. Also not sure if with the above mentioned "pred-dating" you are also referring to all the work done between mid June 2022 and September 28th, 2022 and which has not shown up yet, nor was there any mentioning that this was actually being worked on. And this is stuff going back 10 months by now. With all the talk of being out of drive space and the recent hardware issue, I honestly have very little confidence that this will ever happen. Though I would love to be proven wrong with my assessment... BOINC Cross Project Identifier is not updated correctly on WCG. We have confirmed that the root cause of this issue is linked to the issue where new devices added by users do not appear in a timely manner in the Devices list and My Contribution section of the website. We should be able to resolve this issue between our message passing software and website database shortly. Missing pagination buttons on project statistics history This issue will be resolved with the next content update (e.g. News) to the website. ...HTTP errors causing download issues (In progress) With multiple updates made by our cloud provider to the network node in the data centre, we have successfully mitigated the worst of this issue. However, during peak loads users may still be affected. The “Retry Now” button in the Transfers tab of the BOINC Manager is the workaround to stalled downloads. Well, how do you define "peak loads"? When you released both ARP1 and OPNG WUs last week, this was not just a "peak" issue, but a constant one for the whole time until the release of those project WUs was apparently halted. And not all users have the time to continously hitting the retry button, and with the recent variation of the upload errors, this resulted in an increased use of bandwidth, on both sides. That can hardly be called a work-around...Overabundance of workunits pending validation This is an issue caused by other users who downloaded tasks not completing them by the deadline, stalling the pipeline for others to download their own WUs. We are monitoring situations like this and will begin sending our additional copies to speed up the process. I think this is not a proper assessment of the problem. I haven't specifically checked today yet but a common issue observed since weeks before the hardware crash was that there were WUs stuck with "Pending validation" when all results for the required quorum was in fact returned. So no missed deadline of similar issues as an explanation...Ralf  |
||
|
|
Wingardium Leviosa
Cruncher Joined: May 4, 2020 Post Count: 3 Status: Offline |
@alanb1951
Yes, it helps a lot, many thanks for your kind, detailed explanation. I appreciate it! It feels definitely not bad that my cruncher is considered reliable :-) Cheers, WL |
||
|
|
Jean-David Beyer
Senior Cruncher USA Joined: Oct 2, 2007 Post Count: 339 Status: Offline Project Badges:
|
Toronto is a major film hub. There's plenty of bandwidth over there. It's just getting to it... ;) I thought the problem was when going from Canada (where Toronto is) to USA (where I am) was caused by the delay going through US Cusoms at Niagara Falls or Buffalo.  |
||
|
|
|