| Index | Recent Threads | Unanswered Threads | Who's Active | Guidelines | Search |
 |
World Community Grid Forums
Category: Support
Forum: BOINC Agent Support
Thread: Project is down again - no new work units available |
| No member browsing this thread |
|
Thread Status: Active Total posts in this thread: 59
|
|
| Author |
|
|
Sekerob
Ace Cruncher Joined: Jul 24, 2005 Post Count: 20043 Status: Offline |
After all it seems that some work is being validated (the BOINC stats counter of Firefox is going up) suggesting some servers have come back from the Sunday maintenance run. Cant upload files or report completed tasks either, so must be on that other 50% of the 2 servers.
----------------------------------------For those not aware, WCG upped the default buffer size to 0.3 days from previously 0.1, but that value only kicked in for new members arriving since late 2006. Present members are advised to change their 'default' also to that value, which will put about 7.2 hours of work into the tasks queue (work buffer). The field to change is "Connect to network about every 0.3 days". On the next successful contact BOINC will start downloading the extra work to bridge longer outages.
WCG
----------------------------------------Please help to make the Forums an enjoyable experience for All! [Edit 2 times, last edit by Sekerob at Feb 4, 2007 3:31:53 PM] |
||
|
|
Movieman
Veteran Cruncher Joined: Sep 9, 2006 Post Count: 1042 Status: Offline |
After all it seems that some work is being validated (the BOINC stats counter of Firefox is going up) suggesting some servers have come back from the Sunday maintenance run. Cant upload files or report completed tasks either, so must be on that other 50% of the 2 servers. For those not aware, WCG upped the default buffer size to 0.3 days from previously 0.1, but that value only kicked in for new members arriving since late 2006. Present members are advise to change their 'default' also to that value, which will put about 7.2 hours of work into the tasks queue (work buffer). The field to change is "Connect to network about every 0.3 days". On the next successful contact BOINC will start downloading the extra work to bridge longer outages. Thanks Sek for the info. I'm at 2 days contact time so I have plenty of work stored except on one machine that I just set up last night. Just wanted to let you guys know what I'm seeing. We'll weather the storm..   |
||
|
|
David Autumns
Ace Cruncher UK Joined: Nov 16, 2004 Post Count: 11062 Status: Offline Project Badges: 
|
It sure is out for the count Sek
----------------------------------------This becoming just a bit of a pain, all too regular The problem is with the back off mechanism before you know it (for those blissfully unaware) 2 days crunching goes down the pan   |
||
|
|
Movieman
Veteran Cruncher Joined: Sep 9, 2006 Post Count: 1042 Status: Offline |
It sure is out for the count Sek This becoming just a bit of a pain, all too regular The problem is with the back off mechanism before you know it (for those blissfully unaware) 2 days crunching goes down the pan Don't even say that in jest.. BUT for those of us who are "blissfully unaware" can you explain the issue? Thanks David |
||
|
|
Sekerob
Ace Cruncher Joined: Jul 24, 2005 Post Count: 20043 Status: Offline |
If you watch the message log and have not set your network connection off ad interim, u see the log reporting ever longer times.... it started on my C2D with 1 minutes and now its at backoff-retry in 3:44 hours.... which is why i stop the network attempting.
----------------------------------------Another new feature in 5.8.8 is, that if u go to the transfer tab, and select a file ready to upload and hit the retry button, network connection is made for 5 minutes. If transmission succeeds, it will remain connected until transmissions are finished. Like u, I'm stocked till about when the wulfs start howling (almost full moon here)
WCG
----------------------------------------Please help to make the Forums an enjoyable experience for All! [Edit 2 times, last edit by Sekerob at Feb 4, 2007 3:41:12 PM] |
||
|
|
David Autumns
Ace Cruncher UK Joined: Nov 16, 2004 Post Count: 11062 Status: Offline Project Badges:
|
Hi Movieman
----------------------------------------Unfortunately not jesting Even if you don't retry yourself like a good old ethernet collision it backs off before trying again and then if the servers are still not up it backs off again adding a few more minutes it's kinda exponential so if the servers stay off air for any length of time the output of the BOINC Clients gets seriously hit. It only us "enthusiasts" that notice these things as I say the rest of the BOINC'ers are blissfully unaware. But before you know it gets chilly in my "office" as my room heaters stop crunching. I'm taking this down time as an opportunity to upgrade all my clients to 5.8.8 and set that queue to 5 days worth |
||
|
|
Movieman
Veteran Cruncher Joined: Sep 9, 2006 Post Count: 1042 Status: Offline |
If you watch the message log and have not set your network connection off ad interim, u see the log reporting ever longer times.... it started on my C2D with 1 minutes and now its at backoff-retry in 3:44 hours.... which is why i stop the network attempting. Another new feature in 5.8.8 is, that if u go to the transfer tab, and select a file ready to upload and hit the retry button, network connection is made for 5 minutes. If transmission succeeds, it will remain connected until transmissions are finished. Like u, I'm stocked till about when the wulfs start howling (almost full moon here) I'm still on 5.4.11 on most machines. Just set up another clovertown last night and it's on 5.8.. I've got app 50-60WU stored there to work on. Just an old DX2400 is running it's last 2 WU now.. Is there anyway to transfer WU from one machine to another. Is it a simple matter of taking a file and moviing to X dir to take from machine with huge log to one thats running out? |
||
|
|
Sekerob
Ace Cruncher Joined: Jul 24, 2005 Post Count: 20043 Status: Offline |
The project files and the client_state.xml file are inseparable. I could conjure up a scenario/sequence how to do that, but than i get the blame for it flying in the face if not working. It goes like this:
----------------------------------------1. Suspend all work in buffer (on machine A). 2. Set the network to OFF and set Fetch work to No. 3. Take a full copy of BOINC and is project dirs 4. Copy it to other machine (call it B). 5. Activate those work units u want crunching on machine A. 6. Activate those work units u want crunching on machine B. 7. Let all do their thing. 8. Once servers are up again flush completed work of machine A. 9. Close BOINC A and delete!!!!!! 10. Copy BOINC from B. to A. 11. Abort the work units originally done by A, but suspended for B. 12. Flush work done by B on A. 13. Reactivate the Fetch work 14. Never end with 13 points, so pray on your bare knees on a pinned cushion that it works. This is entirely untried and I deny all and every responsibility for this out of the box thinking (if u can call it that). 
WCG
Please help to make the Forums an enjoyable experience for All! |
||
|
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
I got nothing from the midday update either. Haven't checked my PV's yet but usually have at least 20 waiting.  HI: I just checked my PV and have about 80 in the stream. Most if not all are waiting for a single return from back in late January (?). Looks like a couple may end up as no returns. Hope that doesn't happen because then I'll have to wait for another fellow to re-compute for the no returns. |
||
|
|
Bearcat
Master Cruncher USA Joined: Jan 6, 2007 Post Count: 2803 Status: Offline Project Badges:
|
How much longer before the server is back up and accepting wu? Also, seen a linux file in the files to download as well as an apple file? What are these for?
----------------------------------------Randy
Crunching for humanity since 2007!
 |
||
|
|
|