Where are the 20 WU´s?
Message boards :
Number crunching :
Where are the 20 WU´s?
Message board moderation
| Author | Message |
|---|---|
 NegraBytes  Send message Joined: 23 May 04 Posts: 23 Credit: 22,427 RAC: 0 
|
Hi! I´ve got 20 WU´s on my Results-Page, but I have zero WU´s left in the BOINC-Software! Where are the WU´s? <img src="http://www.boincstats.com/stats/banner.php? cpid=6d6d94c74b5e58374a5943394ccae2df"> ID: 112580 ·  |
|
MikeSW17 Send message Joined: 3 Apr 99 Posts: 1603 Credit: 2,700,523 RAC: 0 
|
You don't give us much to work with, do you? 1) What version of BOINC are you using? 2) Are you dial-up, or always connected? 3) What messages are you getting in the Log? 4) What platform are you on (Windows, Linux etc)? 
ID: 112585 · |
|
NegraBytes Send message Joined: 23 May 04 Posts: 23 Credit: 22,427 RAC: 0
|
>1) What version of BOINC are you using? Ì´m using 4.41! >2) Are you dial-up, or always connected? dial-up! >3) What messages are you getting in the Log? nothing, i saw the 20 WU´s only on the results page! I never saw an message about this WU´s in the log! >4) What platform are you on (Windows, Linux etc)? Windows XP! ..ok? ;-) or do you want more info?! <img src="http://www.boincstats.com/stats/banner.php? cpid=6d6d94c74b5e58374a5943394ccae2df"> ID: 112600 · |
|
Paul D. Buck Send message Joined: 19 Jul 00 Posts: 3898 Credit: 1,158,042 RAC: 0 
|
>1) What version of BOINC are you using? This is a known issue with that version. 4.42 should have corrected this. Because there are so many platforms and versions of the software out there, we need more information to be able to give an answer that is sensible. ID: 112609 · |
Robert Nelson  Send message Joined: 13 Aug 99 Posts: 43 Credit: 3,632,674 RAC: 1
|
Had the problem as described with 4.41 and dropped back until 4.42 came out. Installed 4.42 on several of my computers shortly after it was posted and kept fingers crossed. It has been about 2 days since the install and have just had several ghost WUs appear on SETI web site that are not showing on one of my computers. These ghost units occured well after the change from 4.41 to 4.42 so are not attibutable to 4.41. Note: when problem versions occur, I deinstall the old version before installing a new version so I am very very sure that no left over stuff is hanging out there. I noted that the system seemed to be exhibiting problems communicating with the internet connection. Have since dropped this particular computer back to an earlier version 4.35 which I found to be pretty stable. It appears that when there is a communication problem the hand shaking breaks down between the client and the hub. The hub thinks that the WU was sent but the client thinks that the hub hasn't got work available or a master fetch error occurred. I hope that my 4.42 experience is isolated as it looked like the problem had gone away. The 4.41 problem exhibited itself very shortly after install. ID: 112635 · |
|
Gary Roberts Send message Joined: 31 Oct 99 Posts: 95 Credit: 2,301,228 RAC: 0 
|
Had the problem as described with 4.41 and dropped back until 4.42 came out. Installed 4.42 on several of my computers shortly after it was posted and kept fingers crossed. It has been about 2 days since the install and have just had several ghost WUs appear on SETI web site that are not showing on one of my computers. These ghost units occured well after the change from 4.41 to 4.42 so are not attibutable to 4.41..... I know Paul said this has been fixed in 4.42 but personally I think that is unlikely. This problem has been around for a long time and on different projects. There has been quite a bit of discussion of this problem over at E@H and Bruce Allen, the Chief Developer over there mentioned recently that he and David Anderson were working on a solution involving implementing a "handshake" protocol so that both the client and server would acknowledge the seccessful sending and receipt of a work unit. In simple terms, the problem is due to some sort of timeout issue somewhere in the transmission chain that the server is unaware of. The server thinks the work has gone but it gets lost somewhere along the way. Until I see an announcement from Bruce Allen that the handshake protocol has been implemented, I believe the problem will continue to show up irrespective of BOINC version. I would imagine that the changes required will be both server-side and client-side and it may take some time for all this to happen. I get "phantoms" or "ghosts" all the time in both E@H and S@H work units. Admittedly they are in small numbers and they don't bother me. However I have seen several people now complain that the problem is so severe that the server refuses to send more work based on a large number of "ghosts". The poor client on the other hand is effectively starved for work. Cheers, Gary. ID: 112701 · |
|
Paul D. Buck Send message Joined: 19 Jul 00 Posts: 3898 Credit: 1,158,042 RAC: 0
|
Actually I said it "should" ... :) Well, I am wrong again ... :( I have to admit I have been sticking with the 4.30 and 4.25 versions as they don't seem to be having problems. And I am coward enough that while John McLeod VII is working on the Daemon Work Scheduler I have been sitting on my hands ... Though I am running 4.35 on the Macintosh ... ID: 112788 · |
|
NegraBytes Send message Joined: 23 May 04 Posts: 23 Credit: 22,427 RAC: 0
|
OK, thanks! ...and what about the new 4.43? PS: I´m sorry about my short posts, but my english is not the best :-( <img src="http://www.boincstats.com/stats/banner.php? cpid=6d6d94c74b5e58374a5943394ccae2df"> ID: 112936 · |
RDC  Send message Joined: 17 May 99 Posts: 544 Credit: 1,215,728 RAC: 0
|
I know Paul said this has been fixed in 4.42 but personally I think that is unlikely. This problem has been around for a long time and on different projects. There has been quite a bit of discussion of this problem over at E@H and Bruce Allen, the Chief Developer over there mentioned recently that he and David Anderson were working on a solution involving implementing a "handshake" protocol so that both the client and server would acknowledge the seccessful sending and receipt of a work unit. In simple terms, the problem is due to some sort of timeout issue somewhere in the transmission chain that the server is unaware of. The server thinks the work has gone but it gets lost somewhere along the way. That would be fantastic if they could do a handshake to acknowledge that the WU was in fact sent out by the server and received by the client. I have a few ghost WU's on both S@H and E@H thanks to BOINC 4.41 but pages of them on P@H (all occurring in less than 5 minutes). That's the fun of developmental software however, finding the bugs and getting them fixed. Until the ghost WU issue is fixed, I'm sticking with the older versions of BOINC that work the best on my 2 computers (4.36 and 4.25) as all the ghost WU's are doing is slowing the completion of the WU's since they have to time out on the server and then be resent to someone else later. ID: 112941 · |
|
Robert Nelson Send message Joined: 13 Aug 99 Posts: 43 Credit: 3,632,674 RAC: 1
|
Thanks folks, It is likely, without some sort of handshake that the ghost units will appear from time to time. It appears though that something happened around 4.41 to cause these to come in bunches. RDC's observation that they come in bunches is what I have seen on the newest versions. Prior to the new versions I had not seen the problem at all. From reading some of the other threads on PP&H and Einstein as well as Gary's response, it appears to be a random thing. I noted on my system when the phantom WU's appeared there was no mention in the log of a download event, so at least from an outward appearence it seems that the client was not even aware that something was coming its way. The client did ask for work, but received either a master fetch error, or no work available from project message. It than tried after the waiting time and had the same type of error, this also resulted in a subsequent phantom, fortunatly the scheduler stretches things out, so the string can be broken. Normally I have seen during high traffic volume times the client say it was downloading and then pausing because it had lost connection and then resuming after a period of time, all noted in the log. However recently I have seen that on the .41 and .42 versions when this problem occured, the client seemingly to lose the ability to talk to a number of projects. A change, in my case back to version 4.35 cleared the communication problem and as a result further generation of phantoms, so something appears to be causing the client to get out of sync. A handshake would go a long way to clearing this. BTW Paul let me thank you for the manual. ID: 112966 · |
|
TPR_Mojo Send message Joined: 18 Apr 00 Posts: 323 Credit: 7,001,052 RAC: 0
|
Until the ghosts, beasties, wargs and other nasties have gone I'm sticking to 4.19 ;)  
ID: 112975 · |
|
Paul D. Buck Send message Joined: 19 Jul 00 Posts: 3898 Credit: 1,158,042 RAC: 0
|
BTW Paul let me thank you for the manual. You and everyone else are welcome ... :) I mean, I made it for you! Even better news, I may be getting back to work on it ... though I seem to be going on drunks again ... ah well ... maybe the new doctor can tell me how to stop that ... :) ID: 112976 · |
|
NegraBytes Send message Joined: 23 May 04 Posts: 23 Credit: 22,427 RAC: 0
|
OK, i installed 4.25, and I've got 5 new downloaded WU's and 5 new ghost WU's! ...and the client shows the setiathome_4.09_windows_intelx86.exe and the setiathome_4.09_windows_intelx86.pdb for download, no download status, and I have already downloaded those files!!! i think thats an error on the seti-server, not on the client! <img src="http://www.boincstats.com/stats/banner.php? cpid=6d6d94c74b5e58374a5943394ccae2df"> ID: 113576 · |
©2024 University of California
SETI@home and Astropulse are funded by grants from the National Science Foundation, NASA, and donations from SETI@home volunteers. AstroPulse is funded in part by the NSF through grant AST-0307956.