Message boards :
Number crunching :
To many Download Errors = 0 Credits ... !!!
Message board moderation
Author | Message |
---|---|
STE\/E Send message Joined: 29 Mar 03 Posts: 1137 Credit: 5,334,063 RAC: 0 |
WU 19151546 WU 19146840 WU 19145859 WU 19145177 WU 19140718 WU 19139565 WU 19142482 WU 19136963 WU 19138846 I noticed these WU's in my Account that were granted 0 Credit because of other peoples Download Errors. I successfully completed them along with 1 other person but neither one of us received any credit for them. Why weren't they sent out some more to get the 3'rd or 4'th result returned instead of just aborting them. Myself I can live without the credit because I know after running the projects this long your not going to get credit for every WU. But to me it's things like this that are going to drive the new people from Seti Classic up the wall once they shut Classic down for good. |
michael37 Send message Joined: 23 Jul 99 Posts: 311 Credit: 6,955,447 RAC: 0 |
You are absolutely right. This has been happening way too often lately. See here: http://setiathome.berkeley.edu/forum_thread.php?id=16504 |
Paul D. Buck Send message Joined: 19 Jul 00 Posts: 3898 Credit: 1,158,042 RAC: 0 |
I have asked Dr. Anderson to consider changing the error allocations so that download errors would not be counted as a problem. It may or may not happen. At the moment, this is not an active change (yet) ... but I did ask. If someone wants they could add it to the bugbase as a feature request ... |
STE\/E Send message Joined: 29 Mar 03 Posts: 1137 Credit: 5,334,063 RAC: 0 |
I have asked Dr. Anderson to consider changing the error allocations so that download errors would not be counted as a problem. It may or may not happen. At the moment, this is not an active change (yet) ... but I did ask. If someone wants they could add it to the bugbase as a feature request ... Thats a good idea Paul, I don't think somebody that successfully returns a work unit should be penalized because somebody else had a download error or the server itself messed up when downloading the work unit ... |
Paul D. Buck Send message Joined: 19 Jul 00 Posts: 3898 Credit: 1,158,042 RAC: 0 |
To be honest, I don't understand why this type of split was not made part of the initial design. I would have, because there are two types of errors, work unit/client errors where the dat in the work unit is the most likely cause, and those that could be caused by the server side. They should be accounted for separately. Not sure if I made my point or not. But, what do I know ... :) As a system engineer I always try to account for things in their own way ... |
bobb2 Send message Joined: 5 Feb 00 Posts: 53 Credit: 380,595 RAC: 0 |
By way of explanation and recognition: I just noticed that all these errors occured on WUs sent June 30. There were severe problems with Apache/UPDL server, so that we are tuning for better performance as the load grows. We run clients in-house and also experienced this problem. Here is a tech news item from June 30: June 30, 2005 - 17:00 UTC Last night the upload/download server ran out of processes. This happened because the load was very heavy, which causes adverse effects in apache. When hourly apache restarts were issued (for log rotation), old processes wouldn't die and new ones would fill the process queue. By this morning we had over 7000 httpd processes on the machine! Apparently some apache tuning is in order. |
Paul D. Buck Send message Joined: 19 Jul 00 Posts: 3898 Credit: 1,158,042 RAC: 0 |
By way of explanation and recognition: Yes, we know ... The broader point is that all of those work units may be errored out and never processed because of a server problem. The way that BOINC counts errors, errors due to the fault of the server are orpaned when they have nothing wrong with the internal data. Erroring out and not trying to process a work unit that kills the client, or you cannot get consensus on is cool ... but the system should not take a Work Unit out of production because the server barffed all day long ... |
Steve Cressman Send message Joined: 6 Jun 02 Posts: 583 Credit: 65,644 RAC: 0 |
Also if you check the ones with download errors you will find that it was using boinc client 4.13. The 4.13 client can't continue a d/l that has been interrupted. As long as people keep using 4.13 we will see this problem. IMO 4.13 should have been phased out long ago. 98SE XP2500+ @ 2.1 GHz Boinc v5.8.8 And God said"Let there be light."But then the program crashed because he was trying to access the 'light' property of a NULL universe pointer. |
W-K 666 Send message Joined: 18 May 99 Posts: 19062 Credit: 40,757,560 RAC: 67 |
This problem of the 30 June was noted in at least two threads on the 1 July, Regarding Credits and Q? 7 Client error downloading . I got the impression that when CC ver 5 arrives that will become mandatory but not before. Andy |
©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.