Eric J KorpelaVolunteer moderator
Project administrator
Project developer
Project scientist
Send message
Joined: 15 Mar 05
Posts: 953
Credit: 6,145,212
RAC: 8,150
We changed over to an new web server and didn't quite get the name of the beta download directory right in the configuration files.
____________
[seti.international] Philip J. FryVolunteer tester
Send message
Joined: 7 Jun 09
Posts: 233
Credit: 1,603,894
RAC: 163
Does this mean that we get again download errors?
So *no new tasks* would be recommended?
Or the download of the WUs don't work, because BOINC can't find the file?
And the download will be tried and tried .. - and will work if the path on the server is correct?
- Best regards! :-) - Sutaru Tsureku, team seti.international founder. - Optimize your PC for higher RAC (@ SETI@home Main). - SETI@home needs your help. -
____________
>Das Deutsche Cafe. The German Cafe.<
Sorry to butt in, but I don't think the "no new tasks" option is ever recommended...Unless you are trying to clear your cache to do some maintenance on your PC or something of the sorts.
Cheers!
____________
Per aspera, ad astra!
[seti.international] Philip J. FryVolunteer tester
Send message
Joined: 7 Jun 09
Posts: 233
Credit: 1,603,894
RAC: 163
If your PC get only download errors, like we had for a few days, the daily quota of new tasks go down and down .. - maybe to 1.
If you have no tasks still in BOINC, it will last very long that the quota of new tasks/day will increase.
For to prevent download errors, it's recommended to set *no new tasks*.
But .. - this time this is not the problem.
My machine downloaded the tasks successfully recently - no errors.
- Best regards! :-) - Sutaru Tsureku, team seti.international founder. - Optimize your PC for higher RAC (@ SETI@home Main). - SETI@home needs your help. -
____________
>Das Deutsche Cafe. The German Cafe.<
The period when wrong download URLs were being sent appears to have been from about 2 Oct 2012 at 23:22 UTC to 3 Oct 2012 | 1:19 UTC. Only those hosts which were assigned work during that ~2 hours will have been affected.
Thanks Joe, that explains one, but not the rest that happend in late September. Forgive me, if I missed something.
Thanks Joe, that explains one, but not the rest that happend in late September. Forgive me, if I missed something.
That's sort of the other side of one coin. Those which happened in September the WU file wasn't there, but the host was given the address where it ought to have been.
The sequence for those is approximately:
1. The WUs were split Jan. 19, Jan. 20, or Feb. 2. The initial two tasks for each went on the end of the "Ready to send" queue.
2. In mid March the initial two reached the front of the queue, there was no download problem so most were crunched and validated. But the usual fraction needed to be sent out to another host, so adding another task to the end of the queue.
3. On May 8, the WU storage was moved to GeorgeM, which involved physically copying a large amount of data. I don't recall any difficulty at the main project, but guess that Murphy's Law came into effect for part of the Beta WU storage. That isn't proven, but fits with...
4. About June 24, the additional tasks (from 2.) reached the front of the queue, but attempts to download the WU files gave "file not found" errors. Those errors of course again added new tasks to the end of the queue.
5. About Sept. 24, the additional tasks (from 4.) reached the front of the queue, and caused further "File not found" errors.
Those WUs are all cancelled now, so the saga is over. Everything's golden and will remain so forever (no, I don't believe it either).