Sorry I don't know the correct terminology, but I've noticed that of my four computers, the two which have actually gotten new results to work on since the big slowdown of February 17 both got work from a different major datafile (the 7 Megabyte one that is common across thousands of results) than the ones they had been chewing away at happily for some days.
If this is happening a lot, the extra traffic is really going to slow recovery down.
None of my machines has gotten a second result downloaded yet, so can't see whether it will switch to yet another.
Copyright © 2024 Einstein@Home. All rights reserved.
Scheduler sending work for same major datafile?
)
The problem you pointed out has been fixed.
The project server crashed and when I started the system back up the workunit generator was not running. The result of that problem is that when your computer requested work from the scheduler the workunit generator could not make make new work for your computer that would use the data file you already had. The scheduler then gave your computer some work that was waiting to be sent out but unfortunately that work needed a different data file.
When I noticed the workunit generator was not running I stopped the scheduler to prevent the problem you described. We were able to start the workunit generator and the project is now running without problems.
Thanks.
RE: The problem you
)
Obviously you caught it long before I posted. I just wanted to be sure to speak up in case it had not been noticed.
And, yes, I can see new downloads to my machines are sticking to previously downloaded datafiles.
Let me say how impressed I am with how smoothly you folks keep this operation running.
My thanks to you.
RE: Let me say how
)
I totally agree! Thanks folks!