This can't be right....
6 Sep 2005 18:09:22 UTC
Topic 189841
(moderation:
this WU has one PC reporting an invalid result, out of 3.
Yet the Wu has had credit granted, based on 3 results, one of which is "invalid".
What's the crack?
This can't be right....
)
In fact the pc in question seems to return about 90% of it's Wu's in zero time. It's Hosed. Totally FUBAR...
SO why do i get credit awarded, based on a WU returned by a PC that can't do '1+1' and come up with '2' as the solution?
Come on people, there must be someone out there who knows whats screwing up.
Quorum....3 results that are "simular", so i though. Not 2 that may be simular, and one "SNAFU"...
Anyone care to comment? Please?
They'll find out soon enough
)
They'll find out soon enough that the result is bogus, but I think it is good that you have mentioned it here. It would be a problem if that machine is first, I think then an empty result would be stored.
Of course it defeats the point of having a redundancy of 3.
RE: this WU has one PC
)
There are now three valid results since the 4th has now come in. Please do a keyword search (in message titles) for "validation procedure" and you will find, as the top hit, a full description by BM of how validation works. This was posted less than a day ago in this very list.
Whilst the misbehaving computer is not getting any credit, you are fine and the validation procedure is working as intended. Please note that a quorum consists of three results where the science app completed with an exit status of zero (ie success) and where there is no obvious defect in the result file. This however says nothing about the actual validity of the rwsults produced by that "successful" run. That's the job of the validator.
Cheers,
Gary.