Validate Errors on Work-Unit 86624917
28 Nov 2010 10:24:35 UTC
Topic 195481
(moderation:
Hello everyone.
I did a search on this work-unit and no one seems to have reported anything on it. Work-unit 86624917 (available at time of writing), seems to have a lot of hosts (including mine, the tenth recipient) reporting validation errors. My host didn't record any errors, it seems to have completed the work-unit just fine, and it looks like all those other hosts with validate errors did the same. We can't all be wrong, can we?
I seem to recall there being a problem with validation on ABP work-units a few months ago. Could this be related to that issue?
Soli Deo Gloria
Validate Errors on Work-Unit 86624917
)
It just means there is probably something wrong with the unit itself. In the link you provided in the top section where it gives details about the unit itself is a line that says:
max # of error/total/success tasks 20, 20, 20
That means the project will send that unit to up to 20 different pc's before it gives up trying to let us users try to successfully crunch it. After that the project will run it thru their pc's and try to figure out if it is a bad unit or just what is wrong. In the end it could just be bad, it happens although not very often.
I know about the work-unit
)
I know about the work-unit limit. In SETI, it's set to something like 5, if I recall correctly.
Twenty seems rather high. If it genuinely is a once-off bad work-unit, that's unfortunate for all of us who participated. I considered aborting the work-unit before it started, but I thought, no, I should give it a try. Oh well.
Soli Deo Gloria
RE: I know about the
)
Yeah each project makes its own choice oh how many times to resend it.
Hi! Thanks for reporting
)
Hi!
Thanks for reporting this, I forwarded this to the admins and scientists.
CU
HBE
RE: I seem to recall there
)
Depends on which particular 'problem with validation' you are talking about :-).
Over quite a period of time, I've seen several examples of WUs where the error limit of 20 is eventually triggered. All tasks error out with a 'validate error' unless they are aborted, not returned, or suffer a comp error. The last time I remember seeing one of these discussed was back in early October and this is the response that Bernd gave at the time. It would seem that Bernd already knows about these and has made efforts to minimise their occurrence. Unfortunately, they are still happening.
Cheers,
Gary.
Hi Gary. I wasn't aware of
)
Hi Gary. I wasn't aware of there being more than one. I was thinking of the issue reported in the news a while back, but it seems the thread you linked to is more relevant in this case.
Thanks for your help, too, Bikeman.
Soli Deo Gloria
Hi! Thanks again for
)
Hi!
Thanks again for reporting this one. This is one of a rare case where the input data from Arecibo contains no signal above noise level (usually a sign of a faulty file). The ABP2 app can find nothing at all to send back to the servers and sends an empty result-file, which is treated like a validation error (usually it is, but not in this special case). Anyway the staff has a look on those WUs.
CU
HB
Hello again. I should have
)
Hello again.
I should have checked my other hosts, too. I just noticed this work-unit which my machine completed on 7th November, and it's exhibiting a similar round of validation errors. Thought it would be worth reporting, if it's useful to the scientists.
Soli Deo Gloria