Yep, every host in scheduler backoff due to "internal server error". Can't report work and caches falling.

Ditto for me. Almost a full moon to howl at...

---

Yep, every host in scheduler backoff due to "internal server error". Can't report work and caches falling.

Ditto for me. Almost a full moon to howl at...

Nothing here too. Caches holding. Ready to do my part.

---

Able to report now, at least. Not giving out work yet, though.

---

Hey, with 36 tasks left, I started getting new work. Now to see if it lasts...

Got home to find Linux system out of GPU work due to extreme backoffs on downloads.

---

Grant
Darwin NT

I see they added some ... {rest lost due to some error?}

. . Yep. I noticed that, most of my workload is now old arecibo work ...

Stephen

<shrug>

I see they added some old Arecibo files to the splitters in an attempt to slow down the return rate from all the noise bombs.

And a lot of them here ATM have been the same. :-(

Cheers.

Well here is a few numbers to keep in mind

Results out in the field - 7,536,318

Results returned and awaiting validation - 7,240,993

First time this year I have seen the returns past 7 million.

Results received in last hour - 195,093

Cheers

Well here is a few numbers to keep in mind

Results out in the field - 7,536,318

Results returned and awaiting validation - 7,240,993

First time this year I have seen the returns past 7 million.

Results received in last hour - 195,093

Cheers

Interesting numbers:

Lots of short tasks....

Tom

---

A proud member of the OFA (Old Farts Association).

Yes, I'm seeing shorties (VHAR) now from 27au10ad. I also saw the noise bombs (overflows) earlier from BLC35.

Neither of those are going to help our bad drivers and bad cards. Eric put in a draconian attempt to try and weed those out yesterday:

Hopefully final validation mod to reduce bad results from failing GPUs
If 1 of 2 is overflow, quorum is increased to 3
If 1 of 3 is overflow, results are validated.
If 2 of 2 are overflow, quorum is increased to 3.
If 2 of 3 are overflow, quorum is increased to 4
If 3 of 3 are overflow, results are validated.
4 results are always validated.

so expect lots of resends, too.

If 3 of 3 are overflow, results are validated.

That is a bug.
If first two are ATI graphics cards, which will produce similar bad results, it doesn't matter what the third result is processed on, if it also -9's then the two ATI cards will agree and the correct third result will receive an "invalid".

See https://setiathome.berkeley.edu/forum_thread.php?id=84508&postid=2026843#2026843

I expect there to be many of them in this batch of blc35.... tasks

If 3 of 3 are overflow, results are validated.

That is a bug.
If first two are ATI graphics cards, which will produce similar bad results, it doesn't matter what the third result is processed on, if it also -9's then the two ATI cards will agree and the correct third result will receive an error.

I think we should monitor for a while. If anyone notices this case in the future, please post a link and we can examine. I think the three would still have to be 'weakly similar', so if the early overflow signals were of a completely different type, that could (should?) be flagged as 'invalid'.

If 3 of 3 are overflow, results are validated.

That is a bug.
If first two are ATI graphics cards, which will produce similar bad results, it doesn't matter what the third result is processed on, if it also -9's then the two ATI cards will agree and the correct third result will receive an error.

I think we should monitor for a while. If anyone notices this case in the future, please post a link and we can examine. I think the three would still have to be 'weakly similar', so if the early overflow signals were of a completely different type, that could (should?) be flagged as 'invalid'.

IIRC in the case quoted the results were, 7 pulses & 23 triplets Vs 30 triplets

IIRC in the case quoted the results were, 7 pulses & 23 triplets Vs 30 triplets

The change wasn't notified to me (and it was an automatic, real-time notification) until 8 Jan at 22:57 UTC. I think that's after the example in your linked post. I'd like to confirm it with an after-change case, if you don't mind: I'm pestering Eric quite a lot at the moment about the Anonymous Platform bug (and another related problem which came to light this morning). I'll certainly pass on an error report if it comes to light with confirmation, but let's do the research ourselves, to save Eric the distraction.

If 3 of 3 are overflow, results are validated.

That is a bug.
If first two are ATI graphics cards, which will produce similar bad results, it doesn't matter what the third result is processed on, if it also -9's then the two ATI cards will agree and the correct third result will receive an "invalid".

See https://setiathome.berkeley.edu/forum_thread.php?id=84508&postid=2026843#2026843

I expect there to be many of them in this batch of blc35.... tasks

. . But if all cards agree the sample is an overflow (noisy) then it does not matter as there is no result going into the database right or wrong. If you are worried about the 1 or 2 credits then what can I say ....

. . If the 2 AMD cards are overflows but the third is not then it must be rechecked as it may be a valid result that would be discarded if the AMD cards are accepted.

Stephen

. .

RTS = 48,589, this is still a cause for concern

---

Uploads are getting sticky and backing off in retries. Takes a couple of manual interventions to move them along.

Website is . . .SSSLOOOOWWW . . . .

---

Seti@Home classic workunits:20,676 CPU time:74,226 hours

A proud member of the OFA (Old Farts Association)

FUBAR

---

To add to the general merriment reporting seems to be broken, at least for me.

---