What WUs are coming through here are all crap short ones. Nothing to heat the house. I would prefer some resends instead of these crap new ones.

Maybe this will be a "dry" Weekend (eg. no drinking?)

Called that one hours ago. :^)

---

Fri 31 Jan 2020 07:44:42 PM CST | SETI@home | Project has no tasks available

Splitters have been down for a few hours now.
Getting the odd resend here & there (many of which are just noise bombs).

---

Grant
Darwin NT

There have been a lot of ups and downs lately, but it seems that the downs are winning.

So, is this going to be the new normal for the weekends now?

---

I don't buy computers, I build them!!

So, is this going to be the new normal for the weekends now?

Doesn't look like that. Last week the weekend was the only time of the week my computers could consistently keep their caches full.

---

We are drifting further and further from a situation where the splitters could start running. There's now almost 21 million results in the db and it is still rising despite of splitters having been stopped for 10 hours. So much resends...

About 9 milllion of those 21 million results are probably being held hostage by the ever growing assimilation queue.

---

Interesting

"Results in the field" is DROPING

"Results received in last hour" is SLOWING

AND

"Results returned and awaiting validation" is RISING

Doesn't seem like "awaiting validation" should be going up if "results received" is slowing.

---

Dave

It sure did...

---

Dave

Scheduler was MIA for a while there, but now it's back to "Project has no tasks available"

---

Grant
Darwin NT

OK, we're back - it all went very black there for a few minutes, didn't it?

I'm trying to think through what could possibly have caused all this database bloat. My suggestion of re-running the transitioner to pick up tasks which should have validated, but didn't, seems to have flushed out a few - but not as many as I expected. So what else has gone wrong?

First oddity - WU 3781186004. The middle guy - who seems a perfectly respectable cruncher, with a decent rig and a team member, got his task on 9 December (soon after the limits were raised) - and has done nothing with it. Why? He's returning good work, quickly, now, and has lots of credit at other projects.

Only finger of suspicion I can see right now is 'Driver version 432.00' on Windows 10. And he's returned about 80 good tasks - all of a similar age - in the last day. Did he realise that everything was stuck and downgrade the driver? Could all of this be down to Microsoft (auto update), NVidia (bad driver), and our own long deadlines?

Preserving

8317964641	8873167	9 Dec 2019, 10:04:20 UTC	10 Dec 2019, 4:31:07 UTC	Completed and validated	253.39	128.74	126.77	SETI@home v8
Anonymous platform (NVIDIA GPU)
8317964642	8272778	9 Dec 2019, 10:04:17 UTC	31 Jan 2020, 15:04:19 UTC	Not started by deadline - canceled	0.00	0.00	---	SETI@home v8 v8.22 (opencl_nvidia_SoG)
windows_intelx86
8497103820	8313133	31 Jan 2020, 15:04:03 UTC	31 Jan 2020, 22:52:18 UTC	Completed and validated	765.91	709.91	126.77	SETI@home v8
Anonymous platform (NVIDIA GPU)

Also got a bunch, but not all of them resends. Even got an Astropulse (_0).
Looks like things are moving again ...

Edit: downloads are extremely slow

Tom

And I got a BLC35_1 about an hour ago. The replica has fallen behind again, so I can't (yet) see when it was split - but I hope it wasn't recently.

Edit - it's gone now. 36 seconds on an old, tired, slow CPU. We really should stop splitting these noisy tapes while we're still in deep doo-doo.

Just got a bunch of WU's now, but all are resends _2 or higher.
But downloading them, now that is another thing :-)

I managed to score 50 resends on one of my systems. When they finally downloaded, all done in under 4 minutes. Only 3 of them weren't noise bombs.

---

Grant
Darwin NT

And to add to the issues, it appears that "Result files waiting for deletion" has developed issues for both MB & AP. Both have gone from effectively 0 to over 510k & 13.7k respectively.

---

Grant
Darwin NT

My Inconclusive results are going up too, even though I've only had a handful of Tasks since last night. Last night I had a large number of Inconclusive results that said 'minimum quorum 1' and only listed a single Inconclusive host. I didn't see how a single Inconclusive host task could ever validate. Now, it's very difficult to bring up my Inconclusive tasks lists, but, it seems those tasks are now listed as; https://setiathome.berkeley.edu/workunit.php?wuid=3862758806
minimum quorum 1
initial replication 3

   Task    Computer            Sent                  Time reported                 Status        Runtime CPUtime Credit             Application
8495599283  1473578  31 Jan 2020, 5:02:48 UTC  31 Jan 2020, 21:47:15 UTC  Completed and validated  15.36  12.61   3.59  SETI@home v8 v8.20 (opencl_ati5_mac) x86_64-apple-darwin
8498611906  6796479   1 Feb 2020, 3:00:50 UTC   1 Feb 2020, 4:00:03 UTC   Completed and validated   4.10   1.93   3.59  SETI@home v8 v8.11 (cuda42_mac) x86_64-apple-darwin
8498669733  8673543   1 Feb 2020, 4:01:52 UTC   1 Feb 2020, 5:29:49 UTC   Completed and validated  15.11  13.09   3.59  SETI@home v8 v8.22 (opencl_nvidia_SoG)

So, the single host are now triple hosts, but they are still just sitting there with a number of them showing one or two Completed, waiting for validation hosts, and some with one or two Inconclusive hosts.

Just got a bunch of WU's now, but all are resends _2 or higher.
But downloading them, now that is another thing :-)

I managed to score 50 resends on one of my systems. When they finally downloaded, all done in under 4 minutes. Only 3 of them weren't noise bombs.

How does one tell if the jobs are resends?

---

How does one tell if the jobs are resends?

In this case, we're talking about new - extra - replications of existing tasks , not about resending lost tasks.

You tell from the task name, as shown in BOINC Manager. You need to be able to see the very end of the name - so use advanced view, and make the column as wide as you need. The last two characters are as follows:

_0 - always the first time a workunit has been sent to a cruncher. Every WU has a task _0
_1 - in normal times, usually created at the same as _0 and sent out straight away. At the moment, some are being created and distributed later.
_2 onwards - probably a new replication, because the first two failed to validate (either because they returned different answers, or one of them never returned at all). But again, just at the moment, some results are untrustworthy, so _2 may be created for a safety-check.

And I got a BLC35_1 about an hour ago. The replica has fallen behind again, so I can't (yet) see when it was split - but I hope it wasn't recently.

Edit - it's gone now. 36 seconds on an old, tired, slow CPU. We really should stop splitting these noisy tapes while we're still in deep doo-doo.

OK, it's visible now - WU 3863162100. It turns out to be a replication for a task created as a singleton yesterday morning. And because it overflowed, it's been sent to a third host for checking. And it's gone to a host running Windows 7 - no driver problems. Phew.

Thanks, Richard!

---

Sat 01 Feb 2020 07:08:16 AM CST | SETI@home | Scheduler request completed: got 0 new tasks
Sat 01 Feb 2020 07:08:16 AM CST | SETI@home | [sched_op] Server version 709
Sat 01 Feb 2020 07:08:16 AM CST | SETI@home | Project has no tasks available

The good news is my system(s) are happily crunching along (E@H and WCG). The bad news is "the dry spell" is really dry..... ;)

Tom

---

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