8 of these - all "success, 0 (0x0)" but marked Invalid and "can't find workunit"
=
2 more on computer 7459057 - without the "can't find workunit" moniker but "success" otherwise
=
eg:
Task 4302920074
Name 15my15aa.11300.7602.438086664195.12.179_1
Workunit 1862340547
Created 6 Aug 2015, 14:37:23 UTC
Sent 6 Aug 2015, 18:41:43 UTC
Report deadline 29 Sep 2015, 4:38:27 UTC
Received 7 Aug 2015, 3:48:47 UTC
Server state Over
Outcome Success
Client state Done
Exit status 0 (0x0)
Computer ID 7589235
Run time 3 min 20 sec
CPU time 35 sec
Validate state Invalid
Credit 0.00
Device peak FLOPS 2,128.26 GFLOPS
Application version SETI@home v7
Anonymous platform (NVIDIA GPU)
Peak working set size 145.78 MB
Peak swap size 166.61 MB
Peak disk usage 0.04 MB

Stderr output
<core_client_version>7.4.42</core_client_version>
<![CDATA[
<stderr_txt>

Unable to handle request
can't find workunit

---

"Success" in this context is that the processing has been completed, without error (during the computation that is)
However the results file or the stderr file has been corrupted, or not written.
The work unit "vanishing" is a strange one, to me this suggests that the WU file has been deleted before the task has been run, making it impossible to do the correlation between the results and some part of the data file. I recall that there has been some discussion about missing output files over in number crunchers.

---

Bob Smith
Member of Seti PIPPS (Pluto is a Planet Protest Society)
Somewhere in the (un)known Universe?

As to the Invalid tasks, the most recent thread on this topic is Stderr Truncations. There are several earlier threads and many scattered posts. It appears that the next BOINC version, v7.6.6, will resolve this problem.

As to the "can't find workunit", I've noticed that also over the last couple days. Workunits and tasks normally get deleted from the database 24 hours after validation. Lately it seems that only the workunits are getting deleted but the tasks are not. I don't know why that's happening, but probably isn't anything for us crunchers to worry about.

EDIT: Ah, I see the task deletion topic is now generating some discussion over on the Panic Mode thread, starting with Message 1710406.

Thanks Rob and Jeff for the notes
=
I don't know why that's happening, but probably isn't anything for us crunchers to worry about.
=
Well, on my 4790K-4.3GHz - single GPU/NV-970 sys - Pending bucket is staying at 500>600 since last couple of *weeks(and the DB shuffles + AP and VLAR storms)

RAC took a nice jump during AP storm; but has fallen by 2000+ with all these Pendings (and no new stream/availability of large credits tasks that *seem to validate faster as well, due to their shorter due dates and participation).

On this 4.3GHz machine:
APs run at <39 minutes with the v 7.10 app; MBs at 1-2 hours; CUDAs at 8-25 minutes

I have backed-off to 2x CUDAs instead of 3, to reduce the pressure of those; and free a core to run the other MBs at 8x (instead of 7x)

Can't win tho - if wingmen *and DB don't funnel thru the pipeline

---

I don't know why that's happening, but probably isn't anything for us crunchers to worry about.

No, it all happens on the server. There's really nothing that you can do about it. So sleep on, on two ears this time. :)

---

Apparently so, Jord. (that was a Jeff B quote-sorry, I'm not very good at *proper quoting in this Forum app yet!) ;)

Apparent that is, that Matt's got his hands full - shuffling DB resources and trying to improve performance

(You GO Matt buddy! You do some amazing stuff for us - and are grossly under appreciated)

THANKS for all you do

Edit> Can I get an Amen?

---

"However the results file or the stderr file has been corrupted, or not written.
The work unit "vanishing" is a strange one, to me this suggests that the WU file has been deleted before the task has been run, making it impossible to do the correlation between the results and some part of the data file. I recall that there has been some discussion about missing output files over in number crunchers"
=
Right rob - have seen the "no output" many times before, with no consequence; but the missing WU is in fact a puzzle.
As well as the apparent "effect" on my cache and new work downloadability; along with seemingly long wait to purge or redistribute/resend or acquire another wingman on these Invalids and Errors

---

The output file I'm talking about is the results file that gets transferred to the server as you see the message "Uploading". This file is deleted from your PC as soon as it is transferred so you will rarely see it. These files are about 30k in size.

---

Bob Smith
Member of Seti PIPPS (Pluto is a Planet Protest Society)
Somewhere in the (un)known Universe?

Teaching how to quote, perhaps also useful for others reading here.

Start a quote with a

[quote]

tag container, and then at the end of the sentence you want to quote, or at the end of the paragraph, end with a

[/quote]

Like so:

[quote]Apparently so, Jord. (that was a Jeff B quote-sorry, I'm not very good at *proper quoting in this Forum app yet!) ;)[/quote]

This will make this quote:

Apparently so, Jord. (that was a Jeff B quote-sorry, I'm not very good at *proper quoting in this Forum app yet!) ;)

Start tag containers always use the thing you want to do, like:

[b][i][u][k][quote][code][list][img][url][size][color]

End tag containers always use the same thing you wanted to do, preceded by a forward slash '/', like:

[/b][/i][/u][/k][/quote][/list][/img][/url][/size][/color]

(You'll notice I did not use [/code] in the above, that's because I cannot escape it so it isn't immediately executed. But of course, it'll be used in the same way, end with a slash in front.)

---

OK thanks BOB

---

Teaching how to quote, perhaps also useful for others reading here.

Start a quote with a

[quote]

tag container, and then at the end of the sentence you want to quote, or at the end of the paragraph, end with a

[/quote]

Like so:

[quote]Apparently so, Jord. (that was a Jeff B quote-sorry, I'm not very good at *proper quoting in this Forum app yet!) ;)[/quote]

This will make this quote:

Apparently so, Jord. (that was a Jeff B quote-sorry, I'm not very good at *proper quoting in this Forum app yet!) ;)

Start tag containers always use the thing you want to do, like:

[b][i][u][k][quote][code][list][img][url][size][color]

End tag containers always use the same thing you wanted to do, preceded by a forward slash '/', like:

[/b][/i][/u][/k][/quote][/list][/img][/url][/size][/color]

(You'll notice I did not use [/code] in the above, that's because I cannot escape it so it isn't immediately executed. But of course, it'll be used in the same way, end with a slash in front.)

===
Hey thanks Jord!

---