eOn Client 4.00

[EdK]

Is this still a beta? Why doesn't the Progress bar work? And why does the remaining time stay at "---" while the application is still running?

Does this make sense?

eOn 4.00.png (4.67 KiB) Viewed 361579 times

[skivelitis]

I agree. How can we tell that these are not broken, never ending work units? I grow concerned when tasks run for hours with no measurable progress. How long can one expect this version to take on a Q6600 Win Vista 32?

[Slave_Mac]

Client 4.00 has just failed 23 tasks on my XP machine. Error was instantaneous on all tasks.
Can be viewed while they last, http://eon.ices.utexas.edu/eon2/results.php?hostid=950
error -
Stderr output

<core_client_version>6.12.34</core_client_version>
<![CDATA[
<message>
Incorrect function. (0x1) - exit code 1 (0x1)
</message>
<stderr_txt>

</stderr_txt>
]]>

I have two 4.00 tasks running on a Lion laptop which have not completed yet.
One of these two finished over night w/error,
task 107812231

Stderr output

<core_client_version>6.12.35</core_client_version>
<![CDATA[
<message>
Maximum elapsed time exceeded
</message>
<stderr_txt>

Crashed executable name: eonclient_4.00_x86_64-apple-darwin
built using BOINC library version 6.12.26
Machine type Intel 80486 (64-bit executable)
System version: Macintosh OS 10.7.3 build 11D50
Thu Feb 16 01:21:42 2012

[_MaRiO]

Dawn U eOn !
First Error with credits viewtopic.php?f=9&t=1085 - no answer
Second U release Beta version and error
Now U make alfa version and same error
And this http://eon.ices.utexas.edu/eon2/workuni ... =116974153
21,837.62 - - 21,284.09 - - 7.68 - - eOn Client v4.00

R U ALIVE?

[LAZA]

As i can see, there are NO checkpoints, no progress bar, nor does this make sense!

Beta WU 497477662_0_73 has a runtime (at the moment) 4 hours 54 minutes!
Beta WU 497477662_0_72 and Beta WU 497477662_0_50 have an predicted runtime 5 minutes, 47 seconds!

What the hell is there going on???
PC works all the night without any interruptions but at the morning a want to pause all BOINC processes but i don't want to start over again those 5-hour-WU for nothing.

For Beta this would be halfway acceptable, but since thursday they are declared as stable and this is big BS!

Edit:

Do 16 Feb 2012 12:03:47 CET eon2 Aborting task 497477662_0_73_0: exceeded elapsed time limit 18935.372368
Do 16 Feb 2012 12:03:48 CET eon2 Computation for task 497477662_0_73_0 finished
Do 16 Feb 2012 12:03:48 CET eon2 Output file 497477662_0_73_0_0 for task 497477662_0_73_0 absent

I stopped to process eon2 till this bugs are past!

[Gemjunkie]

Can we get some guidance from admin? What are we supposed to do with these long-running WUs? Especially when they exceed the deadline.

http://eon.ices.utexas.edu/eon2/workuni ... =116971973

2012-02-16_095731.jpg (52.56 KiB) Viewed 361533 times

This project requires way too much babysitting.

[ryeterrell]

Hi everyone,

Let me apologize for the trouble the latest client has been giving you. We know you put a lot of work into making your crunching as efficient as possible, and that it can be a real damper on your efforts when a client misbehaves like this.

I have stopped sending out the jobs that have been causing problems. Those should work their way out of the queue over the course of the day. The shorter jobs you have been processing should take over the queue shortly after that.

Again, I am sorry for the time and effort this has cost you. Thanks for your patience and feedback.

Rye

[Tex1954]

Yuppers to others problems... have OVER 100 e0n 4.0 tasks fail instantly so far and it sure isn't my system doing it...

:D

PS: Yes, tried to detach/reattach... no difference... all bad...

[Gemjunkie]

I would love WUs that fail instantly, I get WUs that run for >24 hrs. and abort themselves for passing the deadline for 0 credit.

[upstatelabs]

I have literally hundreds of these WU that have failed. Fortunately most failed immediately or within 5 seconds. Only a handful ran for 5 or 10 hours before failing, so not a major computational waste, although I imagine a bit of bandwidth. I have a few that say they completed, but the error log says they stopped due to memory segmentation errors or failure to create file errors, etc.
There seem to be WUs in the ready-to-send queue, hopefully these are not more of these.

[Ged T]

I've just aborted running and yet to run eOn Client 4 beta (version 401...) items - They clearly do not work and I prefer, expecially in these austere times, to dedicate cpu cycles (and the electricity cost of exercsing them) to projects that are able to yield viable, post-client results.

There seems to be a school of thought w.r.t. to eOn client software production that ignores the fact that many projects run in the BOINC context: these non-checkpointing (PLEASE get that sorted out - It doesn't matter if the WU are short-lived, but increasingly, that is NOT the case!) and clearly 'minimally' tested clients deny processing to those other collaborating projects - Not good for science overall.

Is it possible that someone, on behalf of the project owners, will confirm through this forum when the eOn client software is worthy? I might be tempted to help UT once more...

[Gemjunkie]

I've just aborted running and yet to run eOn Client 4 beta (version 401...) items - They clearly do not work and I prefer, expecially in these austere times, to dedicate cpu cycles (and the electricity cost of exercsing them) to projects that are able to yield viable, post-client results....

Deselect 'Run test applications' under 'eOn preferences' from your account page. The regular WUs are somewhat less annoying. Somewhat.

It's insane not having check-pointing for WUs where it's feasible and shows a disregard for the volunteer crunchers. There are better projects. See you there.

[Tex1954]

It would appear the developers are allowing these bad WU's to run... have no idea why, but after hundreds of instantly failed WU's, detaching, reattaching, try again... well, time to give up on 4.0 I think.

:(

[upstatelabs]

OK, I've got a whole pile of new failing WUs. Please make it stop!!
It's pretty ridiculous really. Within 10 min of the tap being turned on at around 18 Feb 2012 22:00UTC, the failed WUs started coming back to the project. Why not turn it off right away? Why keep it running for hours and sending me hundreds of these things? Not a very kind thing to do to your volunteer crunchers.

[Tex1954]

I've emailed the folks on the web page in case they are not aware of the problem for some reason...

:|