Long running work unit

Everything about the project RNA World
Nachricht
Autor
Benutzeravatar
gemini8
Vereinsvorstand
Vereinsvorstand
Beiträge: 5898
Registriert: 31.05.2011 10:30
Wohnort: Hannover

Re: Long running work unit

#769 Ungelesener Beitrag von gemini8 » 13.01.2017 21:57

Congrats!
Gruß, Jens
- - - - - -
Lowend-User und Teilzeit-Cruncher

Bild Bild Bild
Bild

Benutzeravatar
Michael H.W. Weber
Vereinsvorstand
Vereinsvorstand
Beiträge: 22419
Registriert: 07.01.2002 01:00
Wohnort: Marpurk
Kontaktdaten:

Re: Long running work unit

#770 Ungelesener Beitrag von Michael H.W. Weber » 14.01.2017 13:08

Mine has completed 12855 hrs of run time at 99.06% completion - so far. :roll:

Michael.
Fördern, kooperieren und konstruieren statt fordern, konkurrieren und konsumieren.

http://signature.statseb.fr I: Kaputte Seite A
http://signature.statseb.fr II: Kaputte Seite B

Bild Bild Bild

robertmiles
XBOX360-Installer
XBOX360-Installer
Beiträge: 86
Registriert: 23.02.2010 18:43
Wohnort: northern Alabama, US

Re: Long running work unit

#771 Ungelesener Beitrag von robertmiles » 27.02.2017 01:50

IanEdwardJames hat geschrieben:I know this is probably in the incorrect thread but it is sort of in relation to my earlier issues....

updated boinc to 7.6.22 and Vbox to5.0.26 Now getting a 'VM Hypervisor failed to enter an online state in a timely fashion'

Does it on both of my Win7 machines.

Tried culling the VM and restarting the computer.

Thank you guys, any help is muchly appreciated.
Looks like what I saw when I was still using a 5.0.* version of Vbox, nearly every time. Upgrading to a 5.1.* version of Vbox fixed it.

https://www.virtualbox.org/

However, see the note on which versions of the VM application are compatible with 5.1.* Vbox - currently, only the latest version (1.18) is.
Zuletzt geändert von robertmiles am 27.02.2017 02:18, insgesamt 1-mal geändert.

robertmiles
XBOX360-Installer
XBOX360-Installer
Beiträge: 86
Registriert: 23.02.2010 18:43
Wohnort: northern Alabama, US

Re: Long running work unit

#772 Ungelesener Beitrag von robertmiles » 27.02.2017 02:05

http://www.rnaworld.de/rnaworld/result. ... d=14952905

This task appears to be stuck at 98.765% progress.

The remaining time estimate counts down from 05:51:40 to 05:51:31, then jumps to 05:51:40, over and over.

The elapsed time is increasing normally.

I plan to let it keep trying to finish overnight, but may have to abort it if there is no more progress by tomorrow morning.

The last checkpoint was also at 98.765% progress.


A few minutes later:

Remaining time estimate still in a 10 second loop, but starting a little higher.

I've set BOINC not to get any more tasks from any BOINC project for now, to check if this is interference from other BOINC project applications.

robertmiles
XBOX360-Installer
XBOX360-Installer
Beiträge: 86
Registriert: 23.02.2010 18:43
Wohnort: northern Alabama, US

Re: Long running work unit

#773 Ungelesener Beitrag von robertmiles » 27.02.2017 02:43

A few sections of the VBox.log.3 file, near the end:

Correction - VBox.log.3 was the OLDEST of the log files, not the newest, so quotes from it do not appear to be useful here.

I deleted them.

The VBox file appears to be the newest of the log files, and indicates that the task is still running.

The stderr file indicates that it is still creating checkpoints every half hour.

Neither of these files mentions whether the progress is still increasing - could that be added in the next VM version?

Jacob Klein
Brain-Bug
Brain-Bug
Beiträge: 564
Registriert: 26.07.2013 15:41

Re: Long running work unit

#774 Ungelesener Beitrag von Jacob Klein » 27.02.2017 03:33

RNA World VM tasks run a program in the VM, cmsearch, that is notorious for not giving a good estimate of runtime. The VM is setup to take that estimate (shown in "Show VM Console"), and use it (possibly a multiple of it, maybe 2.5x or 3x), and if the task is still running after the estimation, hold progress at 98.765% while task continues crunching indefinitely until task completion. Note: This means that "Remaining (Estimated)" is totally worthless and unreliable, for RNA World tasks.

I agree it's a bad design, but that's how it works here. I have several tasks that have been happily crunching along, at 98.765%, for several months, and will eventually complete successfully.

Short story: If it still looks like it's progressing normally, let it finish!

Long story: If the "Show VM Console" looks "normal", and Task Manager shows that the task is using a CPU, then everything is working just fine, and you should let it finish - Do NOT abort it! If you want more details on how to further verify that the task is progressing normally, there is an expert method, that I highly do NOT recommend trying unless you really know what you're doing. That method essentially is: Close BOINC, waiting until all VM-related processes are gone, opening Oracle VM VirtualBox Manager, cloning the VM, restoring the snapshot within the clone, starting that cloned VM, hit Ctrl+C to get to a prompt, type "top" (lowercase), hit enter, note the time spent on the cmsearch process within the VM, close the VM, close Oracle VM VirtualBox Manager, wait until VirtualBox.exe no-longer shows in Task Manager, then you may start BOINC again. If you do this on different days, and compare the times that "top" show you, you'll see it is making progress, because "top" will be showing different values. One of my tasks is over 500 days already, kicking butt!
Zuletzt geändert von Jacob Klein am 27.02.2017 03:49, insgesamt 1-mal geändert.

robertmiles
XBOX360-Installer
XBOX360-Installer
Beiträge: 86
Registriert: 23.02.2010 18:43
Wohnort: northern Alabama, US

Re: Long running work unit

#775 Ungelesener Beitrag von robertmiles » 27.02.2017 03:41

Two of my tasks have a minimum quorum of 2, but no wingmate currently has a copy in progress:

http://www.rnaworld.de/rnaworld/workuni ... id=5986207
My copy finished back in 2012!

http://www.rnaworld.de/rnaworld/result. ... d=14952905
My copy appears to be hung at 98.765% progress. I saw the post saying that this happens if it runs enough longer than the initial estimated time, so I plan to let it run longer, but may suspend it again since I'm about to install a new graphics card on that computer.

Could you check if either or these workunits needs to have another copy sent out?

If there are no plans to send out another copy for the older one, are you sure you want to keep it occupying space on your hard drives?

Jacob Klein
Brain-Bug
Brain-Bug
Beiträge: 564
Registriert: 26.07.2013 15:41

Re: Long running work unit

#776 Ungelesener Beitrag von Jacob Klein » 27.02.2017 03:45

To my knowledge, the project has several "non-VM" completions that they want to compare against "VM" completions. But, because that's a different app within BOINC, they've done strange things like set "initial replication" to 1, with "quorum" to 2, such that a VM task will complete successfully, but not be validated until they can manually compare the result against the non-VM version.

I agree they should go through this backlog. It is frustrating. I too have several that are completed, no active wingmen are working on them, and they're just waiting to be manually validated.

Maybe Christian or Michael can get on that? Isn't it just a matter of freaking using WinMerge to compare the text result? What is the hold-up?!?

robertmiles
XBOX360-Installer
XBOX360-Installer
Beiträge: 86
Registriert: 23.02.2010 18:43
Wohnort: northern Alabama, US

Re: Long running work unit

#777 Ungelesener Beitrag von robertmiles » 27.02.2017 03:54

How do I "Show VM Console"?

I tried opening the VBox Manager; it offered no way to do it, but did say that a newer version of VBox is available, and showed these error messages:

Runtime error opening 'C:\ProgramData\BOINC\slots\9\boinc_93e5ab107f24298a\boinc_93e5ab107f24298a.vbox' for reading: -103(Path not found.).
F:\tinderbox\win-5.1\src\VBox\Main\src-server\MachineImpl.cpp[745] (long __cdecl Machine::i_registeredInit(void)).
Result Code:
E_FAIL (0x80004005)
Component:
MachineWrap
Interface:
IMachine {b2547866-a0a1-4391-8b86-6952d82efaa0}

I suspect that the limit before it appears to hang at 98.765% is only 1.5x, but should be increased. My initial time estimate for this task was about 13.5 days; it appears to be in that false hang after only 19.5 days.

Jacob Klein
Brain-Bug
Brain-Bug
Beiträge: 564
Registriert: 26.07.2013 15:41

Re: Long running work unit

#778 Ungelesener Beitrag von Jacob Klein » 27.02.2017 04:32

robertmiles hat geschrieben:How do I "Show VM Console"?
You can use "View -> Advanced View", then go to the "Tasks" tab, then select the task.

When BOINC is running a VM task, after a little bit (a couple seconds up to a minute), the "Show VM Console" button will appear on the left. If you click it, you are basically using Remote Desktop to remote into the task (you might need to accept a warning dialog).... Then when that window is open, be VERY careful not to type or click anything. Just look, then close the Remote Desktop window. Note: You might need to move the mouse around within the window to wake the remote display. Alternatively, you can hit the "CTRL" key on your keyboard, which will wake the remote display without sending any key commands.

Jacob Klein
Brain-Bug
Brain-Bug
Beiträge: 564
Registriert: 26.07.2013 15:41

Re: Long running work unit

#779 Ungelesener Beitrag von Jacob Klein » 08.03.2017 17:09

HURRAY!!!! :P :P

My longest-running task, has just completed!
http://www.rnaworld.de/rnaworld/result. ... d=14949493
http://www.rnaworld.de/rnaworld/workuni ... id=6341780

It ran for 544 days :o and survived all of my Windows Insider testing that I carefully did with it!
I can't wait for a wingman to complete it, then send me to the top of the "Monsters" page 8)

ChristianB
Admin
Admin
Beiträge: 1920
Registriert: 23.02.2010 22:12

Re: Long running work unit

#780 Ungelesener Beitrag von ChristianB » 08.03.2017 19:08

Congratulations

Antworten

Zurück zu „RNA World Discussions (english)“