Advanced search

Forums : Technical Support : Cosmology@Home Jobs "Unmanageable"
Message board moderation

To post messages, you must log in.

AuthorMessage
Jarried Pille

Send message
Joined: 4 Jan 08
Posts: 3
Credit: 125,796
RAC: 0
Message 22320 - Posted: 22 Dec 2019, 20:25:54 UTC

I believe this is due to my installation of VirtualBox 6.0, but please read on!

I recently started contributing to various projects via BOINC again and I recently have also installed VirtualBox 6.0 for a VM that I use for home automation (Home Assistant - go check it out!). Unfortunately it seems that after installing this for my own personal VM, that this has caused an issue with running the Cosmology@Home jobs that are also running in a virtual environment.

#1 - In the Tasks list under the status column for these jobs, it states "Postponed: VM job unmanageable, restarting later (4 CPUs)"

#2 - When I fire up the VM VirtualBox Manager, I can see these VMs in the list with my own VMs. Some are "Powered Off" and some are "Aborted."

Any easy (or difficult, I'll do what I need to do to get both running, if at all possible) solutions to:

#1 - Actually get these tasks past the few percent that they seem to run before erroring out?
#2 - Remove these VMs from my Manager window?

I tried searching but didn't really come across much for some reason. I thought that this might be a common problem for people who use VirtualBox for other VMs, but maybe I can't search or I'm an edge case. Or both.

-Running on Windows 10
-BOINC v7.14.2 (x64)
-Have not really tried anything to resolve this issue. I need VirtualBox installed to manage my own VM and prefer to have the latest and greatest out.
-All the Cosmology@Home tasks seem to have this issue as each application ends in vbox64_mt (e.g. camb_boinc2docker 2.05 (vbox64_mt).
-Event log message for when each of these jobs/tasks run states "Task camb_..... postponed for 86400 seconds: VM job unmanageable, restarting later."

Thanks in advance!

BOINC on!
ID: 22320 · Report as offensive     Reply Quote
Jim1348

Send message
Joined: 17 Nov 14
Posts: 98
Credit: 4,156,367
RAC: 134
Message 22321 - Posted: 22 Dec 2019, 20:48:34 UTC - in response to Message 22320.  

I started to get this error right after I upgraded to VirtualBox 6.1.0. Someone mentioned that it means that VBox can not respond to the requests from the work units in a timely manner, so they error out.

The solution is to go back to VirtualBox 5.2.34 or earlier.
ID: 22321 · Report as offensive     Reply Quote
Jarried Pille

Send message
Joined: 4 Jan 08
Posts: 3
Credit: 125,796
RAC: 0
Message 22322 - Posted: 22 Dec 2019, 20:57:56 UTC

Jim,

Thanks a lot for that information. I'll try to roll back and I will report back with an update.

I wonder when this project will officially support v6.xxx, as support for 5.2xx will end after July 2020.
ID: 22322 · Report as offensive     Reply Quote
Jarried Pille

Send message
Joined: 4 Jan 08
Posts: 3
Credit: 125,796
RAC: 0
Message 22323 - Posted: 24 Dec 2019, 6:35:08 UTC

Jim,

I installed VirtualBox v5.2.34 and both my VM and the Cosmology@Home jobs seem to be fully working! Thanks for the help!
ID: 22323 · Report as offensive     Reply Quote
maeax

Send message
Joined: 21 Dec 17
Posts: 18
Credit: 1,228,676
RAC: 2,236
Message 22470 - Posted: 18 Apr 2020, 9:26:37 UTC - in response to Message 22321.  
Last modified: 18 Apr 2020, 9:39:21 UTC

I started to get this error right after I upgraded to VirtualBox 6.1.0. Someone mentioned that it means that VBox can not respond to the requests from the work units in a timely manner, so they error out.

The solution is to go back to VirtualBox 5.2.34 or earlier.

Boinc have now Virtualbox 6.0.14 and Boinc 7.16.5 on the Homepage as standard.
This Error come not, when C@H docker is running only!
Have stopped C@H boinc to docker, because of the same Error as told here.
No problems with Virtualbox 5.2.32 before.
The parallel tasks are LHC@Home.
ID: 22470 · Report as offensive     Reply Quote
maeax

Send message
Joined: 21 Dec 17
Posts: 18
Credit: 1,228,676
RAC: 2,236
Message 22471 - Posted: 18 Apr 2020, 13:23:10 UTC - in response to Message 22470.  

C@H vboxwrapper: 26200_windows_x86_64.exe
LHC vboxwrapper:26198ab7_windows_x86_64.exe
vboxsvc.exe is set to lower than normal for testing.
ID: 22471 · Report as offensive     Reply Quote
maeax

Send message
Joined: 21 Dec 17
Posts: 18
Credit: 1,228,676
RAC: 2,236
Message 22472 - Posted: 19 Apr 2020, 1:02:55 UTC - in response to Message 22471.  

2020-04-18 21:46:08 (13468): Guest Log: tau_recomb/Mpc = 298.48 tau_now/Mpc = 15014.2

2020-04-18 21:46:18 (13468): ERROR: Vboxwrapper lost communication with VirtualBox, rescheduling task for a later time.
2020-04-18 21:46:18 (13468): Powering off VM.
2020-04-18 21:46:19 (13468): Successfully stopped VM.
ID: 22472 · Report as offensive     Reply Quote
Jonathan

Send message
Joined: 27 Sep 17
Posts: 140
Credit: 6,936,356
RAC: 8,822
Message 22473 - Posted: 19 Apr 2020, 5:06:24 UTC - in response to Message 22472.  

I have had that problem in the past. The only solution was to run less VM tasks at the same time. Maeax, instead of running single cpu boinc_2docker work units, try 4 to 6 core. You just need less concurrent work units at a time.
There are more posts at LHC@home on this issue and that forum is active. It is pretty quiet here.
ID: 22473 · Report as offensive     Reply Quote
maeax

Send message
Joined: 21 Dec 17
Posts: 18
Credit: 1,228,676
RAC: 2,236
Message 22474 - Posted: 19 Apr 2020, 14:18:18 UTC - in response to Message 22473.  

Jonathan,
had the hope for the upgrade to Virtualbox 6.0.20,
because there are changes in VBoxManage, but no success.
Still the same.
ID: 22474 · Report as offensive     Reply Quote
maeax

Send message
Joined: 21 Dec 17
Posts: 18
Credit: 1,228,676
RAC: 2,236
Message 22476 - Posted: 20 Apr 2020, 1:39:26 UTC

ID: 22476 · Report as offensive     Reply Quote
LumenDan

Send message
Joined: 28 Mar 18
Posts: 1
Credit: 822,771
RAC: 1,145
Message 22477 - Posted: 21 Apr 2020, 13:21:34 UTC

I restrict the number of vbox instances by creating an app_config.xml file in the project data folder.
Located at C:\ProgramData\BOINC\projects\www.cosmologyathome.org\ on my Windows client.

The contents of the file:
<app_config>
    <app>
        <name>camb_boinc2docker</name>
        <max_concurrent>1</max_concurrent>
    </app>
    <app_version>
        <app_name>camb_boinc2docker</app_name>
        <plan_class>vbox64_mt</plan_class>
        <avg_ncpus>3</avg_ncpus>
    </app_version>
</app_config>


  • Only the application named camb_boinc2docker is restricted by this app_config.xml file, the camb_legasy application will continue to operate according to the normal BOINC computing preferences.
  • I have also restricted the number of concurrent CPU instances used by the camb_boinc2docker application (<avg_ncpus>3</avg_ncpus>). This reduces the performance impact of the application when BOINC tasks are allowed to run "Always", i chose to allocate half the number of CPUs that are available to BOINC.
  • Optimal settings will vary depending on the specification and usage of your computer system.

ID: 22477 · Report as offensive     Reply Quote
BulletMagnetEd

Send message
Joined: 17 Jun 11
Posts: 1
Credit: 5,646,695
RAC: 1,837
Message 22488 - Posted: 2 May 2020, 12:39:39 UTC - in response to Message 22477.  

I restrict the number of vbox instances by creating an app_config.xml file in the project data folder.
Located at C:\ProgramData\BOINC\projects\www.cosmologyathome.org\ on my Windows client.

The contents of the file:
<app_config>
    <app>
        <name>camb_boinc2docker</name>
        <max_concurrent>1</max_concurrent>
    </app>
    <app_version>
        <app_name>camb_boinc2docker</app_name>
        <plan_class>vbox64_mt</plan_class>
        <avg_ncpus>3</avg_ncpus>
    </app_version>
</app_config>


  • Only the application named camb_boinc2docker is restricted by this app_config.xml file, the camb_legasy application will continue to operate according to the normal BOINC computing preferences.
  • I have also restricted the number of concurrent CPU instances used by the camb_boinc2docker application (<avg_ncpus>3</avg_ncpus>). This reduces the performance impact of the application when BOINC tasks are allowed to run "Always", i chose to allocate half the number of CPUs that are available to BOINC.
  • Optimal settings will vary depending on the specification and usage of your computer system.


I dropped my CPUs down to "4", and that finally fixed it for me. It kind of sucks to have a 3900X and have to restrict the CPUs to "4" though. Thanks!
ID: 22488 · Report as offensive     Reply Quote
Jonathan

Send message
Joined: 27 Sep 17
Posts: 140
Credit: 6,936,356
RAC: 8,822
Message 22489 - Posted: 2 May 2020, 18:49:07 UTC - in response to Message 22488.  

How many concurrent, 4 cpu task can you run at one time? Are you running any other projects or just Cosmology? Are you running the Boinc data directory off a hard drive, SSD or NVMe?
I was running VirtualBox 5.2 on NVMe and it was real good. I could run four, four cpu tasks at once and didn't get the error. I switched the Boinc data directory to a regular hard drive and VirtualBox to the 6.1 version and now I can't run as many tasks. I can two, maybe three concurrently.

You may want to take a look at your tasks as you have a bunch of 'Time limit exceeded' errors.
ID: 22489 · Report as offensive     Reply Quote

Forums : Technical Support : Cosmology@Home Jobs "Unmanageable"