Hello everyone,

I tried to run some LHC WU's and it did not go well. My tasks end after 10 minutes. I thought it was my old version of virtualbox (6.2) so i replaced it for the one proposed on the virtualbox website. I got exactly the same problem and then noticed the .deb file proposed for my ubuntu version downgraded my virtualbox version to 6. 1 ! If installed the 7.2.6 version and that did not solve their problem.

-I use Xubuntu 22.04 with BOINC 8.2.6.

-My BOINC comes from the distro repos

-My Virtualbox is the "all distributions" version whic bring its own dependencies. I.E: It is supposed not to be dependent on your local librairies or how old they are.

-I set my account to run one task and use two CPU's

Symptoms:

-My tasks use zero or near zero CPU and very little memory. My CPU load is at 85% suggesting i have two threads idle. After 10 minutes, i get a computation error on my task.

Is there a low effort way of knowing what happens ? I am curious. Thanks in advance to whoever will answer me.

Your Stderr says:

[2026-02-27 09:46:47] ** It looks like CVMFS is not installed on this host.
[2026-02-27 09:46:47] ** CVMFS is required to run ATLAS native tasks and can be installed following https://cvmfs.readthedocs.io/en/stable/cpt-quickstart.html
[2026-02-27 09:46:47] ** and setting 'CVMFS_REPOSITORIES=atlas.cern.ch,atlas-condb.cern.ch' in /etc/cvmfs/default.local

Do this
or
If you want to use Virtualbox
Deselect "Run native if available" in your prefs.

Thank you.

Especially for providing a low effort lead. It seems to work out. BOINC downloaded a huge virtual machine, which sort of make sense to me, my CPU load is 100 % or close and i have a task using a humongous chunk of RAM which fits with my souvenirs. And my current task survived 21 minutes !