LHC@home

Tomorrow morning i see on all Computers Atlas Wu's only with 110%-150% CPU Load but ~10h run time. CPU Load Value should be at 350-390% with 4 threads.

I think all Wu's from yesterday are crashed, but without working Alt-F2 it's not so easy to get something displayed.

There must be a problem in the actual set of Wu's older version 2 Wu's worked normal.

Got some VM Text with Kernel Panic displayed.

Some team members got the same Problem.

Got on one VM following error:

https://c.web.de/@309282286350637121/m6p10ZAuQlOkGGaU2RipmQ

I 've seen this error also on a machine at WU restart.

vboxsf --> vbox shared folder driver

best cases:
- upgrade to the most recent vbox version
- reset the project (to get a fresh vdi if your's got damaged somehow)

worst case:
The vbox extensions inside the vdi are not compatible to the running VM kernel any more.
This needs to be investigated/fixed by the CERN developers.

I aborted some tasks. Although set up with 4 cores (out of 8), The VM is using 'only' 20-22% of the CPU.
Consoles F2 and F3 are not showing information, so no idea what the VM is doing - hence abort.

Same problem today. Atlas job do not start.
In the vbox.log is no line Guest Log: *** Starting ATLAS job. (PandaID=...... at startup.

Same here,
it need to roll back to the previous Version from yesterday, without Alt+F3 feature.

I have rolled back yesterday's changes. I'm travelling today so don't have time to investigate what the problem is but will look at it on Monday.

Thank you David.

The problem is now understood and fixed. In the next hour or two new WU will have the fixed version of the top console activated.

This sounds good.

Thanks for the fast reaction on Friday.

On my i7-5820k wu's are failling also the whole weekend. Updated virtualbox now. Hope it will solve the Problems on this machine.