Development for LHC@home

The vboxwrappers have been upgraded to 26155 so that we are now in sync with vLHC@Home.

Please let us know how things are going by posting on the message boards.

Thanks,

Laurence

The vboxwrappers have been upgraded to 26155 so that we are now in sync with vLHC@Home.

Does this mean we can move beyond VirtualBox 4.3.12 on Windows now?

---

The vboxwrappers have been upgraded to 26155 so that we are now in sync with vLHC@Home.

Does this mean we can move beyond VirtualBox 4.3.12 on Windows now?

Yes. You can move up to VirtualBox v4.3.26.
It's advisable to install the extension pack 4.3.26 too.

The vboxwrappers have been upgraded to 26155 so that we are now in sync with vLHC@Home.

Does this mean we can move beyond VirtualBox 4.3.12 on Windows now?

Yes. You can move up to VirtualBox v4.3.26.
It's advisable to install the extension pack 4.3.26 too.

Today I installed 4.3.26 on a SLC6 and a Mint Linux box. On both, the first task soon went into "Waiting to run (Scheduler wait)" mode. I poked around on the Mint machine and found the task slot. stderr says:

2015-03-18 20:58:19 (18361): vboxwrapper (7.5.26154): starting
2015-03-18 20:58:19 (18361): Feature: Checkpoint interval offset (72 seconds)
2015-03-18 20:58:19 (18361): Could not communicate with VM Hypervisor. Rescheduling execution for a later date.

(and there's the same on the SLC6 box).
I'm still running a task with 4.3.12 on my Windows PC, I'll check it out on Friday (I'm off-campus tomorrow).

---

Today I installed 4.3.26 on a SLC6 and a Mint Linux box. On both, the first task soon went into "Waiting to run (Scheduler wait)" mode. I poked around on the Mint machine and found the task slot. stderr says:

2015-03-18 20:58:19 (18361): vboxwrapper (7.5.26154): starting
2015-03-18 20:58:19 (18361): Feature: Checkpoint interval offset (72 seconds)
2015-03-18 20:58:19 (18361): Could not communicate with VM Hypervisor. Rescheduling execution for a later date.

(and there's the same on the SLC6 box).
I'm still running a task with 4.3.12 on my Windows PC, I'll check it out on Friday (I'm off-campus tomorrow).

Recently I had 2 computation errors, because Vbox could not detect the right device/snapshot after a task suspend/stop and a requested resume/BOINC restart.
The first was wrapper 26105 with vbox 4.3.24, the second wrapper 26155 with vbox 4.3.26, also BOINC Manager displaying a VM Hypervisor message whilst waiting to start.
I'll try to reproduce the error.
For that I'll reduce the runtime to 3600 seconds again.

I immediately got "computation error" after switching to the latest version. I had previously removed the old VM by hand. This "computation error" persists even after a project reset!

I immediately got "computation error" after switching to the latest version. I had previously removed the old VM by hand. This "computation error" persists even after a project reset!

Hi Ben,

What happens, if you upgrade to a newer VBOX-version with extension pack installed?
I ran one shortened task without any intervention of my site with success.

http://boincai05.cern.ch/CMS-dev/result.php?resultid=25660

I immediately got "computation error" after switching to the latest version. I had previously removed the old VM by hand. This "computation error" persists even after a project reset!

Hi Ben,

What happens, if you upgrade to a newer VBOX-version with extension pack installed?
I ran one shortened task without any intervention of my site with success.

http://boincai05.cern.ch/CMS-dev/result.php?resultid=25660

Aha, all is explained. I am on a Mac. We have hit the known bug with the vboxwrapper upgrade on MacOSX and Linux, the one that forced us to revert to 26079 for these systems after the upgrade on T4T.

My omission not to advise Laurence and Hendrik about this. We will revert to 26079 on CMS too for MacOSX and Linux. The new wrapper works fine on Windows.

I have sent the stderr details to Rom and Charlie for their debugging.

I'll try to reproduce the error.
For that I'll reduce the runtime to 3600 seconds again.

http://boincai05.cern.ch/CMS-dev/result.php?resultid=25660 - Run without interruption
http://boincai05.cern.ch/CMS-dev/result.php?resultid=25601 - Run without interruption
http://boincai05.cern.ch/CMS-dev/result.php?resultid=25682 - Run with pause, suspend (LAIM off) and BOINC restart
http://boincai05.cern.ch/CMS-dev/result.php?resultid=25588 - Run with suspend and reboot without stopping BOINC first

All success runs. I could not reproduce the error.
I'll ask Rom to check this task: http://boincai05.cern.ch/CMS-dev/result.php?resultid=25156

Before midnight I suspended the task, stopped BOINC and put the machine into sleep mode.
In the morning I woke up the machine, started BOINC and resumed the task.

Today I installed 4.3.26 on a SLC6 and a Mint Linux box. On both, the first task soon went into "Waiting to run (Scheduler wait)" mode. I poked around on the Mint machine and found the task slot. stderr says:

2015-03-18 20:58:19 (18361): vboxwrapper (7.5.26154): starting
2015-03-18 20:58:19 (18361): Feature: Checkpoint interval offset (72 seconds)
2015-03-18 20:58:19 (18361): Could not communicate with VM Hypervisor. Rescheduling execution for a later date.

(and there's the same on the SLC6 box).
I'm still running a task with 4.3.12 on my Windows PC, I'll check it out on Friday (I'm off-campus tomorrow).

I installed VB 4.3.26 on my Windows machine, reset the project, and downloaded a task, It seems to be running "normally".
On the other hand, after resetting the project on my SLC6 box and downloading a new task, it immediately went into the Scheduler wait state.

---