Development for LHC@home

You have to fix the soft_link.

Not sure how to do that. Will need to investigate. For reference in the version.xml we have:

  <file>
    <physical_name>Theory_2019_02_20.xml</physical_name>
    <logical_name>vbox_job.xml</logical_name>
  </file>
  <file>
      <physical_name>cranky-0.0.21</physical_name>
      <logical_name>cranky</logical_name>
  </file>

In vbox_job.xml we have:

        <copy_to_shared>cranky</copy_to_shared>

but cranky contains:

<soft_link>../../projects/lhcathomedev.cern.ch_lhcathome-dev/cranky-0.0.21</soft_link>

I could put this script into cvmfs or the VM image but it should be possible this way.

You have to fix the soft_link.

Not sure how to do that. Will need to investigate.

RTFM

    [ <copy_file/> ]

New version on it's way.

New version on it's way.

lhcathome-dev 20 Feb 16:33:59 No tasks are available for Theory Simulation

Edit: I have a running task now.
Probably the soft link is not working, because it's in the shared map and used by the VM,
so copy_file was helpful.
I also can follow the event processing.

I see it failed. Could not find the output. Another shared folder issue. I will investigate later this evening.

First task ready: https://lhcathomedev.cern.ch/lhcathome-dev/result.php?resultid=2753195

Outcome: Computation error

lhcathome-dev 20 Feb 18:49:12 Computation for task Theory_2279-782636-20_0 finished
lhcathome-dev 20 Feb 18:49:12 Output file Theory_2279-782636-20_0_r696699143_result for task Theory_2279-782636-20_0 absent

What went well:
The application ===> [runRivet] Wed Feb 20 15:44:00 UTC 2019 [boinc pp jets 7000 600 - pythia6 6.427 z2-lep 100000 20] succeeded.
Directly after the successful finish, the VM was closed by the software by placing the shutdown file in the shared directory.
The result tgz file (20kB) was present in the shared folder. I did not see, it was uploaded (nothing in BOINC Manager Event log)

After 5 minutes the task gave a computation error, the folders and VM was cleaned.

Just got this error code:

upload failure: <file_xfer_error>
  <file_name>Theory_2279-791249-20_0_r485709403_result</file_name>
  <error_code>-240 (stat() failed)</error_code>
</file_xfer_error>

I see it failed. Could not find the output. Another shared folder issue. I will investigate later this evening.

It seems that the output.tgz is not moved from the shared folder to the e.g. name Theory_2279-781724-20_0_r1563603091_result file in the project directory to be uploaded.

I see it failed. Could not find the output. Another shared folder issue. I will investigate later this evening.

It seems that the output.tgz is not moved from the shared folder to the e.g. name Theory_2279-781724-20_0_r1563603091_result file in the project directory to be uploaded.

The shared folder should be included in the open_name

<open_name>shared/output.tgz</open_name>

The shared folder should be included in the open_name

<open_name>shared/output.tgz</open_name>

Yep but this will break the native app. Will think of a solution.

The shared folder should be included in the open_name

<open_name>shared/output.tgz</open_name>

Yep but this will break the native app. Will think of a solution.

I don´t understand why that should break the native app.

The app places the output result file in the shared folder and BOINC will handle the rest, I suppose.

I don´t understand why that should break the native app.

The app places the output result file in the shared folder and BOINC will handle the rest, I suppose.

The native app does not use the shared directory but both share the same job template. I have added a link with the native app so the output is found in both locations. Once the queue is drained I will submit new tasks with the updated output path.

Laurence, to confuse you. I returned a valid task

Once the queue is drained I will submit new tasks with the updated output path.

The queue is clear and new jobs have been submitted. I saw that the new version of the native app was working fine with the old job description so let’s see if both app flavors are working with the new job description.

Laurence, to confuse you. I returned a valid task

That is very confusing indeed!

The queue is clear and new jobs have been submitted. I saw that the new version of the native app was working fine with the old job description so let’s see if both app flavors are working with the new job description.

The new job descriptions did not work with the updated native app version so I have just made another attempt.

The new job descriptions did not work with the updated native app version so I have just made another attempt.

Looks like it worked.

This vdi refuses to d/l faster than 4.5K so 421.80MB will take a long time.

Yet if I do a speed test or get anything else it will d/l at over 1.0Mbps

2019-02-21 10:17:34 (8256): Guest Log: VBoxService 5.2.6 r120293 (verbosity: 0) linux.amd64 (Jan 15 2018 14:51:00) release log
5.2.28 is now avalaible or 5.2.8 from Boinc-website.
Is it possible to upgrade?

2019-02-21 10:17:34 (8256): Guest Log: VBoxService 5.2.6 r120293 (verbosity: 0) linux.amd64 (Jan 15 2018 14:51:00) release log
5.2.28 is now avalaible or 5.2.8 from Boinc-website.
Is it possible to upgrade?

On VirtualBox.org page 5.2.28 is said, but on the download site it's called 5.2.26
I have 5.2.26 running for a few weeks without issues.
I also tried 6.0.2 and 6.0.4. I got really more often heartbeat issues and so a waiting VM for 86400 seconds.

Error result: https://lhcathomedev.cern.ch/lhcathome-dev/result.php?resultid=2753359

2019-02-21 14:11:26 (4684): Guest Log: 13:11:26 2019-02-21: cranky: [INFO] Running Container 'runc'.

2019-02-21 14:12:06 (4684): Guest Log: 13:12:06 2019-02-21: cranky: [ERROR] Container 'runc' failed.