You can monitor the usage of the muoncoll resources here.

More information about the use of Tier1 resources can be found in this guide.

Preliminary setup

To access to the grid resources @CNAF you need an account on the CNAF gateway: bastion.cnaf.infn.it. Follow this guide for the instructions on how to get an account at CNAF.

Then use the general purpose UI: ui-tier1.cr.cnaf.infn.it

To use grid resources you need a personal certificate. Instructions can be found here.

User certificate and key are usually saved in the files $HOME/.globus/usercert.pem and $HOME/.globus/userkey.pem resp.

You need also to register to the VO muoncoll: follow this link.

Proxy

To generate a proxy:

voms-proxy-init --voms muoncoll.infn.it --vomslife 24:00 --valid 24:00

This creates a proxy valid for 24 hours with a VO extension. You can check your proxy with the command:

If the job is longer than 24 hours, it is going to be aborted for proxy expiration time. To extend the lifetime of the proxy you need to store a proxy credential in a dedicated store.

On Tier1@CNAF the myproxy store is myproxy.cnaf.infn.it. With this command you should store a credential on the myproxy server, it will ask  first your certificate password, then you have to set a proxy password that should be inserted in the submission file:

The maximum lifetime of long-lived proxies on a MyProxy server is one week (168 hours), and it can be prolonged with 24 hour steps, to achieve this:

Last, you have to modify your condor submit file adding these lines:

The previous described method sometimes failed. So inside the VO muoncoll.infn.it there is the possibility to ask for a proxy with a duration of 48 hours. This means that you can submit job that can stay on the grid's queues (i.e. IDLE + RUN time) for at most 2 days. Remember to add

this option to your submit file: delegate_job_GSI_credentials_lifetime = 0

Condor jobs

A simple guide can be found here.

Submit job

First of all, set GSI as the authentication method:

export _condor_SEC_CLIENT_AUTHENTICATION_METHODS=GSI

There are six computing elements for grid submission: ce01-htc.cr.cnaf.infn.it, ce02-htc.cr.cnaf.infn.it, ce03-htc.cr.cnaf.infn.it, ce04-htc.cr.cnaf.infn.it, ce05-htc.cr.cnaf.infn.it, ce06-htc.cr.cnaf.infn.it.

To submit to the second CE you have to use the command:

where test.sub is the submit file which represents the job.

Query job

To check the job status of a single job use:

You can also get the status of all your jobs, your matched user can be discovered with this command:
condor_q -pool ce02-htc.cr.cnaf.infn.it:9619 -name ce02-htc.cr.cnaf.infn.it -l  <condorID>  | grep Owner

Get output

When the job is finished retrieve the output with the command:

Remove job

If something went wrong remove the job with:

Status

Check the status of the available resources using:

Use singularity image

The muoncoll software is released also through docker images or singularity images stored on the cvmfs area: /cvmfs/muoncoll.infn.it/sw/singularity/

The following submit file:

• requires 4GB of memory;
• sets log, output and error files as unique appending the CondorID;
• asks to transfer back the simulation and reconstruction files: sim.out and reco.out.

The executable file:

• creates the jobHome directory and copy inside the job's script: job.sh;
• uses the singularity command to exec the job's script (NB: mount the jobHome directory as $HOME);
• uses gfal util to transfer the output of the simulation to the SE.

The job's script is executed inside the singularity container:

• sources the init file to set the environment;
• clones the git repository with the example scripts;
• cd into the SoftCheck directory inside the repository;
• runs the simulation and the reconstruction example, saving the output in the $HOME directory.