Running Grid Jobs using projectpy » History » Version 24
Running Grid Jobs using projectpy¶
project.py is a wrapper script that takes a lot of the tedious setup of grid etc out of the hand of the user,
instead using an xml configuration file.
The general instructions for project.py are here:
- Table of contents
- Running Grid Jobs using projectpy
- Getting project.py
- XML file preparation
To use this package in lariat we need to set up the lariatutil product (It should set up lariatsoft automatically).
An example shell session could look like this:
source /grid/fermiapp/lariat/setup_lariat.sh setup lariatutil <current_release> -q <current qualifiers like e7:prof>
To check that you have it set up, try the command
project.py --helpTo actually run grid jobs you need two things:
- an xml file present locally that will configure your job and
- a working .fcl file that will be used by the larsoft instance which needs to be present in your product/fhicl path or, better yet, locally.
XML file preparation¶
Get an example XML file¶
Example XML file for LAriAT batch jobs
Further example xml files can be found in the lariatutil (
lardbt-lariatutil) repository. You can bring it into your working area by doing the following in your srcs directory:
mrb g ssh://email@example.com/cvs/projects/lardbt-lariatutil
There are example xml files are in
What you need to change in the XML file¶
Understanding everything in the xml file:
You will probably want to change the destination directory <outdir> and set the working directory <workdir> as you need.
Release version and your local-only code¶
Make the release match your lariatsoft version:
<!ENTITY release "v01_04_00">
i.e. the release number should be the number of the lariatsoft version you are using. So it has to be present in
If you are using code which is not in the release, include the optional
<local> tag in your
<larsoft> <tag>&release;</tag> <qual>e6:prof</qual> <local>/lariat/app/users/USERNAME/lariatOffline/AwesomeNewModule/localProducts_larsoft_RELEASE_QUALIFIERS</local> </larsoft> BE SURE TO SET THE ACTUAL VALUES OF THE PATH ABOVE!
Or use the tarball option in the file. You have to make the tarball using the script
/grid/fermiapp/products/lariat/lariatutil/(change version)/bin/make_tar_lariat.sh -d localProducts... name_of_the_output.tar if you don't have a local copy.
Rather than list files individually you can use a SAM dataset,
<inputdef> BatchTestRun6326_10Events </inputdef>
A SAM dataset is a selection of files with specific conditions or configurations in common.
More here: SAM datasets
.fcl file preparation¶
The .fcl files need to be in your FHICL_FILE_PATH , so, again they need to be either present in a tagged release or in your localProducts directory.
This also means that after each modification of the .fcl file you need to make install for project.py to pick it up. :(
An example .fcl file is in:
Running your project¶
Once you have these elements you can run:
setup jobsub_client import pycurl project.py --xml <path to your xml file> --stage <your defined stage > --submit
and look for results in your <outdir>
Before you run the next stage you need to create a files.list
normally this would be done by launching:
project.py --xml <path to your xml file> --stage <your defined stage > --check
but, this currently fails for our files, so temporarily the solution is:
ls /path/to/your/output/files/<your defined stage>/CLUSTER_PROCESS/*.root >> /path/to/your/output/files/<your defined stage>/files.list
otherwise the submission will fail by not finding the files.list.
No results. What went wrong?¶
If you don't see output, you run
This lists available logs grouped by the fifebatch headnodes.
For the one that failed,
jobsub_fetchlog -J firstname.lastname@example.org --unzipdir=2586120.0.
This the directory
2586120.0/ full of log files.