Skip to content

zguangyu/harnessed-jobs

 
 

Repository files navigation

harnessed-jobs

Harnessed jobs for running with the eTraveler under the JH

Some environment variables need to be set in addition to the usual set-ups required to run the lcatr code, eotest, LSST stack, etc.:

export SITENAME=<location where the jobs are run>
export HARNESSEDJOBSDIR=<full path to this package>
export PYTHONPATH=${HARNESSEDJOBSDIR}/python:${PYTHONPATH}
export LCATR_SCHEMA_PATH=${HARNESSEDJOBSDIR}/schemas:${LCATR_SCHEMA_PATH}

SITENAME could be, e.g., BNL, Harvard, or SLAC. The site must have a corresponding folder in the config directory, along with the needed configuration files.

The LCATR_SCHEMA_PATH environment variable can alternatively be set by adding a line to the user's lcatr.cfg file:

schema_path = <full path to harnessed-jobs package>/schemas

The lcatr code will add the directories in that line to its search path for finding the schemas.

For jobs that uses the python/DataCatalog.py module, e.g., SLAC/vendorIngest.py, the path to the datacat module must be added to the python path. At SLAC, one would do

export PYTHONPATH=/afs/slac/u/gl/srs/datacat/dev/0.3/lib:${PYTHONPATH}

Testing the harnessed-jobs set up

To test whether a user's environment is properly set up to run the jobs in this package, one could run the setup_test.py script:

$ python ${HARNESSEDJOBSDIR}/tests/setup_test.py

This checks for the needed environment variables and imports the needed packages. An example execution:

$ python ../lsst-camera-dh/harnessed-jobs/tests/setup_test.py
Environment variable LCATR_SCHEMA_PATH not found.
Please set this env var or ensure your lcatr.cfg file has the following entry:
schema_path = /nfs/slac/g/ki/ki18/jchiang/LSST/lsst-camera-dh/harnessed-jobs/schemas

The harnessed-jobs set up otherwise appears OK.

Using package versions:
  LSST Stack: 9.2
  eotest: 0.0.0.8-4-gca943b9
  harnessed-jobs: proto-v1.0-109-g1a38479
$

There is also a General/harnessed_jobs_setup job that does essentially the same thing that could be added to the start of a process traveler to ensure that the package is properly set up.

About

Harnessed jobs for running with the eTraveler under the JH

Resources

License

Stars

Watchers

Forks

Packages

No packages published

Languages

  • Python 79.3%
  • Groovy 16.4%
  • Gnuplot 2.4%
  • Tcl 1.1%
  • Other 0.8%