Skip to content

Latest commit

 

History

History
55 lines (39 loc) · 2.1 KB

running-wmarchive-server.md

File metadata and controls

55 lines (39 loc) · 2.1 KB
Raw

Running the WMArchive Server

Follow this procedure to run the WMArchive Server on your local machine or on a remote node.

On your local machine:

  • Install MongoDB (e.g. brew install mongo on macOS)

  • Install python2.7 and the following packages, e.g. in a virtualenv:

     > pip list
 avro (1.8.1)
 bz2file (0.98)
 Cheetah (2.4.4)
 CherryPy (6.0.2)
 pymongo (3.2.2)

On a remote node:

  • Setup the WMArchive environment by adding the following to ~/.bash_profile:

     source /data/srv/current/apps/wmarchive/etc/profile.d/init.sh
 export JAVA_HOME=/usr/lib/jvm/java

 # Common python packages
 export PYTHONPATH=$PYTHONPATH:/data/wma/usr/lib/python2.7/site-packages

 # Spark setup
 export SPARK_HOME=/usr/lib/spark
 export SPARK_DIST_CLASSPATH=$(hadoop classpath)

# Local WMArchive python packages
export PYTHONPATH=/afs/cern.ch/user/n/USERNAME/WMArchive/src/python:$PYTHONPATH

    Make sure to replace ´USERNAME`.

Both:

  • Clone WMArchive (or your fork) and WMCore.

  • Optionally copy the WMArchive/etc/wmarch_config_local.py.example to WMArchive/etc/wmarch_config_local.py to modify configurations. On a remote machine that runs a WMArchive server already, make sure to change the port.

  • Adapt the ./scripts/run_server.sh and ./scripts/stop_server.sh scripts to your system and copy/symlink them to the WMArchive directory. Make sure to use the wmarch_config_local.py configuration file if necessary.

  • Run ./run_server.sh from the WMArchive directory.

    The script will stop any running server, possibly copy the ./src/{css, images, js, maps, templates}/ directories to ./data/ and then run the server. You will find the log output in the log file you provided in run_server.sh.

To query the remote server: