Operations Guides

There are 2 ways to operate the HPC Gateway processes. This depend if you are “root” user or “hpcgadmin” user.

As "root" user, /etc/init.d method

You have to use /etc/init.d/hpcg command. This command allow to status, start, stop, or restart all HPC Gateway processes (php-cgi, mongo, jetty, cluster) in 1 go:

  • /etc/init.d/hpcg status : status all HPC Gateway processes
  • /etc/init.d/hpcg start : start all HPC Gateway processes
  • /etc/init.d/hpcg stop : stop all HPC Gateway processes
  • /etc/init.d/hpcg restart : restart all HPC Gateway processes

It is possible to manage each HPC Gateway processes, one by one by using:

  • /etc/init.d/hpcg_php-cgi : manage php-cgi process used for dokuwiki system
  • /etc/init.d/hpcg_mongo : manage mongod process, i.e. the MongoDB database
  • /etc/init.d/hpcg_jetty : manage jetty process, i.e the Application Server
  • /etc/init.d/hpcg_cluster : manage Cluster Agent process

The order of start is php-cgi, mongo, jetty, cluster. The order of stop is cluster, jetty, mongo, php-cgi.

As "hpcgadmin" user, hpcg.sh method

“hpcgadmin” can use /etc/init.d/hpcg script. At the end of the day, /etc/init.d/hpcg script calls /opt/hpcg/core/bin/hpcg.sh script provided by HPC Gateway installation. “hpcgadmin” can also use directly the hpcg.sh script.

$ hpcg.sh --help
usage: hpcg.sh --signal [start|stop|restart|status|info] --list [all,php-cgi,mongo,jetty,cluster] --comment comment [--help]


  • -s option can be used instead of –signal option
  • -l option can be used instead of –list option
  • -c option can be used instead of –comment option

The list can be all, a single process or a list of process (comma separated). The signal is applied to all processes given in the list. When a stop signal is requested (stop, restart), it is mandatory to set a comment to explain why the stop has been requested/needed.

The start/stop commands (and the reason) are logged in /opt/hpcg/logs/admin/startstop.log.

Log files are located (by default) in /opt/hpcg/logs. Whenever support is requested, send log directories to the support staff.

Log admin

${HPCG_LOGS}/logs/admin contains logs for:

  • installation/updates
  • cron processes
  • server start and stop

Log mongo

${HPCG_LOGS}/logs/mongo contains logs for mongo

Log jetty

${HPCG_LOGS}/logs/jetty contains logs for jetty

Log clusters

${HPCG_LOGS}/logs/clusters contains logs for cluster agents

If followed as part of the training programme return to topics page.