bwlang / galaxy-admin-utils

Utility scripts for working with local instances of Penn State's Galaxy platform http://wiki.galaxyproject.org/

galaxy-admin-utils

Utility scripts to help with managing local production instances of Galaxy.

Quick deployment of local Galaxy instances

• deploy_galaxy.sh: automatically create a basic local instance of Galaxy

  Usage: deploy_galaxy.sh [ OPTIONS ] DIR

  Creates a new directory DIR and then sets up a Python virtualenv, clones and configures a copy of galaxy-dist, sets up a subdirectory local_tools pointed to by local_tool_conf.xml, and creates a start_galaxy.sh script to launch the new local instance.

  Options:

  --port PORT: use PORT rather than default (8080) --admin_users EMAIL[,EMAIL...]: set admin user email addresses --release TAG: update to release TAG

• add_tool.sh: automatically add tool files into local instance from deploy_galaxy.sh

  Usage: add_tool.sh DIR TOOL.xml [ TOOL_WRAPPER ]

  Simple script to add a Galaxy tool to a local instance previously created by deploy_galaxy.sh. Copies TOOL.xml and optional TOOL_WRAPPER file to subdirectory DIR/local_tools/TOOL/ and adds a reference in the local_tools_conf.xml file. The tool should then be accessible on restarting the local instance.

Backing up/copying Galaxy data and codebase

• backup_galaxy.sh: dump SQL database and rsync database files and codebase for a local Galaxy instance

  Usage: backup_galaxy.sh [--dry-run] GALAXY_DIR [ BACKUP_DIR ]

  Reads information from universe_wsgi.ini file in GALAXY_DIR and generates a dump of the SQL database, plus a "mirroring" rsync of the database files directory and the Galaxy codebase.

  Creates the following directory structure under BACKUP_DIR:

  logs/ Logs from rsyncing the files code/ Mirror of galaxy-dist, local_tools, shed_tools etc files/ Mirror of Galaxy's database/files directory sql/ Timestamped SQL dumps from Galaxy's database

  If BACKUP_DIR is not specified then it defaults to the current working directory.

  If --dry-run is specified then the directory structure is created and the SQL dump and rsync commands are constructed but not executed.

Monitoring and controlling Galaxy processes in a multi-server set-up

• galaxy_status.sh: monitor status of Galaxy server processes

  Usage: galaxy_status.sh

  Must be run from the location were the Galaxy .pid and .log files have been written.

  Reports the status of each server process that has a PID file, either Running (server is alive but not yet serving content), Active (server is available to serve content), or Not running (server is no longer alive).

• run-server.sh: interact with a single Galaxy server process

  Usage: run-server.sh SERVER args...

  Must be executed from the directory where the run.sh and universe_wsgi.ini files are located

  Can be used to interact with a single Galaxy server process in a load-balanced set-up where there are multiple co-operating servers e.g. web0, web1, manager, handler0 etc

  For example to stop handler1 do

  $ sh run-server.sh web0 --stop-daemon

• rolling-restart.sh: restart Galaxy processes in a "rolling" fashion

  Usage: rolling-restart.sh

  Must be executed from the directory where the run.sh and universe_wsgi.ini files are located

  Stops and starts each of the servers listed in universe_wsgi.ini in turn (except for "manager", which must be restarted explicitly using e.g. run-server.sh), waiting for each one to come back online before restarting the next.

  For set ups with multiple handlers and web servers this should mean that Galaxy remains available to end users throughout the restart process.