Operating Confluence

If there is a problem with wikis, obviously this document would not be accessible to you from here. Please copy this content to a location where you store support information. 

Starting and Stopping the Service

• The service can be safely restarted with /etc/init.d/web restart. This restarts apache, tomcat, and other dependent components.
• To restart just the servlet container (just confluence), run /opt/sash-server/bin/sashctl confluence restart
• Wikis uses a multi-container tomcat derivative called sashserver, /opt/sash-server. Therefore, you will see multiple java processes running on a system. You can identify confluence doing an "lsof" on the java or jsvc processes to see which is grabbing up confluence jars in the path to the confluence servlet container (somewhere under /opt/sash-server/servers/confluence).
• Confluence is a resource-intensive application. It can take several minutes for the service to start.

Dealing with Known Issues

Confluence's Internal CRON

There is a known performance issue with an internal Confluence feature called BackupJob. Our confluence is too big to run this successfully. It should be disabled. You can use some of these instructions with other background processes if they evidence problems in the future.

• If this is the problem, the catalina.out will contain errors pertaining to "quartz" and "BackupJob" in the same error statement. /opt/sash-server/servers/confluence/logs/catalina.out
• To stop the current process you have to restart the whole confluence container (see above).
• Since this job should be disabled, check to see if it is disable in the web-management console for Confluence.
  • MIT and OPS members should all have admin privileges for wikis-stage.mit.edu and wikis.mit.edu.
  • Log into the wikis system in question, find the "administration" link at the top of the page.
  • From the Admin console, navigate to the Daily Backup Admin panel in the left-hand navigation.
  • Enter "Edit" mode, and set "Backup Attachments" to OFF.