Java stack (Jira and Confluence) administration

Basic info

Confluence and Jira run in standalone flavour and both have their own Tomcats packaged.
Both are installed on /srv partition (/srv/www/confluence and /srv/www/jira), so disk space issues shouldn't be a problem.
Both run as their own users. Please do not run them as root!

Managing services

A few useful commands (we're using deamontools to make sure Confluence and Jira run). If Jira or Confluence dies, deamontools are responsible for bringing it up again.

<service> = confluence_standalone or jira_standalone, respectively

sudo svstat /etc/service/<service> - shows up/down status
sudo svc -d /etc/service/<service> - stop the service
sudo svc -u /etc/service/<service> - starts the service

sudo mltail <service> - shows last lines from console log
sudo mltail -f <service> - like above, constantly, equivalent of tail -f
sudo mlcat <service> - dumps last 1MB of logs

Above is a preferred way of looking at logs, however, if you're more of a tail -f person, go to:

for Confluence Tomcat logs: /srv/www/confluence/atlassian-confluence-4.1.3/logs
for Confluence own logs: /srv/www/confluence/confluence-home/logs
for Jira Tomcat logs: /srv/www/jira/atlassian-jira-4.4.4-standalone/logs
for Jira own logs: /srv/www/jira/jira-home/log

Potential memory issues, the simple way

Both Jira and Confluence need quite a lot of memory. At the moment, they have settings which should be optimal, but in case of problems (with other apps or Java apps /watch out for OutOfMemory exception/ on sushi not having enough memory) settings should be tweaked. This is done in:

/srv/www/jira/atlassian-jira-4.4.4-standalone/bin/setenv.sh
(in the first line, which right now looks like this: JAVA_OPTS="-Xms256m -Xmx512m -XX:MaxPermSize=384m -XX:+UseCompressedOops $JAVA_OPTS -Djava.awt.headless=true ")
/srv/www/confluence/atlassian-confluence-4.1.3/bin/setenv.sh
(all over the file, setting environment variables /current values listed as well/: JVM_MINIMUM_MEMORY="64m" , JVM_MAXIMUM_MEMORY="320m", JIRA_MAX_PERM_SIZE="256m")

Three most important

-Xms / JVM_MINIMUM_MEMORY - size of memory used at the start, shouldn't need tweaking unless any of the apps won't start

-Xmx / JVM_MAXIMUM_MEMORY - maximum amount of memory that can be taken for storing data

-XX:MaxPermSize / JIRA_MAX_PERM_SIZE - maximum amount of memory for compiled classes, templates, etc

Maximum amount of memory used up by single application can be estimated doing following calculation: (Xmx + MaxPermSize) * 1.5

If you need to, please tweak memory settings with caution, read a bit of Atlassian wiki and stuff like that - don't make random changes.

Confluence: Copy form Production (Sushi) to Staging (Biryani)

To transfer the wiki-related files (executables and data) from sushi to biryani:

root@biryani$ adduser confluence
  -> Home dir will have a different path on biryani than sushi due to
     disk-space allocation issues, but we'll leave symlink as fallback
root@biryani$ ln -s /home/confluence /srv/www/confluence

root@sushi$ cd /srv/www/confluence
root@sushi$ find ! -user confluence
  -> Just in case there was something funny; looks OK

root@sushi$ rsync -va --progress /srv/www/confluence/./ confluence@biryani:/srv/www/confluence/./

To transfer the wiki-related databases from sushi to biryani:

root@sushi$ grep hibernate /srv/www/confluence/confluence-home/confluence.cfg.xml
  -> Determine old mysql db/user/pass, e.g. "confluence_new"/"confluence_new"/"pass"

root@biryani$ mysql -u root -p
  -> Create identical mysql db/user/pass
  create database confluence_new;
  GRANT ALL PRIVILEGES ON confluence_new.* TO 'confluence_new'@'localhost' IDENTIFIED BY 'pass';

root@sushi$ mysqldump -u root -p confluence_new \
  | bzip2 > /srv/www/confluence/YYYY-MM-DD-HH-MM-confluence_new.sql.bz2
root@sushi$ chmod og= /srv/www/confluence/YYYY-MM-DD-HH-MM-confluence_new.sql.bz2
root@sushi$ rsync -va --progress /srv/www/confluence/./ confluence@biryani:/srv/www/confluence/./

root@biryani$ bzcat ~confluence/YYYY-MM-DD-HH-MM-confluence_new.sql.bz2 \
  | mysql -u root -p confluence_new

Finally, you can start the new service on biryani.

Note: The new instance can (at time of writing) be addressed as "http://wiki.dev.civicrm.org:8090/confluence/". Some pages (such as the administrator's "Plugin" screen) may break due to the URL change – fix this by navigating to the Confluence administration area fixing "Configuration => General Configuration => Server Base Url".

Confluence: Upgrade

For full instructions, see http://confluence.atlassian.com/display/DOC/Upgrading+Confluence
As a best-practice, one should first copy Confluence from production (sushi) to staging (biryani) and perform an upgrade on staging.
Navigate to http://www.atlassian.com/software/confluence/download?os=linux and find the download URL for the Linux 64-bit installer
Download and execute the installer on the server, e.g.
- mkdir ~/download
- cd ~/download
- wget 'http://www.atlassian.com/software/confluence/downloads/binary/atlassian-confluence-4.2.2-x64.bin'
- ./atlassian-confluence-4.2.2-x64.bin
After running the upgrade, the Confluence web path may have changed from "/confluence" to "/". This can be fixed by editing "server.xml" and to the <Context> tag, e.g.
- <Context path="/confluence" ...>