So it became apparent that, while sitting on a GlusterFS on two of my nodes, Confluence was dumping 18GB of logs to catalina.out. Unfortunately, there isn't a good way to rotate that file:

https://confluence.atlassian.com/confkb/catalina-logs-are-not-rotated-or-removed-289276264.html

All the while writing to the GlusterFS, which by itself network copies this to the secondary host, atlas01:

atlas02 # du -sh logs
18G     logs

To fix this, created a folder on the host OS called /confluence-logs/logs, then copied the /atlas/atlassian/confluence/logs folder to the OS folder.  Then linked them up and changed permissions:

# atlas02 # /atlas/atlassian/confluence # ln -s /confluence-logs/ logs
# chown -h confluence.confluence logs
# ls -atlrid logs
11784356602385662682 lrwxrwxrwx. 1 confluence confluence 17 Nov  4 20:14 logs -> /confluence-logs/
#

Then started up confluence once more.  Don't forget to do this on the second note too!  The drawback to this, is that when nodes failover, the secondary host doesn't really have a copy of the logs anymore, since it won't be shared at that point.

NOTE:

Additional SQL Logging can be enabled via the UI.

Sum It Up!

Enjoy faster performance!  

Cheers,