There are a few things to check if your THREDDS Data Server is acting
flaky:
1) Have you blocked web crawlers? Something like this should work:
$ cd <tomcat>/webapps/ROOT
$ cat robots.txt
User-agent: *
Disallow: /
2) Have you specified headless server (for systems where X is not
installed) and max memory for Tomcat?
For a 32 bit server, try adding this line at the beginning of <tomcat>/
bin/catalina.sh
JAVA_OPTS="-server -Djava.awt.headless=true -Xms512m -Xmx1024m"
For a 64 bit server with more memory, you could put the max higher
JAVA_OPTS="-server -Djava.awt.headless=true -Xms512m -Xmx2048m"
3) Get the latest THREDDS Data Server distribution from:
ftp://ftp.unidata.ucar.edu/pub/thredds/4.0/thredds.war
and redeploy. Several aggregation bugs that have plagued us for
months were just fixed last weekend.
Keep in mind that I'm not a Tomcat expert, so if others want to weigh
in on these suggestions or provide more info, please feel free!
-Rich