Generating heap/stack dumps for debugging a slow running OpsCenter Daemon

When diagnosing performance issues with the OpsCenter daemon process, stack and heap dumps can be very helpful. If you run into a situation where the OpsCenter process seems to be responding very slowly, the best course of action is to generate some debugging files and attach them to a support ticket so that the DataStax team can investigate. Both a heap dump and a stack dump of the running OpsCenter process will greatly help identify the issues.

The first step in generating these files is to locate the pid file of the OpsCenter process. In a packaged installation (deb/rpm), the pid file should be located at:

/var/run/opscenter/opscenter.pid

In a tarball installation, the file should be located in the top level directory of the OpsCenter install and be named "twistd.pid".

Once you have located the pid file you can run the following commands to generate the debugging files (replacing [PATH TO PID FILE] with the actual path for your system):

sudo -u <user> jstack `cat [PATH TO PID FILE]` > opsc_stack.txt
sudo -u <user> jmap -dump:file=opsc_heap.hprof,format=b `cat [PATH TO PID FILE]`

Replace <user> with the process owner.  By default this is cassandra.  jmap will give an error if any account other than the working user (including root) attempts to connect.

This will create two files in the same directory that the commands were run in, one called "opsc_heap.hprof" and one called "opsc_stack.txt". You should attach these two files to a support ticket to allow the DataStax team to investigate the issue.