Why does the Columbus database fail to start via command line with a "Connection timed out" warning?

When starting the Columbus services and the Columbus database component fails to start with a ''Connection timed out'' warning as shown in the below example, this might indicate an issue with the firewall settings being too restrictive; the database component is unable to establish a communication path with ICE and potentially other components on the local system.

$ /etc/init.d/columbus start

* Starting Columbus

Database: /usr/local/PerkinElmerCTG/Columbus/etc/grid/config.xml read-only No descriptor given. 
Using etc/grid/default.xml icegridnode: failure occurred in daemon: service caught unhandled Ice exception: 
Network.cpp:1226: Ice::ConnectFailedException: conn[FAILED]ed: Connection timed out

Acc: [ OK ]

Web: [ OK ]

Celery: [ OK ]

To confirm if the error is related to the firewall settings, temporarily turn off the firewall and try to start the Columbus services again. You should consult with your local IT and/or network security team before running this test.

The below example is for Columbus installed on RHEL/CentOS 6:

$ service iptables stop

$ chkconfig iptables off

$ sudo /etc/init.d/columbus restart

$ sudo /etc/init.d/columbus status

If, after turning off the firewall and after restarting the Columbus services, all Columbus services including the database component startup, then you should assess and amend the iptables configuration.

For example, to set access for localhost you can use:

iptables -A INPUT -i lo -j ACCEPT