Segfault in Asterisk, Question on High-Availability during fail over

While doing a user acceptance test with three simultaneous calls in a few different queues (not a load test by any stretch of the imagination), our master server experienced a segfault and failed over as expected to our slave node.

However when the slave comes online and restarts Asterisk, it also restarts iptables which prevents our queue monitoring suite (QueueMetrics) from working properly. It also restarts isymphony which we do not use.

We’ve opened a Jira issue with Digium for the segfault itself, but have two critical questions re: FreePBX:

How can we specify what services are started / not started during failover?

Digium is requesting we recompile Asterisk 13 with out opt flags so they can get a more readable backtrace. Will this break FreePBX?

You should not be using Astrisk 13 in production. It is not ready for production and we warn of this at install time. Please use Asterisk 11.

As far as controlling what services start you have no control over that. We start amportal in FreePBX which tells it what services to start. If you yum erase iSymphony and disable the iSymphony module in FreePBX then it will not start with amportal.

We’ve rolled back to 11 for the time being.

According to Digium, Asterisk 13.0 was confirmed LTS and out of beta status ~4 months ago. When will it be considered production-ready in FreePBX?

Yes Digium has marked Asterisk 13 as stable but history has taught us it takes a year plus for major Asterisk re-writes to stabilize.