Issue

Asterisk (TrixBox) server is not functioning properly. Internal calls working fine, but external incoming and outgoing time-out and fail. Looking at Asterisk Command Line Interface (CLI) output (connect with asterisk -r) noticed unusual behaviour:

Normally when incoming call is received, Asterisk outputs number of executed actions into CLI and receiving phone starts ringing in less than a second. On this occasion, however, Asterisk was stopping at following line and hanging for about 7 seconds:

VERBOSE[29592] logger.c: — Launched AGI Script /var/lib/asterisk/agi-bin/recordingcheck

Then, if caller was still on a phone it would go to:

VERBOSE[29602] logger.c: — Launched AGI Script /var/lib/asterisk/agi-bin/dialparties.agi

It would be stuck there for a few more seconds and eventually call would be dropped. In fact occasionally receiving handset would actually ring (probably long after caller has hanged the phone).

Cause

Couldn’t find anything abnormal on actual Asterisk server. Everything, apart from the described issue,  was running normally. Then started looking at company’s underlying network infrastructure and noticed some seemingly unrelated issues. At more closer look I found that one of the domain controllers was down. It happened to be a primary DNS server for the local network.

It appeared that issue was caused by this primary DNS server being down. Although secondary DNS server was available (and Asterisk was configured to use both), missing primary server was not something Asterisk could deal with.

Solution

As offline DNS server had actually developed some hardware issues, I replaced Primary DNS using TrixBox web GUI (System > Network) with one of the working DNS server’s IP. This immediately resolved the problem, and was an acceptable solution until failed DC was brought back to life.

Trixbox v2.6.2.2