We’ve had two occasions so far where the Astrisk Rest Interface (ARI) which is usually contacted at “http://subdomain.domain.com:8088/ari/channels” stops responding after a server reboot. In the Asterisk CLI Module, the entering the “ari show status” command says that it is not a valid command, when usually it reports the status. We’ve had to roll back snapshots pretty far on the Amazon instance to get to a state where it is working again.
We’re having trouble isolating what is breaking ARI, because it seems like we can run plenty of module updates and do Apply Config and the ARI continues to work. Only after a full system reboot do we see that the ARI won’t come back up.
Would anyone have any tips on how we could pin down the issue? Is there any command to force it to try to relaunch, or to see further output on what is preventing it from launching?
The only thing involving ARI that we have configured by hand is a single entry. We wanted to get the “Swagger” Asterisk page to talk to our server and even though all the guides say it will work if we set “allowed_origins=*”, that command did not seem to function as intended, so we manually named the domain in ari_additional_custom:
But, as it turns out, that conf edit doesn’t take effect until after a reboot I think. It seems like the ARI will crash if we have the conflicting commands both trying to load. The =* wildcard and the =ari.asterisk.org can’t coexist.
Now we removed the ari.asterisk.org entry, rebooted, and the =* now does work when accessing the Asterisk Swagger site. I guess we got mixed up because they needed reboots to take effect, or something.