I have 2 x OH2.1 instances connected via mosquito MQTT running on one of the Pi’s for the important stuff, and via mosquito MQTT on the other one for the high traffic stuff that I can afford to lose the odd message on.
I’ve noticed that the odd message on the lower traffic setup doesn’t get through both from one Pi to the other, and also from one Pi to an ESP8266 running ESPEasy.
I’ve just added
mosquitto.qos=1
to mqtt.cfg, and despite adding
max_queued_messages 3000
and also trying
max_inflight_messages 0
in mosquitto.conf I can’t get rid of lots of “too many queued messages” errors in the log on startup when one Pi reports the start of everything to the other one
Is there likely a cure for this, or am I better off giving the new OH2 alpha / beta MQTT binding a try?
When I reboot I usually get at least 20
Dispatching event to subscriber 'org.eclipse.smarthome.model.rule.runtime.internal.engine.RuleEngineImpl@b8c56b' takes more than 5000ms
messages in the log so maybe this is contributing to the MQTT backlog