* had to manually flush db every time I changed callflow or removed/added voicemail
> This is a design decision that I agree is a defect. The cache keeps the callflow in memory to avoid hitting the DB for numbers that are used a lot, but the cache should be flushed when a specific number is changed. We will be coding this up this week.
It is not, in my eyes, an indication that the system is broken it's just an annoying design because if you change a callflow you expect it to work the new way immediately, but in fact it can take time if the number is in use as that resets the timer for the
cache expiration. And of course the first thing people seem to do when playing with this is setting up a number and trying all the functions out.
* had to manually start services after server reboot
> My expectation is that you are starting services yourself after reboot. We have discussed this in the office a bunch and have mixed feelings on it. We have settled on the following, which we'll work on implementing next week:
> We will start services automatically after 60 seconds on the bootup of a server. This is done to allow people who reboot during maintenance but DO NOT want the server to rejoin the cluster to have some time to go in and stop the auto-starting of services
themselves. The details on how we will achieve this remain TBD but that is ultimately why your services do not auto-boot – we don't want a micro server to reboot and immediately start thrashing it's database or otherwise if that's the reason it crashed in
the first place.
This is a bug only with RabbitMQ. RabbitMQ must be started before the other whistle apps. This is a serious bug actually which will be the focus of our attention immediately on Monday – what's really happening is whistle is trying to reconnect to AMQP
any time it loses connectivity, but for some reason it's failing to ever actually do so. Check out your logs in /var/log/local0.log
> Most of the items you list above are design changes we'll be working on next week. Nothing is "broken" or unusable, though, so for the device not registering or making calls we do need to look into your config to see why it's not working.