Hi!
Finally a working, production ready FreePBX Docker container with Asterisk 16. It comes in 2 flavors, with and without Hashicorp Vault integration.
Tested both and they work fine, check it out:
Asterisk 16 went end of life last year[1]. It no longer receives security fixes.
The point is not Asterisk version, but the hard work made on that project, which is publicly available
Sure, but people should be aware of the support level of the version of Asterisk in use.
Same for php7.4 EOL 12/2022
yes if you did take a look at repository, you would notice that it’s 1 year old. It’s easy to upgrade starting from that.
PS: You should be aware that a lot of people out there use obsolete Asterisk versions, just take a look at topics around here.
Would someone please explain the love affair with Docker? If you really want I can post a qcow of FBPX17/Asterisk20 and you can boot it on KVM.
1 Like
We run almost all production applications in Docker because the environment is reproducible and we can get another instance up and running very quickly, it’s easy to make custom images and push it to our private registry. Over time we’re moving to Podman because I like the rootless mode.
Ironically, Asterisk is the only application that we run natively. Even though we’re on an old version, it works really well and is rock solid.
Oh sure it’s completely understandable that if you are DEVELOPING software APPS you would like docker or other container solutions, that’s why after all that Ubuntu is now delivering Chrome and many other apps in containers.
But from the users POV, they don’t benefit one bit from a containerized app vs a “native” app - in fact the app runs slower because of the overhead of the container.
The Asterisk devs don’t seem to feel the same way you do about containerization of Asterisk. Or, for that matter, virtualization of it via KVM or HyperV or any other common virtualization hypervisor otherwise they would deliver VM files. Even though most of the world today runs virtualization very heavily. And I think most users view a phone system as something you setup once - then never mess with again. They don’t care what it runs on or what it requires so long as it works. They also don’t tend to view it as just one application among many on a server, instead it’s more of “put that there in a gold box and never touch it again” And, I think there’s good reasons for that, BTW.
That more lends itself to a total compartmentalized solution than a lightly containerized solution.
I need to clarify some misconceptions. Containers are not virtual machines (at least on Linux). They share the same Kernel as the host OS using the Linux cgroupv2 API and any processing overhead is negligible, if any. There may be some overhead on the network interfaces and mounted directories, but probably not much.
Running asterisk in a container can cause issues related to double-NAT and you need forward ports to the container from the main interface. The interface names and IP addresses might change when you restart docker, and the headaches are endless. Of course these can all be dealt with, but I’m not sure if the extra effort is even worth it, so native it is for me.
I totally understand why the Asterisk developers don’t want to deal with containers, because this creates a nightmare scenario where suddenly you have thousands of people with support queries saying they have one-way audio, and supporting this is not an efficient use of developers’ time.
1 Like
This topic was automatically closed 30 days after the last reply. New replies are no longer allowed.