FreePBX and CentOS base: CentOS is dead from now on


(Dave Burgess) #21

If only there was an operating system that ran on everything from PCs to the Dreamcast… NetBSD supports over 50 different hardware platforms including ARM.

Just sayin’


#22

People are already fretting about a move from CentOS to another Linux.


(Jared Busch) #23

>insert explicitives<


(Michael Hickey) #24

What can anybody do if CentOS is now dead? Fretting about it won’t solve the problem, although it is certainly clear to me why people would be upset by this move. Besides, it might be good in the long run to move on to a different distro. One that won’t become an apparent victim of ‘embrace, extend, and extinguish’ by the likes of Red Hat. To be honest, I’m with cynjut in advocating for the OS base to be one of the BSDs, though I’m more familiar with using FreeBSD than NetBSD myself. Also, asterisk packages exist for FreeBSD, too. If it has to be Linux, then Debian seems to be the obvious choice to go with.


(Jared Busch) #25

Well funny you mention FreeBSD. Some people with the FreeBSD foundation were on the Open Source Lounge meetup this evening. It was a very good discussion.


#26

I can think of a few technical reasons to stick with Linux:

  • dahdi
  • iptables/nf

and a few non-technical reasons

I mean, before people start promoting something, they should at least have a sense that it works, right?


(Jared Busch) #27

Asterisk13, 16, 18 is native in FreeBSD repos, according to the guys on the meeting last night.

Not that I am promoting that as a path forward.

Honestly, I would go with Fedora if it was my choice. But OMG I can hear the screams of others when I just think that.

Debian is likely the best path forward in reality.


#28

My question would be are there any reasons to justify a change to *BSD?

It makes no sense to move FreePBX distro to *BSD. Too many changes that would irritate folks, less market acceptance, and at the end of the day I can’t think of any end-user benefit.

If Sangoma needed to make proprietary changes to the core OS for their appliances, then maybe a BSD licensed base would be justified, but there is nothing I can think of where that would apply.

That said, I’m all for making any changes necessary to facilitate installing non-distro FreePBX on *BSD. I don’t think there would be any insurmountable issues there.

FreeBSD was my *nix OS of choice for years. I ran virtually everything in containers waaaaay before containers were cool, and FreeBSD jails (often running a Linux userland) were the core of that for a long time.


#29

Assuming the kernel modules (dahdi, nf) I mentioned aren’t a concern then someone should do a *BSD implementation of FreePBX and document it on the wiki so that other folks can try it out. The only way to know whether there’s any real benefit is for people to explore it.

Editing to add: the last time I used FreeBSD was about 15 years ago. It felt clunky and had an “old UNIX” smell to it. Since I haven’t had any reason to use it since, my lingering impression is outdated. I’m sure that modern FreeBSD is much nicer, but outside of the possibilities arising from this discussion, I can’t think of any other reason to look into it again.


#30

Dahdi might be, what percentage of installs is it needed these days?

Iptables/nf is irrelevant. The open source firewall module doesn’t even work non-distro Linux at the moment.

I think the biggest issue, as implied on the call, would be generalizing all the assumed linux-isms and hard paths scattered through the code.


(Michael Hickey) #31

DAHDI drivers, utilities, etc, also exist as packages in FreeBSD. So does fail2ban. Also, FreeBSD has three different kernel level firewalls to choose from. To me, that at least shows some potential, provided anyone has an interest in doing a port of FreePBX to one of the BSDs.

Realistically though, going with a different Linux distro as a base for a FreePBX distro, would be the easier/wiser choice. The loss of CentOS is already enough of a disruption as it is. But, one can still dream.

Edit: I was a Linux user fifteen years ago, but I have to admit that FreeBSD has come a long way since then. I can run FreePBX in a bhyve VM, have it installed onto a ZFS mirrored volume, and run daily snapshots on my entire storage pool for just in case. It just works. (for me, at least)


#32

Let’s not conflate ‘FreePBX’ with ‘The FreePBX Distro’

Without doubt a functional FreePBX would probably be run on most anything vaguely unix based , presumable even WSL , as it just needs Asterisk, a webserver, PHP and a Database server (probably only MySQL currently) which will give you a basic functional FreePBX GUI.

Realistically we need to add Firewalling/IDS probably a rudimentary mail service and possibly other utility.

Here’s the rub . . .

The Value added things like provisioning and commercial modules are the cash cows but some need ‘privilege elevations’ and quite reasonably, obfuscation, which are scary for some, We see movement from Zen to IronCube for obfuscation which presumably brings non OS specific hardware and current PHP within range. So The CentOS thing currently pertinent to those that who create “Distros” so being as catholic and open minded as possible will benefit all. However sticking with IronCube probably currently excludes non ‘Linux’ things


#33

Actually ioncube has FreeBSD loader. Realistically, I doubt we would ever see something like sysadmin fully supported - too many real system level differences to address.

A mini-sysadmin stub to handle a couple of basic things like hooks and commercial modules would be doable - which I also hope is considered for non-distro Linux installs.


(Dirk2358) #34

Changing to xBSD in my opinion would be the best way to drop lots of customers / users. The great benefit of CentOS was, that it was pretty easy to switch to RHEL to get professional support. That’s important for companies.

With CentOS (Linux), it is possible to easily use most of commodity hardware and add much more functionality like WLAN APs with a big variety of fully supported WLAN chips.
It’s much more easy to provide own modifications to asterisk by having test installations even on already existing desktop machines.

I think OpenSUSE Leap would be a good solution to switch to: it provides at least 3.5 years stable support and the base is SLES (besides the kernel). And it’s even pretty easy to switch between major releases. SUSE provides a good build service to add and maintain own repositories.


#35

Thank you, I now stand better informed :wink:

(however being told/requested to radically change religion has historically been , shall we say ‘not without pain’ )


#36

I’m not telling anyone to change religion. I don’t think there would be any user benefit and mostly downsides to changing the distro from Linux.

I don’t have any plans to try porting my Debian/Ubuntu install script.

But, I’m all for anyone that wants to use Free/Open/Net/BSD being able to do so if the FreeBSD guys can start submitting PRs.


(Michael Hickey) #37

I have a small, low powered computer that would be perfect for a SOHO environment. The problem with it, is that the kernel in CentOS 7 refuses to work with it. Basically, upon trying to boot it for the first time, it becomes slow as molasses until it stops working altogether. If CentOS 7 refuses to install on it, clearly the FreePBX distro will not work on it, neither. However, Debian 9, and Debian 10 work fine with it. I can then grab all the packages that I need to put Asterisk and FreePBX onto it, and I’m off to the races. However, commercial modules are simply a no-go. Sure, it could be argued that I don’t really need the commercial modules, but the features included with the SysAdmin Pro module, for example, are something I’m willing to pay for, in order to help support the project.

Maybe I’m missing something, but it seems to me that the ability to run FreePBX, on either the FreePBX Distro, or on a distro of my own choosing, AND having the option of activating for commercial modules, could only be a good thing. It would certainly mitigate the impact of having a distro base being killed off, from seriously affecting projects that ultimately rely on it.

Then again, having only one official distro to have to support, would not be an insignificant concern for a company, neither.

Edit: To jerrm: It’s not too hard to install the FreePBX distro in a bhyve VM, running on FreeBSD. You can also mask a NIC from the host, and with PCI passthrough, give it to the VM for the Linux kernel there to handle, provided the host hardware supports VT-d, of course. So, if the FreePBX distro doesn’t switch its base to FreeBSD, I’ll be content to let the more religious among us to do the conversion, as what I have works fine as-is.


(Dirk2358) #38

Gregory Kurtzer, the founder of the CentOS project, started a new project. The goal is to continue CentOS under the new name Rocky Linux. That’s most probably the way to go if Rocky Linux succeeds.


(Tony Lewis - https://bit.ly/2SbDAyc) #39

Here is another announcement of a fork of RHEL. https://blog.cloudlinux.com/announcing-open-sourced-community-driven-rhel-fork-by-cloudlinux


(Michael Hickey) #40

Two RHEL forks? Maybe a move to one of the forks might work.