FOOD FOR THOUGHT Confused About Distros?

[mattseymour]

Having been a long time user of PBXinaFlash always on an intel platform I want to build a simple pbx for a friend. It will be very low call volume and about four extensions, so the Raspberry Pi 2 is an ideal server platform.

I have, however, got confused about where to start with this. I've always found the Piaf/Nerdvittles/incrediblePBX links relatively confusing. It's hard to know whether you're looking at something that's current or three years old, sometimes links are updated and sometimes not. Just an observation...

But anyway, I installed the latest incrediblePBX for raspberry PI with asterisk gui. Great, although I got stuck with a few things and wanted to see how FreePBX performed on the R PI2 platform.

So I downloaded the latest raspbx which includes FreePBX 12... Lovely. Then I ran the IncrediblePBX install script which seems to have downgraded FreePBX to 2.11 and IPtables aren't working.

I guess I'm just a bit puzzled about how to find out what is the definitive latest version of stuff and what works properly with what.

If someone could point me in the right direction, I'd appreciate it.

 

[Jay Deal]

You're not alone, I find the variety of choices presently overwhelming. A chart helps but I can only imagine what it must be like for the newly introduced. I don't know how Ward and company does it, there must be like 40 different current versions of PIAF and IPBX between the different OS's and platforms. And most incredibly, they all seem to work for the most part.

 

[wardmundy]

Ever visited Nerd Vittles? Take a look at the Getting Started Guide. If you're more into the One Trick Pony, pick one and enjoy the ride!

And, yes, the idea is that all of the should work.

 

[mattseymour]

Yeah, I read all of your articles on Nerdvittles.com Must say I find the navigation of your blog site fairly baffling too

 

[wardmundy]

Maybe you're not looking in the right place. Not sure how to make the selection process much simpler. With all due respect, perhaps your friend should cut his teeth on an appliance until both of you are comfortable with Asterisk and the various platforms that are available.

 

[mattseymour]

Well... For starters... Much as I really appreciate the getting started table you've put together, it suggests that IncrediblePBX 12 will work with Raspbian but as far as I can see it doesn't. It needs Ubuntu 14.04.

I think my problems come from having information in too many places and not knowing what's up to date.

This all comes across as being critical, sorry about that. It really isn't my intention. I've got on brilliantly with IncrediblePBX and love the work done by all here. I've been using Piaf over the last few years, so I'm happy with Asterisk and know what I'm doing with networking etc.

I think I'll go with RasPBX for now as it will do everything we need (and more). It'll all be secured by hardware firewall so should be ok without the IncrediblePBX security extras (which I really do like btw).

 

[mattseymour]

Incidentally the js-200fx looks the perfect product for the job I have in mind, unfortunately they're not available in the UK, so the Pi works out much cheaper.

 

[synack]

Also, not to be critical, but myself and a coworker were just talking about this the other day. While we all love choice, more choices makes the task of choosing the right recipe much more difficult. Add dependencies for this and that, and I can imagine it's a gargantuan task for the uninitiated.

I would like to see (and possibly assist with if necessary) a platform independent system where you choose your backend, frontend and features, and click "go" so to speak. For me, that would truly be incredible.

 

[wardmundy]

That's a pipe dream, I'm afraid. We do sort out dependencies in all of our installers. But, remember, we are dependent upon open source products with literally thousands of developers. Sometimes somebody breaks stuff. The landscape changes almost daily... except for RasPBX which hasn't changed much in the last 15 months. Not being critical but why anyone would pick a 15-month-old VoIP platform is a real head-scratcher. But... to each his own!

We get rafts of about Asterisk 12 or FreePBX 12 not working as expected. Can you spell E-X-P-E-R-I-M-E-N-T-A-L? If you want a stable platform, stick with an LTS release of Asterisk (HINT: 11) and not one that goes end-of-life at the end of this year. If you want a stable version of FreePBX with no surprises, stick with 11 for the time being. If you want to be a pioneer, by all means have at it. But don't whine when it blows up or does something unexpected.

We have numerous VoIP solutions that are functional on many different platforms, and we have documentation and a roadmap that will help folks choose the one that's right to meet their particular needs. HINT: Things don't run the same on a $35 Raspberry Pi as they do on a $3,500 server. Guess why?

If you can't figure something out with the available tutorials and this forum, hire a consultant. That's what they get paid for. If that's too rich for your blood, spend a nickel a day (that's right, 5 cents) and test several different versions on Digital Ocean until you find one you're comfortable with. You even get a $10 credit to play for free for two full months if you click on our referral link. If that's too hard, buy an appliance. We write about them all the time, and there are many to choose from. This week there's one for under $30.

As for Nerd Vittles being too complicated and difficult to navigate and too hard to decipher what's up to date, and blah, blah, blah...

• Newer articles are more up to date than older ones. We don't typically revise ancient articles!
• Articles are arranged chronologically, newest first. VoIP projects age in dog-years not human-years.
• Articles include their published date and updated date at the end of each article.
• A list of 250+ articles in chronological order with hotlinks accompanies each article. Below 50 = OLD!
• A roadmap of current releases on numerous platforms is available. Hotlinks to tutorials included.
• A tag cloud is available to find articles by subject. Results arranged chronologically.
• A search option is available in the right column to find articles using key words.
• Hot links in current articles always refer readers to related topics.
• When all else fails, Google it -> keywords to find site:nerdvittles.com or keywords to find site:pbxinaflash.com
• If none of this helps, look in the mirror. You may not be smarter than a 5th grader.

 

[synack]

Whoa there cowboy. This is not a critique of the wonderful products you and your team put out. I for one am very happy with the work you guys do.
I don't expect handouts and will gladly roll up my sleeves to help out in anyway I can.
All the choices you guys are providing has brought around the inevitable fragmentation. It might be clear to YOU but many of us are scratching our heads trying to figure out what's what.
I'm only suggesting that we broach the subject with some discussion on how we can reduce the fragmentation while keeping choice, which is so important to many of us.
What I describe above may be Utopian, but what is a dream without a goal? Even if we only achieve a fraction, would it not be better than nothing at all?

 

[wardmundy]

synack: Sorry, my comments were not aimed at you. You just happened to be the last to post before my tirade.

We're all for improvement. Just let us know what would be helpful and how to achieve it.

 

[synack]

Just throwing some ideas about here. Definitely open for discussion, and smack me if i'm totally off base.

For a start, how about teams of individuals dedicated to providing "base" platform images. These base images must be able to run a single PiaF/Incredible installer made for all platforms to be considered "supported".

Each platform would provide a set of scripts(API?) that would be used by the installer. This API would also describe the platform capabilities to the install script, which would be transformed into run-time install option menus.
The API could either be provided pre-installed on an image, or installed separately depending on what makes the most sense to everyone.
I envision the install script would decide it needed "x" applications and the API would be responsible to install them and their dependencies.

I firmly believe the PiaF/IPBX installer should remain high level, and the platform API would take care of the nitty gritty for that platform.
Supporting new platforms should be as simple as porting the API to that platform.

Since the chosen front end would be dependent on what was chosen for back end, the installer would need to dynamically update it's sub menu's accordingly.

The end user experience would be: Download appropriate image for hardware platform and distro of choice. Download/run the single "up-to-date" PiaF/Incredible installer(s) and/or API, select choices available.

I know this is extremely high level, but I'm hoping it's a catalyst for discussion.
I suspect this should be continued in another thread in another category.
Dev? I have a dream?

I'm going to take a stab at an alpha version for BBB with some version of incredible pbx 12, as that is what probably interests ME most at the moment. It will likely take a few weeks as I can only dedicate spare time right now. I would appreciate any feedback and "wishes" as it's easier to consider them at the start then to integrate later.
The first "downside" I foresee would be support for new frontend/backends as it may require a complete API update across all platforms. The danger here is obviously moving the fragmentation down to API level instead of the installer, which will get us no further ahead. Having said that, keeping the API generic enough to only abstract out the particulars of the distro/platform *should* reduce the number of API updates.

 

[wardmundy]

That's pretty close to the existing PIAF 3 model except it does not yet support Debian/Ubuntu/Raspbian. Even so, I think darmock would tell you it's a bear to maintain. Problem is that RH and Ubuntu and Linux kernel and Asterisk and FreePBX "improvements" break things in major ways sometimes. And babysitting another layer in the install process can turn into a full-time job. Not saying it can't be done but I think you'll be trading one pile of headaches for a fresh pile.

 

[darmock]

PIAF 4 model will change the paradigm slightly and have a unified installer for all supported variants of asterisk. Thus you will download a base install program and then in the setup config you will choose what flavor of asterisk you want and then it will be downloaded and installed. I have always been somewhat hesitant to build a base model and then a kitchen sink model and put it in the same installer. A LOT of people only want the base model sans IPBX. I understand the intent but...... Currently the base model is based on the Red Hat/Centos/SL/et al model which is what a lot of people want. The other debian/ubuntu/raspian/sunos variants are very much in the minority. That being said once PIAF4Base is out I can see building an ubuntu base version however most of the people that help out with the base PIAF are RH/CO/SL/et all developers.

In PIAF4 we are also going to migrate to a not every software revision of asterisk will get a autoconf file. Every time a new variant of asterisk comes out the dev team has to install the new version and check that all of the options that are configured in the make menuconfig actually will run etc. This gets to be a pain as each variant of asterisk and 32/64 bit versions have to be installed and separate autoconf files need to be created. (We learned long ago that there ARE differences at times between 32/64 for many reasons)

A simple base/ipbx installer is an admiral goal but I am not sure of how this could be accomplished easily so it bears some more thought.


I am going to put PIAF3 development to a halt in the near future after we release the white version of the product with the base system locked down by iptables. Then it will be finish up paif4 and get it out for testing



Tom

 

[synack]

That's very much what I was hoping for, but coming at it from a slightly different direction.

A single monolithic installer would be good, but supporting multiple OS and hardware platforms might get tricky as you guys suspect. My idea for an "API" was more about isolating as much of the base platform out of the installer as possible. That way, an enterprising individual can build platform support outside of main package. I'm still in the preplanning stages of it. I've looked at the current installer and am cherry picking routines and variables here and there for inclusion. I'm sure the alpha will be very much bare, but with a reusable framework.
Any chance of a pre-release 4 so i'm not building on a soon to be deprecated package? Or should I just wait and re-evaluate the necessity once the new version comes out?
Also, i'm thinking of adding gentoo support along with BBBubuntu. Gentoo lends itself to this and versioning much better than other distros, and well... it's my fave.

 

[shetu]

That's a pipe dream, I'm afraid........]

fire on word.

 

[prattmd2]

Just a thought .....

How about an install decision tree being available.

1. Pick hardware: VirtualBox/ Raspberry Pi 1/Raspberry Pi 2/Standalone server
2. Based on that selection, choose a compatible base distro: If you had chosen Raspberry Pi2 for example, Raspian would show up. If standalone, you would see CentOs, Ubuntu
3. Select your GUI: Incredible PBX, Asterisk GUI
etc etc

Maybe that would make which version to install more logical for some

Just my humble opinion

 

[wardmundy]

Great idea. I'll work on it.

 

[wardmundy]

prattmd2: First cut at an Asterisk decision tree. Suggestions are more than welcome! Click on image to begin...

 

[vic555]

This is nirvana for the technically challenged.