Anyone know why there is no out-of-the-box package so that one could simply do an apt-get install freepbx ?
Has this been considered?
i would also like.
Greetings,
The process for adding a package to the repository is easy in theory but requires some packaging experience especially for a project as detailed and demanding as FreePBX.
The first step is to create the package and there is plenty of documentation on the web. Once you have a good *.deb that works you should contact the Ubuntu Masters Of The Universe (MOTU) which you can review your *.deb and upload your package to the Ubuntu repositories.
I have personally tried and failed to get Asterisk and FeeePBX working properly on Ubuntu but then again I never tried very hard. I find CentOS to be much more stable for a server and the ease of installing from the ISO beats a Ubuntu server install.
https://answers.launchpad.net/ubuntu/+question/38819
If you have any additional questions then let me know.
Is the demand for the Asterisk - FreePBX combination low on Ubuntu for some reason? Is Asterisk unstable on Ubuntu?
My understanding is that Ubuntu is the most widely used distribution of Linux.
I would add the package myself if I fealt comfortable enough with Ubuntu, Asterisk, and FreePBX to do it but unfortunately this is beyond my capabilities which is why I was hoping that this would be of value to others prompting someone or folks who is/are capable to add it.
I think the addition of a Debian based package especially for Ubuntu would extend the userbase and popularity for FreePBX.
I, for one, have never seen the utility of worrying about installing Asterisk/FreePBX on UBUNTU, DEBIAN, FREEBSD or anything other than the standard CENTOS release. What could be easier than grabbing an existing ISO and loading it up on a computer in an environment that has an installed userbase in the hundreds of thousands.
The real difference in the various LINUX distros is in the GUI, but you really be installing that on a production telephone server anyway.
BF
Many drivers for telephony hardware (the Digium cards, Xorcom Astribank, PRI Cards, etc.) have been developed primarily (or in some cases, exclusively) for Centos and not other Linux distributions. Do you want to write new drivers?
The biggest philosophical disagreement I have in regards to the distribution discussion is because the IT Community has recently ascended to take control of telephony servers and apps. For years, there was a telephone shop and an IT shop. The telephony guys lived by the “Five Nines” or more rule and simply did not tolerate misbehaving hardware or downtime. The IT folks were allotted more flexibility in systems and apps being down from time to time. Now that telephony and IT tend to be combined in most shops, there seems to be much less emphasis on keeping the phones working than there was in years gone by. A different operating system on a single box or cluster of boxes seems to terrify some shops, even though they don’t need to mess with anything except to backup and restore they systems on occasion.
If you have a server running a stock distribution that just works with the apps and hardware required to make the phones work, you should not have to know too much about the underlying operating system. Distros like PBX in a Flash have pretty well taken all the guess work out of it. These hardened distros allow you to place Asterisk and FreePBX in the same realm as phone systems from the big boys like Avaya or Cisco. IT shops have long made exceptions to their operating system requirements for telecom servers. Avaya Communications Manager runs on Red Hat Enterprise and many a Windows shop has Avaya servers running in them.
My whole point is that you don’t need to, nor should you, muck around with the telecom server. Modifications and modules for Asterisk or FreePBX are not normally written at the operating system level so modifications are much easier. Good distros have scripts that do all the hard work for you already built into the distro. The overwhelming majority of distros use Centos for a very good reason–it works! Multitasking other GUI based stuff in competition with call completion on the same server is not a wise decision.
I simply would not be willing to stake my paycheck on running Asterisk or FreePBX on an operating system which is not already time-proven by the masses and working in a solid, reliable fashion.
We can opine about the pros and cons, but at the end of the day support of the community is still overwhelmingly with Centos. I see Ubuntu questions pop up in forums that go mostly unanswered due to lack of community expertise. Perhaps if pioneers take it on to make a bullet-proof Asterisk/FreePBX/Ubuntu distro as good as the Centos one, then you may gain some ground. In the meantime, it is certainly not worth the headaches, heartaches and even loss of jobs that might happen if you roll out a less than solid telephone system. At least, it isn’t worth it to ME. Do you want to risk it?
Certainly for someone who wants to set up a dedicated machine just for the pbx, intents to do all pbx configuration via a GUI, and has not invested their time into experience, etc. into a particular distribution, would most likely find using an ISO on the “standard” OS – I guess that is CENTOS – may make perfect sense and be easiest.
However, user requirements differ by the diversity of people.
A reason some users might value running FreePBX with the Asterisk packages on Ubuntu may include some of the following:
- Requirements for running apps, tools, utilities, etc. unrelated to the PBX on the same box
- Investment of time, resources, etc. into the learning curve of a particular distribution (e.g., Ubuntu)
- Standardization of all tools on the same platform (OS) to streamline the maintenance (by employees and contractors)
- Requirements to use existing customized scripts, etc. on a platform that one might want to utilize for the PBX as well (and not have to port over all the scripts, etc.)
Consider: Why does one company or university or individual standardize on a Machintosh versus Windows versus Sun Unix versus RedHat versus Ubuntu versus CentOS. Users and requirements are diverse.
For someone used to and standardized on Ubuntu, for example, running apt-get install freepbx is much much much * easier * for that person than obtaining an ISO for a distribution that they are unfamiliar with, and doing any additional learning, porting of scripts, etc. on that distribution, etc.
I would believe that there must be some user demand for FreePBX on Ubuntu because:
- There are packages for all of the other major componenets related Asterisk already on Ubuntu
- Searching around, there have been a number of folks that have tried to get FreePBX running on Ubuntu apparently with significant difficulty. Some successful, some not.
When first creating this thread, I was wondering why FreePBX was the only one without a package for Ubuntu. Whether it is because it has been too complex to get running successfully on Ubuntu; Whether there just is not enough demand among Asterisk users on Ubuntu for the FreePBX GUI; Or some other reason.
For a package, would it need to be driven from the FreePBX community or from the Ubuntu community?
Again…I point out…A telephony server should be dedicated to just that…telephony.
In order to save on assets, the GUI should not be installed on a telephony server. Indeed, the telephony server is usually headless with administration being done remotely via a browser or SSH.
That being said. There is no concrete difference in linux at the command line level. Right now I’m using 3 or 5 different servers with various versions of Linux: Centos, Suse, RedHat, Ubuntu, etc and don’t find any real differences. My first Unix experience was in 1980 with a Radio Shack Model 2 computer that ran a version of Unix. What I learned there still serves me well today.
As inexpensive as servers are, I can’t think of any reason to not to dedicate a server to telephony.
BF
Certainly from the development and maintenance of freepbx itself, limiting the number of platforms that it is supported on is much easier from the software devlopment perspective. That is why most commercial software development companies desire to limit the number of platforms that they support.
From the user perspective, however, requirements of users differ. For this reason, software companies often need to support two or more platforms (e.g., Microsoft .NET based as well as Java-based). All of the employers (enterprise software development firms), I have worked for have done just this specifically because propsective customers were addament about choosing only products that adhered to their major requirements (e.g., a Microsoft .NET based IT dept. or a Unix, Java-based IT dept.). This is true even though the servers were dedicated for only that enterprise software (license costs of the products being in the hundreds of thousands and sometimes millions range).
FreePBX, like other open source, is community driven and thus of course a different animal. However, user requirements still differ among people. The main driver is often familiarity and maintenance costs and time.
Linux distributions are certainly fundamentally similar. However, there are differences. For example, it appears that for CentOS, Asterisk combined with FreePBX will run together right out of the box – with a pre-configured ISO ready to go. But, many folks have apparently tried to get the same combination working on Ubuntu and have from what I have read had a tough time and for many ultimately unsuccessful outcome. Thus, there are differences between CentOS and Ubuntu and they are apparently significant enough as they relate to the Asterisk - FreePBX combination.
So even if the box is dedicated to telephony only, support for a number of (even if two) distributions for the Asterisk - FreePBX combination has value. I would guess that there probably are other distributions (other than CentOS) where this combination is already supported. http://www.voip-info.org/wiki/view/Asterisk+OS+Platforms lists a large number of platforms that Asterisk itself is supported on (including Debian --> Ubuntu). Whether the addition of support (i.e., a package) for FreePBX in addition to Asterisk’s existing support on Ubuntu would make sense of course is dependent on the degree of demand for the Asterisk - FreePBX combination which seems to remain unclear.
you are quote correct on that.
As such, if there is a part of the community that wants to create a package they would be very welcome to do such.
There are plenty of people running FreePBX on Debian based distros and FreePBX is designed to define the different directory structures to allow such. When hard coded directories are reported with suggested patches, they are almost always fixed. I recall only one case that was not for a very real reason that kept it from being able to be fixed and there were no proposals to overcome the issue. (It was in fact changed and then reverted.)
I have heard concerns by some Debian based community members wrt to the FreePBX Module Admin system that ultimately “breaks” any packaging system (RPMs included). However, it is very possible to do an initial install with a package like an RPM or equivalents as many distros do, and FreePBX picks up from there keeping the system up to date for its needs. It does nothing to get in the way of critical components like Asterisk which can be maintained with packages when available.
I have never had to do anything with drivers in FreePBX. I use Debian on all of my FreePBX/Asterisk servers. I have no problems NEVER. Standard PHP and Apache along with Asterisk install of Debian packages and that’s it.
Many people use Debian for the purpose or concerns of Licensing of applications. If you want to be 100 percent sure that you will not run into a license issue, you better not use Centos.
Centos does not guarantee that there are non-free software in the distribution.
The point I want to make is all of my hardware that I have used for telephony have worked on Debian as well as others. If you can’t compile it on Debian then you can’t compile it on Centos, unless they are using a very old version of Centos. This is usually the case, they continue to us old insecure version with no patches and push the distro out for years after the version was put out.
If you want to talk about solid I work with Purdue University and they run over 5000 servers with Debian installed mainly for security reason. Being that RedHat takes too long to resolve issues they don’t want to “risk it”.
All hardware that works on Linux and has drivers are very rarely not supported under Debian/Ubuntu.
In fact the Ubuntu Servers distro makes installing kernel modules for hardware almost a no brainer.
Can you show me real stats that says Debian/Ubuntu Server has not been proven by the masses?
Yes the ISO is for those that don’t know anything about an OS, but anyone that wants to understand security and what will keep them from being sued should always know what they are running.
I have used Centos and I hate working with it, it always seems so clunky. Everything requires a GUI to get something setup. I never could understand why a server needed a GUI. GUI’s pose security risk that we don’t want to risk.
Debian gives you the opportunity to install the smallest install (around 500 megs with Asterisk and FreePBX, try that with Centos)
Please if you have never tried or worked with something before, try to be a little more open. The Telephony of yesterday was closed, try not to do that with this too.
Just my 2 cents.
[em]
Also if someone could offer something, I would be willing to build a package for [size=15]Debian/Ubuntu Server.[/size][/em]
You can look at Debian VOIP for more of Debian VOIP support
debian/rules.
First, everyone’s comments and input have been fantastic and greatly appreciated.
I. From a users’ perspective:
From a users’ perspective, kenn10’s comments on the drivers along with his and Bill/W5WAF’s points on the value of sticking with the masses are very compelling in consideration of minimizing headaches and hastles by new users of FreePBX + Asterisk to “go with the flow” which apparently is CentOS.
Similarly, pshempel makes a further strong case of the value of meeting the diverse needs of a wide class of users. In this case: licensing issues, maintenance preferences (e.g., preference for direct file configurations over gui), smaller footprint of the server install, etc.
II. Increasing the community value of FreePBX + Asterisk combination perspective:
From this perspective, further diversifying the community base in terms of platforms that the combination is supported on (i.e., pre-configured package installs) would most likely further solidfy FreePBX as a diverse, multi-platform, “standard” application.
- The user base for the combination would be expected to expand. The rate of acceleration unclear at present.
- FreePBX + Asterisk combination would be made further robust resulting from a more diverse set of user feedback including feature patch contributions, enhancement suggestions, bug identification and resolution, greater overall security due to wider field of attacks, vulnerability identification and resultant hardening of the application and integration, etc.
- Higher likelihood of standardization of the FreePBX + Asterisk combination due to wider availability to folks who value using a particular distrobution such as Debian/Ubuntu. Otherwise, would leave open greater probability for alternative GUIs other than FreePBX to be used with Asterisk on those distributions. Perhaps something like gastman which has a package on Ubuntu would be furthered more than if FreePBX was also available in a package on Ubuntu.
III. Whether an Ubuntu package makes sense for FreePBX:
How would we evaluate whether making such a package would make sense? Any thoughts on how to go about due diligence for making this justification?
IV. Ubuntu package creation team:
If a “go” for an Ubuntu package is justified, how would we go about putting together a team to turn this into a reality?
I personally would be interested in contributing. However, as mentioned above, my own technical abilities for doing so are considerably limited. There would need to be some sort of a team; involvement of folks like pshempel who have successfully configured the FreePBX + Asterisk combo to work on Debian/Ubuntu would of course be extremely valuable for making this a success.
Others’ thoughts?
Hi
I would be willing to do this, building a package for FreePBX for Debian/Ubuntu Server would not require a team. This would only require about 40 hours of work setting up dependencies and about another 20 for testing.
I have been building packages for Debian for about 10 years now. I even attempted to get a Debian proper of FreeSwitch, now that is a challenge and would require 2 people to get that worked out with all of the dependencies it has. FreeSwitch really needs a Debian proper done of it to get it into the main Debian repositories. FreeSwitch would really take off then with the thousands of users that would take advantage of this.
I don’t have very much free time. But if someone would put up a small bounty I would be willing to have something done in the next 2 weeks. Without a bounty it would take about 6 weeks to get done.
Thanks
pshempel - this will be an extremely valuable addition to FreePBX, Debian, and Ubuntu community projects. I just checked out the info on FreeSwitch. Looks like you built the following? http://wiki.freeswitch.org/wiki/Installation_Guide#Ubuntu_Packages
Looks really impressive!
I personally am quite cash-constrained at the momemt and am trying to start a company with no income and 5 mouths to feed (including my own - and my wife is pregnant and thus she counts as 2). Basically, any diversion of cash right now comes out of grocery and gas money. Please understand, I am not embarassed by this situation, just in start-up mode here. 60 hours of your time is significant and understandable that a cash contribution would help you justify a shorter-time frame. So I would like to contribute but I cannot offer very much cash and I would not want to make a suggestion that would be offensive to you. I am guessing that I would be the only one on this thread willing to contribute as such? Again, I am reluctant to post an amount in this posting because it would be fairly low due to my circumstances and I would not want to be offensive by the low amount. Would you feel comfortable posting an amount that would help you justify shooting for the 2-weeks timeframe?
Thanks
I would think that there are at least a few persons here that would like to see this come to fruition.
Let say we put up a bounty and I set it at [em]$250.00~[/em] and complete this in say three weeks. If we can’t get that within the next two weeks, I will just begin work on this and set the time to finish within the next six to eight weeks.
I to am also building on a new business and at the same time working a full time job. This is taking a very large amount of my time now.
If that works for you let me know. I want to see FreePBX working in as many places as possible myself. I have been using Debian since version 1.8 and have found it to be one of the best packaging systems around.
Looking forward to your response
That is fantastic!
Let’s see if there are any other folks who also respond about contributing.
Thanks
u get my support. Get a freepbx for ubuntu server!
is it $250 per contributor or combined?
I use Ubuntu for all my servers, and for the occasional PBX, having to dive into CentOS is not always easy or straightforward.
I do not have much spare cash, but would be happy to help as a beta tester. I am reasonably proficient in Ubuntu administration and would like to give some time back to the community.
kindly give me conclusion remarks
depending on
- stablity
- number of user support( load )
which Linux distribution is more Valuable Ubunto OR Centod OR RED HAt
I’m new to Asterisk and computer-hosted telephony, but I configure and maintain web servers, internal file servers, and firewalls, generally about two dozen active systems at any given time, both for my own purposes and for clients.
For firewalls, I use either SmoothWall or pfSense, depending on whether or not a public DMZ is needed. SmoothWall includes their own custom hardened Linux kernel, pfSense is built on OpenBSD. I don’t care, and really don’t need to know, and I certainly would never suggest that either of them switch to a different distribution to match my whims.
For web servers, I find that CentOS gives me plenty of life for a given server during which security upgrades are available. It’s reliable. I’m not saying it’s the best possible, just that it works very well for that task and it’s unlikely that somebody else is going to come along with a compelling reason to change anytime soon.
Alas, to run Netatalk on CentOS requires manually recompiling the kernel, a task I prefer to avoid. But Debian installs both Samba and Netatalk by default in their server configuration, and so far I haven’t had a bit of problem with it in over twenty server-years of operation. Again, it’s hard to imagine a compelling reason to change.
Asterisk, as bsgcic points out, can be installed on a range of systems. I don’t want to learn Asterisk well enough to do that, I want a system that installs easily and runs reliably. One of the projects that purports to provide that is FreePBX. If they were running on some bizarre OS, particularly if it were one without a proven track record, I probably would look at something else. As long as the OS choice is credible, I don’t care what it is.
If someone wants to create a port of FreePBX on another distribution, that’s up to them. I think it would be a waste of effort, given the opportunity to build on the current project, but OSS is all about freedom. I really would hate to see any of the energy currently going into improving FreePBX to be diverted into that effort, however, as it doesn’t provide any benefit at all.
If Ubuntu wants to include a solid FreePBX package in their distribution, more power to them. But nobody else has any reason to make that effort. I’d prefer that people didn’t waste the time of the FreePBX team nattering about it.
Van