How Asterisk is soiling itself

Part of the good things about open source is that you can “get the source”. If somethings wrong, you can fix it.

That is, if you know how to hook all the pieces together. If the documentation is correct and valid

Why, if asterisk has dropped suport for zap, does this page still talk about getting a copy of zaptel?

Why if you use SVN to get the source, will it not build?

[quote]dahdi_cfg.c: In function âspanconfigâ:
dahdi_cfg.c:376: error: âDAHDI_CONFIG_NTTEâ undeclared (first use in this function)
dahdi_cfg.c:376: error: (Each undeclared identifier is reported only once
dahdi_cfg.c:376: error: for each function it appears in.)
dahdi_cfg.c:380: error: âDAHDI_CONFIG_TERMâ undeclared (first use in this function)
make[1]: *** [dahdi_cfg.o] Error 1[/quote]

Maybe the plan is to make more money on support by making it necessary, but seriously guys, I’m at the point, I’m ready to sell my asterisk hardware and just go get something that works. I’ve got more than 10 hours into this, and I still don’t have a setup that works or even has a hope of working anytime soon. Getting the source tarballs doesn’t help, since running the code from the source tarballs produces errors like:

[quote]sudo dahdi_genconf chandahdi
Use of uninitialized value $type in hash element at /usr/local/share/perl/5.10.0/Dahdi/Config/Gen/ line 160.
Use of uninitialized value $type in hash element at /usr/local/share/perl/5.10.0/Dahdi/Config/Gen/ line 161.
Use of uninitialized value $type in hash element at /usr/local/share/perl/5.10.0/Dahdi/Config/Gen/ line 162.
Use of uninitialized value $type in string eq at /usr/local/share/perl/5.10.0/Dahdi/Config/Gen/ line 166.
Use of uninitialized value $type in concatenation (.) or string at /usr/local/share/perl/5.10.0/Dahdi/Config/Gen/ line 167.
missing default_chan_dahdi_signalling for chan #1 type at /usr/local/share/perl/5.10.0/Dahdi/Config/Gen/ line 167.

Sure, you guys may know how to work around these problems because you deal with every day, but it makes it very difficult for someone who last set up an asterisk system five years ago, and needs to do it again with the same hardware due to a motherboard failure.

Asterisk NOW!, doesn’t detect the phone hardware, so that option seems to be out as well. I’ve been in the software business for 20 years, and products that don’t work when you get them seldom get a second chance, if I didn’t have five years experience with asterisk, I wouldn’t consider it any more at all.


Ok, I understand you if you have only 10 hours using linux and asterisk, at the beginning the final goal looks pretty imposible. But how can we solve many issues in our installations? Well first is read the documentation, the best place is read all the comments in the config files.

If you still with questions the next step is read about your config file in the web, maybe, then look at many forums to try solving the issue.

In the company that I work we integrate VoIP-solutions with asterisk, and have many servers with many different hardware vendors, sangoma, openvox, digium and rhino. And no problems, even we have customers with 4 years without request support.

I recommend to you build your asterisk from source, everything from source dahdi or zaptel, libpri, openr2, asterisk and the addons. With that you can control everything in the config files. if you want to add a GUI use FREEPBX.

You can build asterisk in many Linux Distros but we always prefer CentOS.

I’ve built 2 previous asterisk systems, and when I did, zaptel was the model for the telephony boards. Dahdi replaced that, and when it did, the config changed, but the documentation didn’t change with it. The dahdi tools don’t work, from either the tarballs, or subversion.

Build it from sources, easier said than done when the code won’t compile because things are not defined.

I’ve been working with Unix since 1986, writing kernel drivers for the tcp stack and FDDI. I’m no novice to building things. But when the source is broken, it takes far too long to come up to speed on all the things you need to get it to work. Especially when the documentation is incomplete. We’ve had no trouble with our system over the last 5 years, but when the hardware died, we figured it was time to move up to the most recent release, so that when the next hardware dies 5 years from now, we won’t be so far behind the curve. Looks like this was just a poor release to try coming up to speed on.

For a project to be viable for other people to build it, releases should work/compile, without having to mod the source because people failed to test the code they “released”. The error messages in my previous post speak for themselves.

You seem to have quite a lot of programming experience. So, why not contribute to fix the broken codes, i.e. if you think the source codes are broken?

Because at the moment, I’m more interested in getting a phone system working. Code that is released, should work.

I’m not interested in coding a phone system, I have other fish to fry, that I actually find interesting and tasty. I put my first asterisk system together when asterisk was pretty much just Mark Spencer and one other guy. I thought it was cool, but I haven’t ever wanted to do telecom code.

I hate phone calls too much.

I wish you luck to wait for your solutions here. :mrgreen:

you can still use zaptel.

how about a iso like asterisknow, switchvox or PIAF?

puttyWe finally managed to get a running system off the AsteriskNOW! iso. Took around 12 hours to get it configured and running. One of the problems being that genconf always generates ISDN settings T1 cards, and FreePBX generates a lot of the conf files automatically, so just editing them by hand was off-limits. By doing this, we failed to move up to the 1.6 Asterisk. The fact that the AsteriskNow! iso doesn’t use 1.6, indicates to me that my post is fairly on target about things not being ready, since even this released version doesn’t use 1.6.

The Centos distro on the AsteriskNow iso, is disowned by the centos group, so the slow boot time (10+ minutes) is something we’ll have to live with for the moment. Since we have a running system, I’ll probably try to get back on-top of Debian, but considering all the issues I had with compiling it may be months before have sufficient time to get a nice setup.

My initial impressions with Asterisk were very favorable, I used it as voicemail attached to the DK280, it was rather straight forward, AND IT WORKED! and that was when all the configuration had to be done by hand. Now, nearly 8 years later, it’s harder to get a working system, due to complexities in the configuration that do not appear to be documented anywhere. This is in spite of the fact that there are now web-based systems for managing your Asterisk set up.

Toshiba used concepts in their setup, that appear to be missing from the asterisk setup, especially in the web tools. things like phantom extensions that could be made to appear on the phones, that could be targets of incoming DIDs, even though I haven’t used a Toshiba system in the last five years, some of those concepts are still crystal clear to me, whereas parts of the configuration in Asterisk are still fuzzy, even though I just managed to make the system run.

Part of the reason people like Windows and Mac is that it’s fairly easy to get the thing up and running. These tools did not rise to preeminence, because they had a better marketing budget, because the best marketing in the world can’t make a product it’s difficult to use easy. But they did deliver on ease-of-use, ease of setup, and it just works. Right now most people buy phone systems from a vendor, because that’s the only way to get one to work. The concepts are too foreign to most people. Having worked on software for 20 years, writing kernel drivers and application code, I can deal with foreign concepts and difficult topics, but I can’t deal with is software that is counter intuitive and poorly documented. And tools that lie.

The inspiration behind Asterisk is a good thing, but in a lot of ways the execution leaves a lot to be desired.

Sorry man, but I don’t think like you; What about the THOUSANDS of installations in the world fully functional with asterisk, why some governments move their telephony system with asterisk?.

I’m not a Linux expert, even a kernel coder, but we have at least 30 servers in my country fully functional, so I don’t know why you try to soiling Asterisk Software, if you want an in-hand solution Buy a SwitchVox server, use Trixbox, but don’t sell to us the Idea that Asterisk and the documentation is a crap.

The heart of Trixbox is also Asterisk. :mrgreen:

SwitchVox = asterisk with commercial support

The heart of Trixbox is also Asterisk. :mrgreen:[/quote]
Yes I know, but i recommended because is only insert a CD next next done it. And switchbox because its Asterisk with pay support

The heart of Trixbox is also Asterisk. :mrgreen:[/quote]
Yes I know, but i recommended because is only insert a CD next next done it. And switchbox because its Asterisk with pay support[/quote]
Well, OP doesn’t seem to be content with Asterisk. So, recommending other products based on Asterisk is useless. Perhaps, OP should write his/ner own PBX system like FreeSwitch. This way, no one is to blame. :mrgreen:

Maybe you should actually read my original post, I’m not looking for solution this insert a CD, next, next, done. I’ve already gone to the trouble to download the software and attempt to compile it.

What my post was about, is the fact that software is being released that doesn’t work, won’t compile, and has issues. All of these can be corrected by taking more care with what is released, and by keeping documentation current with code.

One of my logged issues, has already been fixed by the developer, question is, how did it get released?

Another issue, is the fact that dahdi_tool has been broken all the way back to what was zap_tool (earliest record I can find is ~2005). This fact is well known (at least among developers) and yet nothing has been done about.

I’m not disputing that Asterisk can and does work, I’ve used it now for several years, but the quality of releases, especially the current releases seems to be lacking. and saying maybe I should write my own, is just another way of sticking your head in the sand and ignoring the fact that there are issues that need to be fixed.

The current Asterisk NOW! iso seems to have problems with dual core machines, but performs very well on single core machine. Does anyone know why? Is anyone even aware of the problem?

[quote=“mark0978”]One of my logged issues, has already been fixed by the developer, question is, how did it get released?[/quote]If you don’t want to contribute, don’t send in your fixes. This way, nothing will get released.

I agree with you, especially DIgium is keeping a lot of release versions, i.e. 1.2.x, 1.4.x, 1.6.0.x, 1.6.1.x, 1.6.2.x, etc. However, when it comes to free/open source software, I will try to not to make any complains unless I can contribute. I am no programmer and as such I won’t make complain, except be happy to use it.

If you don’t stick your head in the sand, you won’t know how it feels. So, go ahead to stick your head in the sand. :mrgreen:

AFAIK, no one seems to be aware of the problem, except you. So, why don’t you launch your own investigation and report back to here.

I have already logged 3 bugs, 1 of which has been fixed, so I am being proactive. The more people that are aware of the problems, the more likely they are to be fixed…