Vodia PBX

We want to brush up the end user documentation and also translate it into at least the major language. End users are difficult to satisfy with a Wiki page

If you are have external users it does pay back to have a static public IP address. I would run one NIC on that public IP address and just to transparent routing (no NAT), that makes the setup pretty easy and your life much more comfortable. If you have private users spending another private IP address does not cost you anything and also makes your life much easier (no need to run calls from the private IP through a firewall).

No, you can also do that with one NIC - but then things get more complicated. You need to use the settings "IP Routing List" and put something like "192.168.0.0/255.255.0.0/192.168.1.2 0.0.0.0/0.0.0.0/123.124.125.126" there. This means "if you are sending to the 192.168.x.x, use the 192.168.1.2 IP address, otherwise if you send it to anywhere else, use 123.124.125.126". And you need to program your firewall so that it forwards SIP and RTP to the private IP address (192.168.1.2 in this example). Not very user friendly, but in that environment user friendlyness is difficult.

It is difficult for a large organization to offer only excellent products... I am sure there is a configuration that tells the switch that anything except 0 is a bad codec. Probably there is a IOS command somewhere that tells the switch to reject any packet that is not only the specified codec. Unfortunately, I am not a CCSSE so I can't tell you what command is neccessary to make the device SIP-compatible again...

Try disabling the pass-through mode. This seems to be an ongoing pain in the neck, especially with the CS410 PSTN gateway. You can do this from the web interface (once that you are logged in as admin) with http://1.2.3.4/reg_status.htm?save=save&am...s_through=false.

Most probably you have a routing problem with the RTP (SIP seems to be okay, maybe because you forwarded the RTP traffic). Check out http://wiki.pbxnsip.com/index.php/Office_w...ic_IP_addresses, there you can see how you can use DMZ to get the setup you want. If you have Wireshark on the PBX host, you will be able to see what is going wrong.

I assume that response comes back to a SDP offer in an outgoing INVITE? If that is the case, the 400 Bad Request response is just an indication that whatever equipment they are using, they don't understand SIP and SDP properly yet... What the PBX says is: "Send me A-law, U-law, G729, G.726 or GSM" and "My first preference would be A-law, my second u-law" and so on. If the PBX would only allow A-law then it would not send the other codecs, right? The correct behavior would be to pick the codec they like and send the answer back to the PBX. Anyway, what you can do it set 0 in the trunk codec list or even better, in the global codec list of the PBX (under the ports section). Then the PBX will only offer u-law, and that provider gets his way!

http://www.pbxnsip.com/software has it!

It is already available.

There were a couple of important fixes around the daily sending of emails that caused problems in several places. I think everyone running 2.1 versions should move to that version. The release notes are available at http://wiki.pbxnsip.com/index.php/Release_Notes_2.1.5.

If you don't know from which IP address a request will come to the PBX and you want that this request lands on the trunk, then you must not use a outbound proxy. However, then you must make sure that the PBX is able to identify that the call goes to the trunk - see http://wiki.pbxnsip.com/index.php/Inbound_...ifies_the_trunk for more on this topic. The outbound proxy is also the "inbound proxy"... You can make the PBX bind only to one IP address, but you would do that in the port settings of the PBX. See http://wiki.pbxnsip.com/index.php/Port_Setup. You can use static registrations if you want to send a call to a specific IP address when an extension is being called. See http://wiki.pbxnsip.com/index.php/Extension#Registrations for more on this topic. But usually this is not a good idea. Practically all (S)IP phones support registrations.

http://www.pbxnsip.com/download/pbxctrl-debian3.1-2.1.5.2357

As these kind or problems was causing more and more head scratching we added more logging in 2.1.5 so that you can see what actual ERE being sent to the pattern matching subsystem.

I think the problem has been found - it was sending the license warning to the user. If the license is valid less than 30 days and you have set the email address for the admin it made boom. Workaround - take the admin email out (or get a poermanent license). We will make a 2.1.5 release today that fixes that problem.

It can happen that in the first attempt the email did not go out. Then the PBX would increase the number of unsuccessful attempts. That same email might go out later in a successful attempt, incrementing that counter. The message Last email could not be sent simply says that you need to investigate what is going on. I would recommend to turn only email logging on and see what the PBX has to say in the log.

I would recommend to always use the outbound proxy setting, even if it is identical to the domain name. It just makes it 100 % clear where the request should be routed to. The only exception I see is ENUM, and accepting calls from the wide open Internet - but I guess that is not on you list.

Ouch. It seems that the file system still has the messages XML files after the users deleted them. There was this version (2.1.3?) that was not flushing the file buffers all the time, and those messages are probably a remainder from this version. There must be a correlation between the files references in the messages XML files and the WAV files in the recordings directory. If a file referenced in the XML does not exist, it is safe to remove that XML file. The latest version does not have this problem any more, the problem will go away eventually when users go through their messages.

Looks to me like the PBX for some reason rejects to dial the provided number. Does the dialplan allow that number? Also I don't see the DTMF in the log, but I guess that is because of the log settings.

Well it should append the new log messages. What OS are we talking about? Are you sure?

I have seen a prototype a few days ago. We expect to have a beta version by the end of January.

Maybe just post the whole packet. Also pay attention to the upper and lower case, some versions were a little bit picky about that.

Okay, we identified two possible reasons for a crash. One was obviously a problem with a session reference in the CDR list - if there was a http session that times out, the PBX would use a reference to that session which possibly accesses memory that is not available any more. The other possible reason for a crash was a similar problem with buttons that needed translations (e.g. for the DND button, and agent login/logout state). There is a new build at http://www.pbxnsip.com/download/pbxctrl-2.1.5.2355.exe (manual upgrade). It would be great if we could verify that this was a problem.

You can ignore those messsages or move or remove the ports from the ports section (admin/settings/ports). 443 = https 69 = tftp

No. In the EU people don't like to dress and clean up so that they can get into a conference call.

Any logs? We are going through the whole session logic and cleaning that area up. Maybe there was something else not 100 % clean. Seems that also other places like buttons got "infected"...