Jump to content

Vodia PBX

Administrators
  • Posts

    11,110
  • Joined

  • Last visited

Posts posted by Vodia PBX

  1. yes. so are you saying that if the ntp server which checks the time every so often doesnt get a reponse the time on the PBX will reset to JAN 1? 00:00 ??

     

    because it was working before hand.

     

    On the CS410, the PBX first uses ntpclient to set the initial time on the Linux level. Then the time is not set any more on the OS level. The PBX uses NTP just to keep track of time drifts, but does not change the time on the OS any more.

     

    When you change the NTP server from the web interface, you need to reboot the system before that takes effect.

  2. So do I need to set up something in the ATA or the CS410?

     

    In the ATA. I remember there was an option to disable re-registration upon a new registration - which makes sense because why would you kick out other registered phones just when the ATA reboots.

  3. I do see an RTP Timeout in the logfile viewer when the call is cutoff:

    [5] 2008/10/03 09:36:23: RTP Timeout on 9336d5f6@pbx#29825

     

    Ok, that makes it clear that the PBX disconnects the call because it did not get any media any more. Did you use the PnP configuration for the phones? Seems strange that the phone does not send any keep-alive traffic (not even silence indicators).

     

    Maybe you can do a Wireshark trace when the situation happens to see if the phone really keeps completely quiet.

  4. I have seen many issues with Snom 360s and using domain names. They tend to lock up, or do weird things about every other day if you use a name, however if you use an IP, they will run for an average of a month or 2 without the same weird issues.

     

    Well, you can still use an IP address in the the outbound proxy and at the same time use a domain name in the "domain" setting.

     

    If the phones are so sensitive about DNS, then you should also try to make DNS as stable as possible. DNS problems may also exist for firmware update and other services locally on the phone.

  5. Do we need to manipulate dialplan in OCS or PBXNSIP or what else is needed to do this?

     

    The only think that comes to my mind is to use the ANI field. ANI is for outbound, tel:alias (or just a regular alias) for inbound. But I am not the big OCS expert, maybe it is easier to manipulate the caller-ID in OCS. And don't forget the gateway. Most gateways have great flexibility rewriting the caller-ID.

  6. No it is not SIP aware. Just a Microtek. I should also mention im running that beta software. probably should roll back and see what happens.

     

    Even for a beta I would be surprised, as this is a fundamental routing function. But anyway, maybe just try the roll-back version.

  7. I just dont want individual domain tenants to see the log information for the whole box.

     

    Oh, okay. Actually if they log in as domain admin then they don't see the log at all.

     

    The big question is if you want the domain admins to spend too much time with troubleshooting anyway. So far our answer was "nah"...

  8. ok I receive a call below.. but the IP is from the router NOT from the original gateway. I am getting this on a couple CS410's here also: (PS IP changed) So im getting 404's unless I make a trunk with the router IP then it works.

     

    [0] 2008/10/01 21:32:59: SIP Rx udp:208.187.9.9:5060: <---- IP of gateway

    INVITE sip:8015063400@208.187.69.69 SIP/2.0 <--- correct Invite

    Via: SIP/2.0/UDP 66.7.99.99:5060;branch=z9hG4bK45e8aa75;rport <--- correct source

    From: "DAVID BURR " <sip:8014038529@66.7.99.99>;tag=as12012ea7

    ...

     

    [5] 2008/10/01 21:32:59: Trunk test sends call to 8015063400 in domain 999.nexsip.com

     

     

    is this a bug? why would it say the router was the source IP? most of the time it works just fine. then suddenly i get 404 errors then it starts working again. This is happening in multiple locations.

     

    Well, this is not a bug of the PBX... What router model is it? Maybe it is SIP aware? Check if there is something that you can turn off. Routers do not change IP addresses, unless they are doing NAT. But then they should change it only for outbound traffic, not inbound.

  9. however when i send a call out from the domain the SRC is the IP of eth0. which makes sense, however is there any way to make the IP SRC be the domain instead of the eth0 IP?

     

    That way i can control the LD costs from the carrier instead of having to sort it out myself now.

     

    You should take a look at the ICID parameter in the trunk (see http://wiki.pbxnsip.com/index.php/Trunk_Settings). Maybe that is a better way to keep track of where the calls come from.

  10. I am having a few CS410's that are loosing their time. Box this morning has an uptime of 17 hours. however I ssh'd and ran 'date' and it says: Tue Jan 1 03:13:45 UTC 2008

     

    Which means 3 hours ago, the date reset for no reason.

     

    Sounds like a problem with DNS or NTP itself. Is the PBX able to resolve pool.ntp.org?

  11. The addres is filled in but I replaced the phone number with the bold text (the bold text contains the phone number to dial but I removed the phone number). Doing this with another phone it works fine.. except from this snom 360 phone.

     

    It must be somehow related to this specific snom 360 instance. I don't believe it is related to the TAPI.

  12. I have just ventured into setting up multiple domains in the hopes of hosting if all of the testing goes well. I am having some issues with the second domain. For example sake, lets say that I setup the DNS records as follows.

     

    75.136.163.71 = voip.company1.com

    75.136.163.72 = voip.company2.com

     

    Under the PBXnSIP domain setup, I created a domain with the FQDN as the domain name, and I set the IP address as an alias. I am having the following issues when testing.

     

    When I put a call on hold, the caller does not hear hold music, and they cannot take them off of hold. Upon inspection of the packets, I see the IP of company1 within the payload of the packets sent from the PBxSIP machine (not the from address)

     

    When TFTPing down phone config files, The request from the phone goes to the IP of company 2, however the IP of company 1 responds, and some firewalls block this since the data is coming from a different IP than the phone made the request to.

     

    I know it is possible to only bind the PBX to specific IPs, however is it possible to do it at a domain level?

     

    I realize it is possible to point both DNS names at the same public IPs, however there are other (non voice) aspects of my hosting that sharing an IP will cause issues with

     

    I would not mix domain names and IP addresses. You can just use one IP address for all domains. Just make sure that the phones use the domain names in the SIP requests. You should tell the users to use an outbound proxy, it can also be the domain name, it can also be another DNS name or even the IP address (though I would not recommend that, maybe one day you want to use the server to another address).

  13. In our hosted environment, so far the PnP functionality is working very well for us in most cases. However, we have run into a couple of issues. The main problem relates to the fact that the VLAN IDs are provisioned globally. We can not set the VLAN tag on the server, because we do not have a common VLAN ID which will be set across all the clients on the server, and if we do not set a value, it erases the VLAN tag set on the phone manually. It would be nice if this can be overridden per extension or even per domain, but we would settle for just having a switch so the server does not attempt to provision the VLAN tag.

     

    Yes, that makes sense. Maybe we should move this setting into the extension. The whole VLAN topic is kind of moving target to me right now, with 802.1X and carrier Ethernet. I am really not sure what the best way it, maybe we have to try a couple of things.

     

    We also ran into an issue where a carrier was blocking port 5061 on the TCP side, luckily they had a secondary carrier that they could route their VoIP traffic over which wasn't blocking the SIP TLS port. It would have been nice if we could change the transport method (udp, tcp, tls) per domain instead of globally on the server.

     

    Well, you can have two or more TCP/TLS ports open on the same server. The PBX can deal with that.

     

    There is also a small security problem in the form of an information disclosure, since the 'snom_3xx_phone.xml/admin_pin' and 'snom_m3.cfg/VOIP_SETTINGS_PIN_CODE' settings are global, they are the same for all the domains. This is not a big deal for us right now, but at some point it would be nice to set the admin pins per domain.

     

    We already made it possible to use domain admin PIN and passwords, but that is not in the provisioning files for the phones. We keep that also on the radar.

  14. Another one I can't find

     

    Whereabouts is this? I can't see a setting on the extension for it? Also, how does this fit in with dialog permissions? We previously implemented this "how to decide who can pick up a call" with rather complicated dialog permissions strings to prevent certain people picking calls outside their team.

     

    What would be really nice is what the Mitels do, and make one of the softkeys into a "pickup" button when a phone in their group is ringing. I'm not quite sure SIP allows for that sort of thing though :lol:

     

    Yea, that 3022 build did not have it, we made a 3023 build in the meantime.

     

    SIP? That's not the question here... The problem is that this might be a performance killer. I would like to leave it the way it is for right now and see how it is being received in the real world.

×
×
  • Create New...