Elastix: phones lose registration

gjacobse

This was edited.

The 8 is the last part of the site name - blanked to remove site / client details.

scottalanmiller

Oh, how much packet loss is there?

gjacobse

Something that has been bothering me a bit is MAC addresses.

While it is not a 'end all' tool, I used AngryIP to scan the vLAN .100 - .189 to find all the active IPs. I added MAC access resolution to the scan and when it was done, all 25 IPs had the same MAC address.

I've logged into half of the phones and the phone GUI shows that they are different per phone.

Doing an ARP -a give this:

Interface: xxx.xx.5.111 --- 0xc
  Internet Address      Physical Address      Type
  xxx.xx.xx.100         00-01-e8-d7-ab-5f     dynamic
  xxx.xx.xx.101         00-01-e8-d7-ab-5f     dynamic
  xxx.xx.xx.102         00-01-e8-d7-ab-5f     dynamic
  xxx.xx.xx.103         00-01-e8-d7-ab-5f     dynamic
  xxx.xx.xx.104         00-01-e8-d7-ab-5f     dynamic
  xxx.xx.xx.105         00-01-e8-d7-ab-5f     dynamic
  xxx.xx.xx.106         00-01-e8-d7-ab-5f     dynamic
  xxx.xx.xx.107         00-01-e8-d7-ab-5f     dynamic
  xxx.xx.xx.108         00-01-e8-d7-ab-5f     dynamic
  xxx.xx.xx.109         00-01-e8-d7-ab-5f     dynamic
  xxx.xx.xx.110         00-01-e8-d7-ab-5f     dynamic
  xxx.xx.xx.111         00-01-e8-d7-ab-5f     dynamic
  xxx.xx.xx.112         00-01-e8-d7-ab-5f     dynamic
  xxx.xx.xx.113         00-01-e8-d7-ab-5f     dynamic
  xxx.xx.xx.114         00-01-e8-d7-ab-5f     dynamic
  xxx.xx.xx.115         00-01-e8-d7-ab-5f     dynamic
  xxx.xx.xx.116         00-01-e8-d7-ab-5f     dynamic
  xxx.xx.xx.117         00-01-e8-d7-ab-5f     dynamic
  xxx.xx.xx.118         00-01-e8-d7-ab-5f     dynamic
  xxx.xx.xx.119         00-01-e8-d7-ab-5f     dynamic
  xxx.xx.xx.120         00-01-e8-d7-ab-5f     dynamic
  xxx.xx.xx.121         00-01-e8-d7-ab-5f     dynamic
  xxx.xx.xx.122         00-01-e8-d7-ab-5f     dynamic
  xxx.xx.xx.123         00-01-e8-d7-ab-5f     dynamic
  xxx.xx.xx.124         00-01-e8-d7-ab-5f     dynamic
  xxx.xx.xx.125         00-01-e8-d7-ab-5f     dynamic
  xxx.xx.xx.126         00-01-e8-d7-ab-5f     dynamic
  xxx.xx.xx.127         00-01-e8-d7-ab-5f     dynamic
  xxx.xx.xx.128         00-01-e8-d7-ab-5f     dynamic
  xxx.xx.xx.129         00-01-e8-d7-ab-5f     dynamic
  xxx.xx.xx.130         00-01-e8-d7-ab-5f     dynamic
  xxx.xx.xx.131         00-01-e8-d7-ab-5f     dynamic
  xxx.xx.xx.132         00-01-e8-d7-ab-5f     dynamic
  xxx.xx.xx.133         00-01-e8-d7-ab-5f     dynamic
  xxx.xx.xx.134         00-01-e8-d7-ab-5f     dynamic
  xxx.xx.xx.135         00-01-e8-d7-ab-5f     dynamic
  xxx.xx.xx.136         00-01-e8-d7-ab-5f     dynamic
  xxx.xx.xx.137         00-01-e8-d7-ab-5f     dynamic
  xxx.xx.xx.138         00-01-e8-d7-ab-5f     dynamic
  xxx.xx.xx.139         00-01-e8-d7-ab-5f     dynamic
  xxx.xx.xx.140         00-01-e8-d7-ab-5f     dynamic
  xxx.xx.xx.141         00-01-e8-d7-ab-5f     dynamic
  xxx.xx.xx.142         00-01-e8-d7-ab-5f     dynamic
  xxx.xx.xx.143         00-01-e8-d7-ab-5f     dynamic
  xxx.xx.xx.144         00-01-e8-d7-ab-5f     dynamic
  xxx.xx.xx.145         00-01-e8-d7-ab-5f     dynamic
  xxx.xx.xx.146         00-01-e8-d7-ab-5f     dynamic
  xxx.xx.xx.147         00-01-e8-d7-ab-5f     dynamic
  xxx.xx.xx.148         00-01-e8-d7-ab-5f     dynamic
  xxx.xx.xx.149         00-01-e8-d7-ab-5f     dynamic
  xxx.xx.xx.150         00-01-e8-d7-ab-5f     dynamic
  xxx.xx.xx.151         00-01-e8-d7-ab-5f     dynamic
  xxx.xx.xx.152         00-01-e8-d7-ab-5f     dynamic
  xxx.xx.xx.153         00-01-e8-d7-ab-5f     dynamic
  xxx.xx.xx.154         00-01-e8-d7-ab-5f     dynamic
  xxx.xx.xx.155         00-01-e8-d7-ab-5f     dynamic
  xxx.xx.xx.156         00-01-e8-d7-ab-5f     dynamic
  xxx.xx.xx.157         00-01-e8-d7-ab-5f     dynamic
  xxx.xx.xx.158         00-01-e8-d7-ab-5f     dynamic
  xxx.xx.xx.159         00-01-e8-d7-ab-5f     dynamic
  xxx.xx.xx.160         00-01-e8-d7-ab-5f     dynamic
  xxx.xx.xx.161         00-01-e8-d7-ab-5f     dynamic
  xxx.xx.xx.162         00-01-e8-d7-ab-5f     dynamic
  xxx.xx.xx.163         00-01-e8-d7-ab-5f     dynamic
  xxx.xx.xx.164         00-01-e8-d7-ab-5f     dynamic
  xxx.xx.xx.165         00-01-e8-d7-ab-5f     dynamic
  xxx.xx.xx.166         00-01-e8-d7-ab-5f     dynamic
  xxx.xx.xx.167         00-01-e8-d7-ab-5f     dynamic
  xxx.xx.xx.168         00-01-e8-d7-ab-5f     dynamic
  xxx.xx.xx.169         00-01-e8-d7-ab-5f     dynamic
  xxx.xx.xx.170         00-01-e8-d7-ab-5f     dynamic
  xxx.xx.xx.171         00-01-e8-d7-ab-5f     dynamic
  xxx.xx.xx.172         00-01-e8-d7-ab-5f     dynamic
  xxx.xx.xx.173         00-01-e8-d7-ab-5f     dynamic
  xxx.xx.xx.174         00-01-e8-d7-ab-5f     dynamic
  xxx.xx.xx.175         00-01-e8-d7-ab-5f     dynamic
  xxx.xx.xx.176         00-01-e8-d7-ab-5f     dynamic
  xxx.xx.xx.177         00-01-e8-d7-ab-5f     dynamic
  xxx.xx.xx.178         00-01-e8-d7-ab-5f     dynamic
  xxx.xx.xx.179         00-01-e8-d7-ab-5f     dynamic
  xxx.xx.xx.180         00-01-e8-d7-ab-5f     dynamic
  xxx.xx.xx.181         00-01-e8-d7-ab-5f     dynamic
  xxx.xx.xx.182         00-01-e8-d7-ab-5f     dynamic
  xxx.xx.xx.183         00-01-e8-d7-ab-5f     dynamic
  xxx.xx.xx.184         00-01-e8-d7-ab-5f     dynamic
  xxx.xx.xx.185         00-01-e8-d7-ab-5f     dynamic
  xxx.xx.xx.186         00-01-e8-d7-ab-5f     dynamic
  xxx.xx.xx.187         00-01-e8-d7-ab-5f     dynamic

JaredBusch

Is there NAT involved in their VPN?

Either that or some other routing problem.

I bet that MAC is the router.

If so randomize the SIP port on all the phones. Your problem will go away.

JaredBusch

@g.jacobse said:

00-01-e8-d7-ab-5f

A lookup shows that as Force10 Networks.?
http://www.coffer.com/mac_find/?string=00-01-e8-d7-ab-5f

gjacobse

@JaredBusch said:

@g.jacobse said:

00-01-e8-d7-ab-5f

A lookup shows that as Force10 Networks.?
http://www.coffer.com/mac_find/?string=00-01-e8-d7-ab-5f

Confirmed. Switch as the other end of the VPN.

JaredBusch

@g.jacobse said:

Confirmed. Switch as the other end of the VPN.

Change the phones to use some port other than 5060 for the SIP registration.
Make sure they are all different as you cannot reuse the same port behind a NAT and this issue is looking to me like that.

JaredBusch

Well, that or make who ever runs the network figure out what is happening on their side.

coliver

@JaredBusch said:

@g.jacobse said:

Confirmed. Switch as the other end of the VPN.

Change the phones to use some port other than 5060 for the SIP registration.
Make sure they are all different as you cannot reuse the same port behind a NAT and this issue is looking to me like that.

Would the usage of a STUN server solve this issue?

scottalanmiller

@JaredBusch said:

Well, that or make who ever runs the network figure out what is happening on their side.

That is proving difficult.

scottalanmiller

@coliver said:

Would the usage of a STUN server solve this issue?

Should not be any NAT.

JaredBusch

@coliver said:

@JaredBusch said:

@g.jacobse said:

Confirmed. Switch as the other end of the VPN.

Change the phones to use some port other than 5060 for the SIP registration.
Make sure they are all different as you cannot reuse the same port behind a NAT and this issue is looking to me like that.

Would the usage of a STUN server solve this issue?

Yes, but you would have to set one up internally as they obviously have everything inside the VPN.

JaredBusch

@scottalanmiller said:

@coliver said:

Would the usage of a STUN server solve this issue?

Should not be any NAT.

Should not be any NAT, but it sure is acting like it if everything is reporting the same MAC.

scottalanmiller

Yeah, that is very fishy.

gjacobse

Statement I was given regarding the MACs

...as an aside, this has nothing to do with the, or any, firewall.  This is normal layer2/layer3 handoff in any network...

NetworkNerd

To give background here, the Fortigate has a Juniper L3 switch connected to it. Supposedly the config on that guy has not changed, but I don't think we really know.

Dashrender

@JaredBusch said:

@scottalanmiller said:

@coliver said:

Would the usage of a STUN server solve this issue?

Should not be any NAT.

Should not be any NAT, but it sure is acting like it if everything is reporting the same MAC.

Is that right? Wouldn't you always loose the originating MAC if you go through a router of any kind? i.e. a VPN?

JaredBusch

@Dashrender said:

Is that right? Wouldn't you always loose the originating MAC if you go through a router of any kind? i.e. a VPN?

It depends on the VPN. It can easily bridge at either Level 2 or Level 3. One will let everything pass through, the other would show the MAC form the bridge. Either way, as long as there is no NAT, it should not interfere. That is true.

Just the description of the problem fits more to a NAT scenario.

gjacobse

Speaking with a 3rd party network tech who is one site. We have three phones aren't connecting - turns out that one of the POE ports was bad...

One is bad,.. more is likely. replace the switch!

gjacobse

Looks like more than one port is bad as far as POE goes.

got down to it being either the phone itself or the port, but ports were 'testing' okay. Defaulted the three phones that were missing and they now have IP addresses where we can get to them and program them.

Sad news is that during all this,.. they broke the VPN and we can't get in.

Interestingly on that note, they have a Fortigate 60D at that site. they can't update the firmware to do the testing they need. Hmm.. Firmware update failed... make you wonder.

Glad to see we didn't lose three phones though. Losing a port on the Juniper is bad enough (just the POE side it seems)... but still.