Solved Slave Bridge PBX can't see presence of just 1 location

[johnporteous]

Hi All,
I have a weird situation that arose this weekend when I migrated an existing PBX to a new server and changed the OS.

In this issue, there are 3 self hosted PBXs, geographically spread. Prior to the migration (done a restore on the new VM & OS and worked a charm), all 3 PBXs could see the presence of one another and call with zero issues.

Here is how I have it configured (when I say 'can see' I mean extensions are listed in the phonebook etc).

PBX A - Is Master to PBX B and Master to PBX C. Can call and see both B and C
PBX B - Is Slave to PBX A and Master to PBX C. Can call and see both A and C
PBC C - Is Slave to PBX A and Slave to PBX B. Can call both A and B; can see PBX B....BUT Cannot see A.

That is the issue; PBX C cannot see PBX A.

When I restored, I simply changed the IP of the Bridge where required but it didn't seem to like it so I deleted all Master and Slaves between the 3 and started again. In fact I have done it several times, methodically going through settings to ensure I had missed any, done side by side comparisons of settings for the same reason. Everything appears to be exactly set correct, just the presence won't work.

The firewall passes tests, no issues making outbound calls, rules are correct as tested calling between all PBXs.

I had this issue a couple of months ago and it was cleared by simply adding :443 once. This time around as it's a migration, as part of the new install from 3CX you have to nominate the HTTPS port which I accepted as default 5001. I wondered if that can be an issue, but it's not because PBX B has been set with the FDQN:5001 and can see PBX A.

I hope that makes sense, please feel free to check with me if you need further info. I'm at my wits end trying to get this one resolved, and would sincerely appreciate any help you can offer me.

 

[johnporteous]

That's a good question; interestingly no I can't. And I can reach mgmt console of PBX B from PBX C. But that confuses me a bit as the public IP and FQDN are out in global DNS; I don't run any DNS servers (all 3 PBXs are sat in Azure); what would cause that?

 

[Alphabetic]

Well, if you ping it and it comes back with the old IP - DNS.

If you ping and it doesnt reply, firewall.

 

[johnporteous]

LOL....well of course; I wrote that without thinking - I'm frazzled LOL

 

[Alphabetic]

(╯°□°）╯︵ ┻━┻

 

[johnporteous]

Well I'll be darned; Good old MS Azure strikes again. @kieferschild that was the issue. PBX C is still doing a lookup for PBX A under the old IP. Didn't cross my mind that it could be the case, as I got clouded that B could see A. I'll get that sorted soonest and confirm. Many thanks for your perseverance.

 

[AWS2P]

if you nslookup each FQDN are they resolved and with the right IP

 

[johnporteous]

Kieferschild has already mentioned that. They do everywhere except on PBX C which is the issue it seems. For some reason, Azure DNS has not updated with the new IP of PBX A.

 

[Alphabetic]

it still hasnt updated?

ipconfig/flushdns&&ipconfig/registerdns

 

[johnporteous]

After forcing the DNS to update on PBX C and resolve to the new IP of PBX A when doing an NSLOOKUP, that made no difference. Also the firewall has 5001 open. So back to square one it seems.

 

[Alphabetic]

After forcing the DNS to update on PBX C and resolve to the new IP of PBX A when doing an NSLOOKUP, that made no difference. Also the firewall has 5001 open. So back to square one it seems.

Can you access the management console.

 

[johnporteous]

I done the ipconfig flush ages ago, Didn't update. In the end I resorted to changing the DNS servers in Azure to point to Google and Cloudflare instead of Microsofts DNS and it updated instantly but the issue remains. I may rebuild the bride between PBX A and C to see if that clears the issue

 

[Alphabetic]

I done the ipconfig flush ages ago, Didn't update. In the end I resorted to changing the DNS servers in Azure to point to Google and Cloudflare instead of Microsofts DNS and it updated instantly but the issue remains. I may rebuild the bride between PBX A and C to see if that clears the issue

Forget the bridge.


until you can access the management console, dont bother with anything else.

 

[johnporteous]

I'll revisit in the morning, it's very late here and it's been a very long day.

 

[Alphabetic]

Check Windows firewall on PBX A too.

 

[johnporteous]

PBX A is fine remember, it has a functional bridge with PBX B as master/slave.

 

[Alphabetic]

PBX A is fine remember, it has a functional bridge with PBX B as master/slave.

humour me.

 

[johnporteous]

LOL, sure thing. but it will be in the morning, I'm off to bed. Many many thanks for tomorrow, I'll continue tomorrow.

 

[johnporteous]

So I have checked the PBX A Windows firewall; it is still all off, as it is on PBX B and C. This was due to when installing the 3CX instance the 3CX firewall tests fail unless you turn off Windows Firewall. In Azure, the normal ports are open and match those open on the other PBXs.

Switching the DNS to Google/Cloudflare, restart, then switch back worked to resolve the NSLookup. But referring to your earlier point about access to the mngmt console, I still cannot browse to it nor can I ping it. However, none of the PBXs can ping one another, despite working bridges, internal calls, etc. I can ping other VMs/Services we host in Azure. As mentioned earlier in the thread it seems like a DNS issue but I'm darned if I can find it or a solution, especially when B & C Bridge functions correctly, as does A & B.

 

[SweetAction]

The management console port (5001 or 443) is used for status in Bridges (and other things).

If you can't load the management console then the status won't work. Forget ping - it's possible to block ICMP while allowing TCP/UDP on the correct port - focus on figuring out why you can't load the management console page from PBX C's network.

 

[johnporteous]

That's exactly what I'm doing, and it's driving me nuts LOL