New install - cloud hosted - no voice

[Gergorian]

I have a new 3CX install that I have hosted in the cloud. I've got things working but callers can neither hear nor be heard even on internal (between extensions) calls. I am assuming it is due to needing some ports opened but I'm not sure which ports and where to open/point them to since there isn't an on-premise system. Previously I've done local installs so I had to do some port forwarding to the system that was internally handling calls (both 3CX and FreePBX). Phones are connected via STUN and are being recognized as the extensions I've configured in the 3CX Dashboard.
They are using Ubiquiti Edge Router. Do I just need to open specific ports pointed to each phone or is there a "global" setting I can do to allow the phones to properly communicate?

Thanks for any assistance.
- Greg

 

[Evolute IT]

I don't think there's another firewall unless UFW is installed. You can check using "ufw disable".

At this point, create a new droplet with a new license and simply see if it works. Be sure to NOT add it to the firewall. You want a plain droplet.

I have a 3CX on DO and it works perfectly so it is strange. Also, what is the Region where you hosted your instance?

 

[Gergorian]

UPDATE: So a "clean" install the firewall tested fine. So I created another new install and restored a backup from my failing system and now THAT firewall tests fine. I did these exact same steps when setting it up originally so not sure where the firewall got messed up but at this point it seems to be working.
I can now call via the web interface for a client and can hear the caller but not sure about the users being heard yet. Phones have not found the "new" server (same FQDN, same license but new IP of course). Hoping having them power cycle the phones will get them to locate the new system and then we can test further.
- Greg

 

[cobaltit]

Power cycling won't change the TTL unless you control the DNS server those phones are using and can flush the cache there.

 

[Gergorian]

I can access the network, router is a Ubiquiti EdgeRouter. Not sure where but will see if I can find a command to clear it's DNS cache for the entire network.

 

[Evolute IT]

I can access the network, router is a Ubiquiti EdgeRouter. Not sure where but will see if I can find a command to clear it's DNS cache for the entire network.

Some routers don't do DNS and rely on either an external BIND or MS DNS or simply Google DNS like my network here. This way, DNS updates are much quicker than the basic ISP DNS.

 

[Evolute IT]

UPDATE: So a "clean" install the firewall tested fine. So I created another new install and restored a backup from my failing system and now THAT firewall tests fine. I did these exact same steps when setting it up originally so not sure where the firewall got messed up but at this point it seems to be working.
I can now call via the web interface for a client and can hear the caller but not sure about the users being heard yet. Phones have not found the "new" server (same FQDN, same license but new IP of course). Hoping having them power cycle the phones will get them to locate the new system and then we can test further.
- Greg

Sometimes with DO, it can happen. It might also be the region you select (if the new one is in another datacenter). It can also be simply the Debian install was corrupted after messing with it to troubleshoot. Those things can happen, that's why we backup everything!