Clustered Vicidial Server Setup Not Connecting Calls 11.0.1
Posted: Wed Nov 15, 2023 1:57 pmHi,
We are migrating from a single server build to a clustered server build:
Single-Server Setup:
ViciBox_v11.x86_64-11.0.1.iso | Vicidial 2.14-892a Build 230726-0922 | Asterisk 16.30.0-vici | Single Server | No Digium/Sangoma Hardware | No Extra Software After Installation | Intel(R) Xeon(R) | CPU E3-1220 v3 @ 3.10GHz, 4 Core, 4 Thread, 8 GB RAM DDR3 1066/1333
Clustered-Server Setup:
ViciBox_v11.x86_64-11.0.1-md.iso | Vicidial 2.14-892a Build 230726-0922 | Asterisk 16.30.0-vici | Clustered Server | No Digium/Sangoma Hardware | No Extra Software After Installation | Web and Database Server: Intel(R) Xeon(R) E-2236 CPU @ 3.40GHz, 3400 MHz, 32GiB (2 x 16GiB) DIMM DDR4 Synchronous Unbuffered (Unregistered) 2666 MHz (0.4 ns) | Telephony Server 1: AMD Ryzen 5 PRO 4650GE with Radeon Graphics, 1666 MHz, 16GiB (2 x 8Gib) SODIMM DDR4 Synchronous Unbuffered (Unregistered) 3200 MHz (0.3 ns) | Telephony Server 2: AMD Ryzen 5 PRO 4650GE with Radeon Graphics, 1400 MHz, 16GiB (2 x 8GiB) SODIMM DDR4 Synchronous Unbuffered (Unregistered) 3200 MHz (0.3 ns)
Our clustered server is three servers and all of them are behind a Fortinet firewall with the Vicidial firewalls completely open. Services should be ported correctly.
During after hours testing, we changed our internal IP address for the clustered server setup to respond to our active SIP trunk and started testing calls.
For purposes of this post, I’ll use random internal IP addresses to represent our setup:
WEB01 192.168.0.15
DIALER01 192.168.0.16
DIALER02 192.168.0.17
We are using MicroSIP and have tried variation in the account settings, but revert generally to:
SIP Server: 192.168.0.15
DOMAIN: 192.168.0.15
Sometimes we can get a manual dial call out and it will ring a cell phone, but upon accepting the call, it will immediately drop. I don’t remember which scenario caused the outbound connection with the drop. All others would not ring out. The inbound attempts get a busy message.
Every server IP configuration in MicroSIP brings up the “Answer” popup, but upon clicking it, there is no success tone with the message “you are the only agent in this session,” as we would expect. It gets a drop beep and no number shows on the MicroSIP display panel, indicating the softphone is not connected to the dialer properly.
When not logged in as a test agent and while going to our campaign view in admin directly and trying a test outbound call, the call will go to a cell phone, show the proper caller id, but upon accepting the call it immediately drops and no sound transfers either way.
We restored the database from our single server build to maintain unique settings and our many agent logins. We otherwise followed the clustered server build steps and settings.
We notice that we can pick the server we want on the phone extensions themselves. This may be part of our MicroSIP connection issues that we haven’t tried troubleshooting with. However, we do not understand conceptually how picking a single Dialer IP as the phone Server IP would work when using the cluster to switch between available dialers during overflow. And, we can’t test until after hours and our window is limited.
Any help with settings would be appreciated.
What are our best resources for clustered settings for inbound and outbound call connections and MicroSIP softphone setup? What settings from a single server build could be causing our calls to not connect fully?
We are migrating from a single server build to a clustered server build:
Single-Server Setup:
ViciBox_v11.x86_64-11.0.1.iso | Vicidial 2.14-892a Build 230726-0922 | Asterisk 16.30.0-vici | Single Server | No Digium/Sangoma Hardware | No Extra Software After Installation | Intel(R) Xeon(R) | CPU E3-1220 v3 @ 3.10GHz, 4 Core, 4 Thread, 8 GB RAM DDR3 1066/1333
Clustered-Server Setup:
ViciBox_v11.x86_64-11.0.1-md.iso | Vicidial 2.14-892a Build 230726-0922 | Asterisk 16.30.0-vici | Clustered Server | No Digium/Sangoma Hardware | No Extra Software After Installation | Web and Database Server: Intel(R) Xeon(R) E-2236 CPU @ 3.40GHz, 3400 MHz, 32GiB (2 x 16GiB) DIMM DDR4 Synchronous Unbuffered (Unregistered) 2666 MHz (0.4 ns) | Telephony Server 1: AMD Ryzen 5 PRO 4650GE with Radeon Graphics, 1666 MHz, 16GiB (2 x 8Gib) SODIMM DDR4 Synchronous Unbuffered (Unregistered) 3200 MHz (0.3 ns) | Telephony Server 2: AMD Ryzen 5 PRO 4650GE with Radeon Graphics, 1400 MHz, 16GiB (2 x 8GiB) SODIMM DDR4 Synchronous Unbuffered (Unregistered) 3200 MHz (0.3 ns)
Our clustered server is three servers and all of them are behind a Fortinet firewall with the Vicidial firewalls completely open. Services should be ported correctly.
During after hours testing, we changed our internal IP address for the clustered server setup to respond to our active SIP trunk and started testing calls.
For purposes of this post, I’ll use random internal IP addresses to represent our setup:
WEB01 192.168.0.15
DIALER01 192.168.0.16
DIALER02 192.168.0.17
We are using MicroSIP and have tried variation in the account settings, but revert generally to:
SIP Server: 192.168.0.15
DOMAIN: 192.168.0.15
Sometimes we can get a manual dial call out and it will ring a cell phone, but upon accepting the call, it will immediately drop. I don’t remember which scenario caused the outbound connection with the drop. All others would not ring out. The inbound attempts get a busy message.
Every server IP configuration in MicroSIP brings up the “Answer” popup, but upon clicking it, there is no success tone with the message “you are the only agent in this session,” as we would expect. It gets a drop beep and no number shows on the MicroSIP display panel, indicating the softphone is not connected to the dialer properly.
When not logged in as a test agent and while going to our campaign view in admin directly and trying a test outbound call, the call will go to a cell phone, show the proper caller id, but upon accepting the call it immediately drops and no sound transfers either way.
We restored the database from our single server build to maintain unique settings and our many agent logins. We otherwise followed the clustered server build steps and settings.
We notice that we can pick the server we want on the phone extensions themselves. This may be part of our MicroSIP connection issues that we haven’t tried troubleshooting with. However, we do not understand conceptually how picking a single Dialer IP as the phone Server IP would work when using the cluster to switch between available dialers during overflow. And, we can’t test until after hours and our window is limited.
Any help with settings would be appreciated.
What are our best resources for clustered settings for inbound and outbound call connections and MicroSIP softphone setup? What settings from a single server build could be causing our calls to not connect fully?