FreePBX 17 UCP Interface

[microchipmatt (Canada)]

Question: Is the error pictured in this screenshot:

A. Because the commercial modules have not been ported yet
B. Also because it is not licensed on the test server

I imagine one of those reasons applies.

[lgaetz (Lorne Gaetz)]

Neither of these things. It’s because either the UCP node server daemon is not running, or the route from the browser to the UCP Node service (default ports 8001 & 8003) is blocked. Any errors on the dashboard?

[microchipmatt (Canada)]

UCP Daemon is running, I have a hardware edge device, but this test FreePBX 17 beta VM is sitting on the same network as my machine behind it. No firewalls of any kind on this subnet between my machine and the freepbx box. By Default there is no Software firewall on Debian 12.5, so nothing should be standing in the way of UCP:

Internally my Computer and the FreePBX box should have free run of each other. The only thing running on this Debian 12.5 install is FreePBX 17. I’ve never had this problem before, usually on CentOS I would install, and everything just worked. It did install because I can get to it, but not sure why this is an issue.

Further testing, this is me going to it on the Debian machine it is installed on:

[microchipmatt (Canada)]

This is in the /var/log/asterisk/ucp_error.log. over and over. Usermgr was updated today:

2024-06-05 20:14 -06:00: }
2024-06-05 20:14 -06:00: There was an error with MySQL Connection
2024-06-05 20:14 -06:00: Error: connect ECONNREFUSED ::1:3306
2024-06-05 20:14 -06:00:     at TCPConnectWrap.afterConnect [as oncomplete] (node:net:1549:16)
2024-06-05 20:14 -06:00:  From event:
2024-06-05 20:14 -06:00:     at /var/www/html/admin/modules/ucp/node/node_modules/mariadb/lib/connection.js:141:13
2024-06-05 20:14 -06:00:     at new Promise (<anonymous>)
2024-06-05 20:14 -06:00:     at Connection.connect (/var/www/html/admin/modules/ucp/node/node_modules/mariadb/lib/connection.js:130:12)
2024-06-05 20:14 -06:00:     at Object.createConnection (/var/www/html/admin/modules/ucp/node/node_modules/mariadb/promise.js:38:17)
2024-06-05 20:14 -06:00:     at connect2database (/var/www/html/admin/modules/ucp/node/lib/freepbx.js:74:25)
2024-06-05 20:14 -06:00:     at /var/www/html/admin/modules/ucp/node/lib/freepbx.js:35:4
2024-06-05 20:14 -06:00:     at ChildProcess.exithandler (node:child_process:414:7)
2024-06-05 20:14 -06:00:     at ChildProcess.emit (node:events:517:28)
2024-06-05 20:14 -06:00:     at maybeClose (node:internal/child_process:1098:16)
2024-06-05 20:14 -06:00:     at Socket.<anonymous> (node:internal/child_process:450:11) {
2024-06-05 20:14 -06:00:   errno: -111,
2024-06-05 20:14 -06:00:   code: 'ECONNREFUSED',
2024-06-05 20:14 -06:00:   syscall: 'connect',
2024-06-05 20:14 -06:00:   address: '::1',
2024-06-05 20:14 -06:00:   port: 3306,
2024-06-05 20:14 -06:00:   fatal: true,
2024-06-05 20:14 -06:00:   sqlState: 'HY000'
2024-06-05 20:14 -06:00: }
2024-06-05 20:14 -06:00: There was an error with MySQL Connection

So when I first set this server up, I imported my backup config from another FreePBX server, would that matter…I;ve never never had that problem before…but just want to make sure.

[microchipmatt (Canada)]

Interesting, this appears to be my EXACT issue:

FreePBX 17 on Debian 12: IPv6 port binding for mariadb

Hi, I followed the instructions in How to Install FreePBX 17 on Debian 12 with Asterisk 20 (dated Nov 27, 2023) using the netinst image of Debian 12.4 and everything went really smoothly - thank you for the detailed walk-through. It case it helps anyone, I had to make (2) minor changes: Uninstall the FreePBX Firewall module, which fails to find a PHP include on Debian. It appears to be CentOS specific. I’m using ufw instead. Update the bind-address in /etc/mysql/mariadb.conf.d/50-serve…

Nope, that didn’t work either.

One other point of note, UCP server it CONSTANTLY restarting

That definitely does not seem right.

[pramarajan (Ramarajan P)]

It looks like there is an issue with IPv6, so you may need to enable precedence ::ffff:0:0/96 100 in /etc/gai.conf

[bug]: install script disables IPv6

[microchipmatt (Canada)]

Okay, I will try that and see if the problem is fixed and report back.

[microchipmatt (Canada)]

@pramarajan

/etc/gai.conf

The problem persists. Sigh. I have Installs of FreePBX 14, 15 and 16 on cent OS on the same network segment and the same subnet, that I am accessing from the same computer, and those all work. As stated earlier, no edge device or firewall in between them, The ones on CentOS work fine, the FreePBX beta UCP has stopped working since the last update as far as I know. Can anyone think of anything else. I;ve followed whatever articles I can find about this error.

[microchipmatt (Canada)]

reinstalled UCP with:
fwconsole ma uninstall ucp
fwconsole ma delete ucp
fwconsole ma downloadinstall ucp
fwconsole reload

I get to the login interface of UCP as I did before, but now I can’t even log in. So that probably wasn’t the right thing to do. Restored snapshot. I;m starting to think I should just wait for general release, but If I run into this again I;ll need to know how to solve it…I fell like its something small, but I can’t see it. Even worse UCP was running when snapshotted now after restoring snapshot UCP is not running.

[microchipmatt (Canada)]

So I nuked my test server, Created a new one, reinstalled FreePBX 17 Beta, with the newest install script. Script runs perfectly, no errors, FreePBX comes up, and the UCP Daemon is up, UCP is functioning. I import my backup config. My extensions ARE chan-sip, I let backup and restore convert my trunk and extensions to PJSIP, that goes well. But after Restoration is complete, the UCP Daemon no longer comes up. And UCP gives the SAME ERROR.

[kgupta (Kapil Gupta)]

Hi @microchipmatt I have raised [bug]: FreePBX17 - UCP daemon not running error · Issue #224 · FreePBX/issue-tracker · GitHub to track of your issue, could you please share the ucp logs “fwconsole pm2 --list” output in that jira.
thanks

[microchipmatt (Canada)]

I will indeed.

[microchipmatt (Canada)]

Sadly I thought this was fixed. It is not, there is still a problem with UCP. Hyper-V, Azure and ESXi all produce the same results above after my data restore. I have filed a ticket. There are probably some things I should have checked on Vultr I more than like missed. I keep hoping GA comes out and this issue won’t continue, but we’re close enough now I think it will be still. We’ll see what the ticket produces. Hopefully they can help me solve it on FreePBX 17 on Debian. As stated above FreePBX 16 works perfectly with my data restore. So just need to figure out what’s going on with FreePBX 17, my restore and UCP.

[microchipmatt (Canada)]

Okay, we have a solution!
After data import:

• Recreated the self-signed certificate in Certificate Manager and set it as the default. This fixes UCP and ALL errors
• For Fail2B I recreated any log files that are needed in jail.local on the system after data import. Fail2Ban starts at this point and Everything is working.

As a fix in FreePBX 17, and UCP, the development team is looking at Setting the newly created Certificate after data import, if there is only one, as the default.