SBC Cluster possible on Raspberry Pi 4 (Raspbian Buster)?

Status
Not open for further replies.

MSCO

Gold Partner
Advanced Certified
Joined
Nov 14, 2020
Messages
26
Reaction score
31
Greetings,

I've spend a couple of hours today on trying to get an SBC Cluster working with two Raspberry Pi 4's (4GB) but to no avail..

Is it even possible to have an SBC Cluster on Raspbian Buster (v10)?

As I see conflicting data in the guides:

• "be two (2) PCs or Raspberry Pis with the same Debian or Raspbian (9/Stretch) Linux version and updates."
• "A Raspberry PI 4 B4 or B8"
-> I tried installing Raspbian Strecht Lite on RPi 4 but it can't boot with it, as it requires a newer OS.

I have tried several things by following these topics:
- https://www.3cx.com/docs/installing-pbx-raspberry-pi/
- https://www.3cx.com/docs/sbc-high-availability-cluster/

A single SBC works perfectly fine, but when I configure SBC 1 as the first node and then run the script on the SBC 2, it just doesn't work.
The log that SBC 2 shows upon finishing the configuration show both IP's are [online] but when I do crm status or crm_mon it says there is no active cluster / waiting for it to get online.
(I don't have the exact logs in front of me, the Raspberry Pi's are in another location at the moment)

The 3CX management console shows the Virtual IP in the details of the SBC entry and the phones can connect fine.
However when I turn off SBC 1, phones can't connect anymore. SBC 2 does not take over.

I first did an installation by updating/upgrading everything on the Raspbian and tried it again without the updates, but it made no difference.

I've done all the things that the guides speak of:
• Set the same timezone
• changed hostname (& password)
• set them to a fixed IP (.253 & .254) on the same LAN
• ran the SBC script and used the same Authentication ID from Management console
wget https://downloads-global.3cx.com/downloads/misc/d10pi.zip; sudo bash d10pi.zip
• Installed the cluster from the last question from that script.
• Selected the correct "first" or "second" node.
• Entered the IP of the other SBC when required
• Added the Virtual floating IP and is not in use elsewhere on the network (.248)
• Added the default gateway & 8.8.8.8 as ping resource
• the summary shows correct data
• Did the same thing on the SBC 2 but of course by choosing it's the second node, entered the IP of the first SBC when required, used the same virtual floating ip, ...
• the summary shows correct data

If it's at all possible to install, I can get and post the logs, but I'll have to do another installation from scratch (reinstall raspbian etc) as they are currently installed as two seperate SBC's but the secondary/back-up serveraddress of our phones don't seem to work too well for this..

Thank you!
 
Last edited:
This is actually limited to Debian 9 stretch at the moment as stated in the guide

https://www.3cx.com/docs/installing-pbx-raspberry-pi/
  • be two (2) PCs or Raspberry Pis with the same Debian or Raspbian (9/Stretch) Linux version and updates.
We do not as yet have any information on when this is to become available for Debian 10
 
  • Like
Reactions: MSCO
Aha, thank you for the quick answer.
 
This is actually limited to Debian 9 stretch at the moment as stated in the guide

https://www.3cx.com/docs/installing-pbx-raspberry-pi/
  • be two (2) PCs or Raspberry Pis with the same Debian or Raspbian (9/Stretch) Linux version and updates.
We do not as yet have any information on when this is to become available for Debian 10
I have tried two times on Raspberry Pi 3B+'s (completely from scratch) but I have the same issue.
First time without doing apt update && apt upgrade, second time I did do that.

I have it setup as
node 1 = 10.11.0.11
node 2 = 10.11.0.12
virtual ip = 10.11.0.13 (This is shown on the 3cx Management console as well)

This is the output when finishing the first node:
sbc 1 result.png

This is the result of finishing node 2:
sbc 2 result.png

I have added the install log (redacted the 3cx.be domain)..
Really seems like it should just work with what I'm doing.

If you want some other log file, I can definitely add it here
 

Attachments

I've added the output of journalctl -l -u pacemaker.service -u corosync.service aswell if anyone wants to take a look and see what is going wrong :confused:
 

Attachments

Hi MSCO, try running crm commands as root or using sudo
 
Hi MSCO, try running crm commands as root or using sudo
I clearly need the weekend, it was indeed just that, thanks! :)
Now I know our existing Mitel phones (not provisioned or anything) don't work flawlessly with the failover. They need a restart before internal & incoming calls can be made (only external outgoing calls still work before restart), but I'm sure that's just because they aren't provisioned/supported.
 
Status
Not open for further replies.

Forum statistics

Threads
111,982
Messages
590,121
Members
164,909
Latest member
Jacob.Ive