SBC keeps going down and up

[Evolute IT]

Hi guys,

We are experiencing a problem at a customer's site. For some reason, the SBCs we have on-site are going down and up every few minutes, but only if they have phones registered via them.

For example, we have 2 SBCs, .211 and .212. Both are 18.0.30, running on Debian 9. All installed more than a year ago.

.211 had 6 phones behind it and it was doing that. I swapped them all to use .212 and now, that one is having the issue but not the other one. (I see the uptime in the 3CX console.)

The issue has been only for the past 2-3 days, Internet connection on both sides is perfectly fine.

(PBX is latest v18 U1, running on Vultr installed via 3CX ISO.) (If it changes anything, all remote phones are Grandstream GRP2612P and GXP1630, all on latest supported firmware.)

 

[ChrisC_3CX]

Did the .212 SBC already have some phones provisioned to it and was working fine until you added the 6 phones from .211? How many IP phones in total are now on the .212 and how many BLFs in total would you say there are behind that same SBC?

Also, what are the specs of the SBC host, is it perhaps a RPi?

It does sound like it might be one or some of the 6 IP Phones causing this. If the SBC drops frequently enough, maybe you could start turning one device off at a time to hopefully determine which one is causing the issue. I'd recommend start turning them off one-by-one, and once the SBC stabilizes, start turning them on again (except the last IP phones you turned off) in order to determine if it's only one or multiple devices causing the issue. You would of course have to wait in between turning devices on or off.

 

[Evolute IT]

Did the .212 SBC already have some phones provisioned to it and was working fine until you added the 6 phones from .211? How many IP phones in total are now on the .212 and how many BLFs in total would you say there are behind that same SBC?

There are only 6 phones. Before that, the .212 had 0 phones. It was originally a HA SBC that we simply uninstalled and stuck to having 2 SBCs. We are unsure whether HA SBC is still a supported thing as we haven't heard from it. As for BLF, maybe 20-25 BLF total across phones.

Also, what are the specs of the SBC host, is it perhaps a RPi?

Atom 1.6Ghz, 2GB of RAM, 500GB HDD. It's not a Pi but those are our old machines we used for that. Many are running across our customers and we never had this problem.

does sound like it might be one or some of the 6 IP Phones causing this. If the SBC drops frequently enough, maybe you could start turning one device off at a time to hopefully determine which one is causing the issue. I'd recommend start turning them off one-by-one, and once the SBC stabilizes, start turning them on again (except the last IP phones you turned off) in order to determine if it's only one or multiple devices causing the issue. You would of course have to wait in between turning devices on or off.

This could be an option, but I am unsure why it would do that. I would need to do that after hours.

 

[ChrisC_3CX]

It was originally a HA SBC that we simply uninstalled and stuck to having 2 SBCs.

If this is the case, I recommend postponing what I mentioned before and go straight to re-installing the SBC from scratch first. This because the 3CX SBC HA installation has many dependencies in regards to packages meaning there's a good chance this instance has remnant configs and packages from the SBC HA possibly causing issues. Use the latest 3CX ISO which will also install Debian 10.

Use the same IP address to have all devices automatically connect and then completely disconnect and shutdown the .211 SBC to make sure it's not interfering in any way. Give it another shot and let me know what happens.

 

[Evolute IT]

We will! I will setup a new machine and install it on-site next week.

In the meantime, I actually splitted the phones between the two SBCs, and for now it seems to hold. GXPs on one, GRPs on the other.