Intermittent Failure of 3CX

[Alan Figgins]

Hi All,

This is a standard install of 3CX on Debian using the ISO download from 3cx.com.

I have a situation where users are reporting an intermittent failure of the phone system.

For example, this morning, it was working at, say, 9am, not working at 10am, then working again at 11am (times just made up for illustration).

I am assuming they are right, so how do I go about finding out what is going on? I have had a look at the activity logs. They seem quite impenetrable, but presumably if I get into it, I will be able to understand them if I know what to look for.

Any suggestions on where to start would be gratefully received.


Thanks,

Alan.

 

[cobaltit]

Without knowing what 'failure' constitutes it's hard to say. You also don't mention if this is a brand new install or an install that's been flawless for some time. Is is bare metal or VM? Is it local to the users or remote/hosted?

 

[Alan Figgins]

Hi cobaltit,

Good points - It is a new install, on bare metal, and is local to the users.

I have asked the users for more detail (exactly what happened in terms of incoming or outgoing, if outgoing, did they get a dial tone, if incoming, was the phone ringing but nobody was there on pickup etc) but I might not hear back from them until Monday.

Thanks,

Alan.

 

[cobaltit]

I've seen NTP issues on HyperV cause intermittent issues to bare metal rules that out. The event log should show you things like registrations which may give clues. Otherwise you'll have to wait for feedback to see what 'failure' involves. If it's the inability to call out or receive calls I'd have them test internal calls or even the echo test to narrow it down to a trunk issue perhaps.

 

[AWS2P]

When according to the customer they indicate that 3CX does not work ,in same time do they have their Local network working? access internet was up? No intermittent power down ?

 

[Alan Figgins]

Thanks guys - I will find out on Monday, and begin the toubleshooting as you have suggested.

Alan.

 

[CentrexJ]

Dial tone is meaningless on voip phones.

When the system is ‘failed’ can the phones call each other and/or the voicemail extension?

If so it means 3CX is up but the sip trunk or gateway is down.

 

[Alan Figgins]

Hi All,

The issue was entirely with inbound / outbound calls - they said internal worked fine, so I figured it was either the SIP Trunk / Gateway or a connection issue.

I ran some tests, and the internet connection was not stable - sometimes great, and sometimes very slow / laggy, so called the ISP who are getting the fibre line looked at.

Hopefully that will solve the problem.

While I wait to see / hear what happens, are there any tests that you perform on a client's internet connection prior to installation (if a new install) to make sure it is stable and fast enough?

I can do a standard 'SpeedTest' and would certainly do that, but that seems to me to be only a point-in-time test, whereas something more extended over a period of a week (say) would make sense as well. I am envisaging leaving a PC connected to whatever point we would be connecting our router (off which the 3CX Server and all the endpoints) would be connected, and leaving it doing something for a period of time.

Most of our clients have uncapped internet connections, and so do we, so total data throughput won't matter in and of itself while testing.


Thanks,

Alan.

 

[cobaltit]

Throughput is RARELY the issue, it's all about the quality. There's a number of tests you can do, some are free, some aren't. Just depends on the customer and how invested you want to get. You can do something like MTR or PathPing for free. You can buy a program like PingPlotter. You can use a paid service like VoIPSpear or you can get really in-depth with something like VxSuite.

 

[Alan Figgins]

Thanks cobaltit - I have a bit of research do do I guess, but presumably you would recommend doing at least the free options?

Alan.