Solved Unstable SBC

Status
Not open for further replies.

eddv123

Titanium Partner
Advanced Certified
Joined
Aug 15, 2017
Messages
2,855
Reaction score
478
I have a single SBC on a site which is up but disconnects on regular intervals of up to 30 minutes. I believe it is a networking issue (firewall or local internet provider) but the error messages I am getting a a little inconclusive so I would like to know what they mean.

The onsite uses a 3rd Party IT provider which is also restricting me to what I can look at as a reason for issues onsite.

Both 3CX and the SBC are on Windows (Windows firewall is off, no anti virus software running, it is up to date on .NET Framework and Windows updates - which auto update is turned off also).

Entries in the SBC logs include some lines which I have not seen before including:

TCP transport connection error (10038): An operation was attempted on something that is not a socket.

Followed by the standard: Bridge failure 'Too long inactivity' on TCP connection: while processing TCP
https://www.3cx.com/community/threa...ty-on-tcp-connection-every-few-minutes.56251/


Searching for the first error on here does not produce anything so this is a first it would seem.

System error 10054: An existing connection was forcibly closed by the remote host.' on TCP connection: secure read

Seeing this also, but not too regularly, would be useful to know what it means however.

I am running tests on the local firewall/network for Latency and packet loss but would be grateful for any additional thoughts. I can provide full logs but not posting on a public forum for obvious reasons.
 
Hi Ed,

Wondering what version Windows this is running on and what build of windows.

Also SBC version please if you can share
 
@JohnS_3CX

Its a version 16 SBC and system, the Windows machine is Windows 10 PRO.
 
From the errors mentioned it might be a couple of things I can think of, either something else is installed on the machine messing up the traffic (network filter driver of some sort - traffic shapers, antivirus filters, capture software etc), or the connection of the machine to the internet is very bad/compromised/unreliable due to something after the PCs NIC port. It could also be corruption on the OS, but I have not data to support that.

If you can do so though, install the SBC using the Debian ISO to eliminate any Windows related messiness (if you don't wanna hunt it down and save some time). That would leave the rest of the path beyond the PC as possible points of failure.
 
A couple of things were the cause of this issue - both network I am happy to say.

After discovering the client site had miraculously changed the network setup without prior consultation (questions asked after traceroute and a bit of scanning from the SBC machine) 2 extra layer 3 devices had been added between the SBC and edge router (link to the cloud PBX).

At least one of these was closing the TCP port used for the tunnel as well as causing DNS related problems: Connection error Could not contact DNS servers for _3cxtunnel._tcp

As most of us know if the SBC cannot get DNS resolution the tunnel drops. Resolved by moving the SBC back to the network edge.
 
Glad to hear it was resolved Ed, thanks for updating the thread with your solution
 
Always a network problem - never 3CX :)
 
Status
Not open for further replies.

Latest Posts

Forum statistics

Threads
111,934
Messages
589,822
Members
164,813
Latest member
divdigital