v16 call drops

[ccs1]

Hey All,

we are getting a number of issues seemly since the v16 upgrade and wonder if anyone else is getting the same.

Setup
we host 3cx instances for clients in our DC, no specs, network / firewall have changed majorly (i.e. we havn't replaced equiptment everything is pretty much as its always been)
Most clients are connected with a SBC to run the onsite phones.

some of these clients are now reporting call drops / one way audio in the middle of a call this can be anything from a couple of seconds into the call to 5 minutes (or more), we have run the firewall checker (all passed) using a offical SIP Provider, handsets are Yealink T46. historically we had SBCs on shared services, we have now deployed dedicated SBCs (Windows NUC, i5 8 (or 16gb) of ram (whatever we had on hand) and SSDs) but are still having the issues.

we have tickets with Support but its taking a while and clients are getting frustrated.

We have experienced this in our own office with handsets (but not softphones) which would suggest its a networking/SBC "issue" somewhere but we've gone through everything with a fine tooth comb and as i said nothing (practically) has changed since v15.5

any thoughts welcomed

 

[poynter]

Finding a similar issue too with T46S dropping transferred calls or sending it to voicemail.

 

[tbilal]

Seeing the same behavior in my enviornment using softphone/web client. Firewall configuration was followed per their documentation. Looking to keep using v16.

 

[treefiddy]

Is this only behind the SBC?

 

[tbilal]

Is this only behind the SBC?

No it is not.

 

[treefiddy]

Ah ok. If it was I'd say to try not configuring the failover server in the setup and just rely on the DNS record. I've had similar problems and changing this seemed to fix the issue.

It seemed like the SBC was attempting to use the failover IP on slight packet drops; not adhering to the time limits set in the config.

 

[Chris Fella]

we have one client reporting this, but they are have had ongoing issues with their leased line provider and so we thought it was down to that.

Is there any info on a fix?