Search results

Not a clue, the reality is though that hardware is very much a thing. The most I have had on a single box is 3000 concurrent calls. I had 750 using a cluster of three arm cubieboards probably 10 years ago. In the real world, none of these stats are meaningful. You can have one single client...

By default it uses UDP 16384 to 32768. That's the way you should leave it. The server would die long before you eve filled those. What will usually kill your server, long before max concurrent calls is CPS, calls per second. Just scale out as you need to.

As Adrian said, there are some linux tools but its always going to be very, very messy and unless you had stopped the disk being written to, you will lose stuff anyway. I don't even think the tools that do recover actually get the full filename.

No easy way to do this, do you not have backups?

I've just turned it off, it was a mistake my side. I'm trying to combat the increasing amount of spam, mainly from gmail users that we seem to be getting. I have created a group, verified members that has permission to post without moderation and I added you to it but unfortunately had left...

The 200OK has the contact: Contact: <sip:*97@wan.cloud.lan:5060;transport=udp> So the ACK will be being sent to wan.cloud.lan in this case and I expect that is not where you want it?

Homer does not capture RTP

I'm unsure but I know asking for rport indicates that the device is behind NAT

Try enabling rport on your device if you have it.

What will usually kill your server is calls per second. By the time that will become a problem, you'll have wheelbarrows full of money so you won't worry about it anyway, I look forward to the day when I reach that stage :) I have come across servers with 2500+ domains on them and they are still...

I don't think there is any way to reliably determine any of this. It depends entirely on what features a client is using. You could have one single client consuming the resources that 10 other clients consume together, it all depends on what they are doing.

Its basic networking 101 port address translation, google it.

I have only had a quick look but about halfway down that dialplan I see an Action Hangup, is that not going to hangup the call?

If you really do only need a p-asserted-id then that is easy, in the advanced gateway settings set the SIP CID type to pid then stop and start the gateway.

This looks better /usr/bin/fs_cli -x "reload mod_callcenter" Just try it, you'll work it out.

I too was in this position and instead of using fusionpbx and to simplify setup too I just went down the road of building an opensips server to handle it. We have some 2000 users on it now, its been going well over two years and has never skipped a beat. The server load is next to 0.00 too. I...

That should do you, get the full path to fs_cli Usually: /usr/bin/fs_cli Get it with: which fs_cli

Something along the lines of: fs_cli -x "reload mod_callcenter"

You do not need to exclude tables from the replication, see the following post, you can include local files that are local to the server to set the local ips: https://www.pbxforums.com/threads/fusion-on-google-cloud-compute-or-aws.4862/#post-19023

The first major thing there is that the profile should be set to external not internal. Whenever you make changes to a gateway you must stop it and start it again.