Failover IP based

[nicoonline]

Hello,


We are a sip provider and we provide 2 Proxies for redundancy.
We support IP Based trunks and we use DNS SRV to allow our customers to have better failover in case of problems.
I have tried several configurations on 3cx and I can't get rid of the 32 second delay before switching to the alternate proxy.


How can we expect an end user will wait 30 seconds and not hanging up. There are many ways to find out if the proxy is UP or DOWN.
If the SBC does not respond within 5 seconds, you can assume that it is down. You can also send OPTIONS packets to check if there is a response.


I also notice that 3CX does not respect the SIP RFC (regarding SRV requests). In case a SIP server does not respond, the packet should be sent to the second host immediately and not end the call and wait the next one.

How do you implement a correct failover in your installations ? Again this timeout is the biggest issue in my opinion. i have worked with at least 10 others vendors and i never saw this.

 

[YiannisH_3CX]

Hello @nicoonline

The 32 seconds delay comes from the RFC which states that to consider a server down that server must not respond for 32 seconds. Unfortunately that delay cannot be avoided.

 

[nicoonline]

Hello Yiannish, What is strange is that you are the only vendor who understands the RFC like this
Cisco, Avaya, Siemens, Aastra, Panasonic and others have understand that 32 seconds is not understandable by any end-users