We are having an issue with some VoIP phones in one of our offices, they periodically fail to register, we can still ping them internally but cannot make calls etc. Forcing them onto a new IP reservation then rebooting fixes this. We have run DHCP from the router, server & set them IP's statically but seemingly makes no difference.
On running a few packet captures from the MX84 on the LAN interface, I can see traffic between one of the non-functional phones and the external IP of the VoIP provider, including some outgoing only SIP packets-no responses coming back in. However, on the WAN side capture I cannot see these outgoing SIP packets (nor any incoming) for that internal IP address.
On assigning a new IP and rebooting the phone and repeating the captures, SIP packets for that phone are showing on both the WAN and LAN interfaces as expected and the phone works.
The phone provider suspects a NAT issue, we have no 1:1 or 1:Many NAT mappings set up. It seems as if the outgoing SIP packets are reaching the router and not being forwarded out over the internet, however the phone can successfully make TCP connections. Also this issue is only periodic, and affects all phones which have to be manually assigned a new IP to get them working again.
Leaving me a bit baffled as to what is causing this and why the SIP packets are not exiting the router (or at least why I cannot see them). Any help would be much appreciated on this.
From your description, it is possible that the MX is dropping the packets before forwarding on the WAN. Our support team needs to take a look at the live flows and logs from the MX when the issue is happening. I would suggest opening a support case to work on this.
My initial thoughts indicate this could be due to out-of-order packets being sent on the MX LAN and causing the packets to be dropped. Let me know the MX firmware on which you saw this issue.
The out-of-order packets are dropped in any firmware 14.36 or later. That specific bug is fixed in 15.23 beta firmware. But, our support team can check the logs and confirm whether you are hitting this issue or not.