Upgrade from 2.4.4. to 22.0x results in LAN traffic intermittently dropped for OpenVPN clients
- pfSense 2.4.4 AWS image with around 100 - 150 OpenVPN clients functions normally.
- After moving to 22.01 or 22.05 (be it an upgrade or re-deploy of the virtual machine), VPN clients experience traffic delays due to packet loss. This only seems to become noticeable when there are more than 10 VPN clients connected. The packet capture in the ticket was taken with about 95 clients connected, and it shows TCP retransmissions and duplicate acknowledgments between both the client filtered for (the 172.x.x.x address), and the domain controllers (the 10.x.x.x addresses), also on AWS. The loss is not constant, so there are sections of the capture which show no problems.
- Multiple upgrades and new deployments have been performed with the same result.
- Both importing the configuration and reconfiguring a fresh deployment has the same result.
- DCO is disabled.
- Status changed from New to Closed
I'm closing this. We've not been able to replicate it, we've not seen it with other customers, and with 23.01 around the corner, it doesn't make sense to put effort into an end of life release. Rather than attempt to upgrade spanning multiple years of upstream updates, the better solution is to back up the configuration file, reflash directly to the latest stable release, and import the configuration.