Counter ping from different subnets does not work at the same time. Have you met this?

Question

Computer networks

Situation: two networks with different addressing, each has a gateway on two different interfaces on Kerio Control 8.5.3 and each network has a client.
I create a rule in Kerio that the traffic between interfaces is possible.
And I observe the following picture: clients can contact each other, application software on each machine sees each other, but the simplest PING does not work simultaneously in both directions. In turn - it works (and then, after a few minutes of a break), but simultaneously launched on two machines towards each other - no.
As a result: when the machines are on the same network, simultaneous counter ping works, and when the machines are on different networks, then no.
Checked tracert - yes, the same situation. Simultaneous launch - the result is only one (randomly selected machine), and tracert launched in turn - both requests are successful.
I see this for the first time, and you?
Maybe it's the features of routing on Kerio Control? Checked on versions 8.2 and 8.5.
Or is this generally normal?
Experienced networkers, tell me!

Answer 1

It became interesting, it never occurred to me to do a counter ping. Checked on different VPNs (PPTP and OpenVPN) and different systems (Linux, WinServer) in all possible cases. Pings go in all opposite directions without jambs.
Never worked with Kerio Control, because. I state that the gateway on Windows is bad manners, but it seems to me that it is in it.

Answer 2

Cisco equipment gives the same result when pinging from different subnets

Answer 3

IMHO an unambiguous bug with all symptoms.