Nevermind, tested by starting each Ignite pod 1 by one and then clients 1 by one, still fails after few minutes of run-time :(
There is something we're missing that happens in k8s setups I guess -- Sent from: http://apache-ignite-users.70518.x6.nabble.com/
