I have discovered what was wrong with my cluster. I have analyzed logs on both machines and found something like this:
join validation on cluster state with a different cluster uuid LcitTe7KQFCEF8Ww_MuU_g than local cluster uuid or6tpJqPRqSzOdea_eYA7g, rejecting
According this link I have checked data path ..\elasticsearch\data\nodes\0 there were some data, I deleted them all, restarted ELK stack and cluster started working and changed status from yellow to green.