In our instances pipeline jobs stuck very often after restart. It seems like pipeline detects that it needs to continue and tries to start execution where was interrupted but nothings is executed, it only looks like it is executed from UI - computer is occupied by job, build is in running state, but it seems that it is in endless waiting cycle.

The interesting is that this state is probably saved persistently because I do not catch it every time, but when it is caught, I can restart jenkins as many times as I want and it will never  recover successfully (always stuck).

I believe that problem is somewhere in program.dat, but it is not easily readable as xml for human so I am not sure where the difference is. 

I did the same with the previous 2 runs and they were able to recover, but the third one did not.

A add the screen of the successful recovers (or unsuccessful - one ended by failure but it did not stuck, which is success in this case) and the screen of build with issue. Since it seems to be persistent (as I described) I archived jenkins home and add it as well and the jenkins war as well. Archived jenkins home contains the versions of pipeline as well.

Thank you for help!