tree bac5424d9d3cf070faa804f2dedfab3c8d8ad161
parent 1313ee3d9029cfd7fa0ee85fb4a767ae1b9ed8c4
author markap14 <markap14@hotmail.com> 1624639249 -0400
committer GitHub <noreply@github.com> 1624639249 -0400
gpgsig -----BEGIN PGP SIGNATURE-----
 
 wsBcBAABCAAQBQJg1gcRCRBK7hj4Ov3rIwAAzCYIAEPuVBH5bV7CVUN7pn4UmBv7
 61Y+fNS/rA2ZMf2JZ/VFZmHeKVZo2FlqdLJ81tW9rbXcg91EX3PbZaO5APFUrZg0
 1+xkCc7RlRtG+m2cMRe8pU/JtSJx1PQTn0j9UqHji+9uXcNITf398t1I5LgWVwzy
 di1w9iYOgZfjqw6ihGuxkDBqtw4cK2x/qwAo4FUUJCz3FxDQHCbVslbVTWNF8Ehk
 RoHXqnB+dhHzMAUXbjx3W8MkLvUFQLXKamf6OTpBy7IywGlzNtyNWBZiQd59x64v
 UaCAOxSkPnCZj0el9ZUCfFyzGmwIwNuklfvQjqbO1DBcHIO15rev3A7H0uwo6tQ=
 =r6Jp
 -----END PGP SIGNATURE-----
 

NIFI-8731: If a processor is running but made invalid due to a parameter update, it ends up in a state of STARTING. If the parameter is then updated again, it transitions the state to STOPPING but the processor is not fully stopped yet. At that point, the parameter is updated and the processor is attempted to be started again. Fixed this by keeping the number of active threads to >= 1 if processor is STOPPING in order to convey that it is not fully stopped. Also addressed a few minor bugs discovered in the process: when stopping a processor, if status == invalid, it should be skipped instead of waiting for the status to become stopped since it never will be. In the DTO's run status use Stopped instead of Invalid if there is at least 1 active thread / if stopping but not stopped. When considering if a processor has transitioned to the desired state for parameter updates, do not consider validation status if still transitioning to stopped or if the desired state has already been reached. Added new system tests to verify behavior. (#5180)

This closes #5180