Github user jerrypeng commented on the issue:
https://github.com/apache/storm/pull/2442
@revans2 Sorry for not being more clear about what I meant in regards to
using the heap instead of stack space. Currently, the algorithm uses recursion
to backtrack which means it is going to use space on the call stack. This can
cause a stackoverflow if the stack size is not large enough to handle
TOPOLOGY_RAS_CONSTRAINT_MAX_STATE_SEARCH. Thus, you might need to configure the
jvm stack space (-Xss) in Nimbus to appropriately match what is expected to be
set for TOPOLOGY_RAS_CONSTRAINT_MAX_STATE_SEARCH. What I am suggesting is that
if there is an iterative implementation of the recursive algorithm, we would
not need to put some many frames on the stack but in heap so that you don't
need to configure the stack space for the Nimbus Daemon. Not something
critical, but rather something to think about.
---
