hello all

i have been taking a look to the rtpengine.c code and i found that the issue has relation with the queried_nodes_limit which had 50 defalut value, so before it was reached in the loop, we received a request/response retransmision doing all the same.

anycase, the issue it's because when allow_op is applied, the select_rtpp_node gets a node even having the node disabled by the first attemp (having the rtpengine node down), and it does the loop.
if it's not applied, we dont get a node at the retry.

but adjusting the number of queried_nodes_limit to a value to be reached before a request retransmission arrives, i guess it should be enough (¿?)

best regards
david escartin