Hi Gurus,
I am looking to develop a discovery service feature on top of nginx.
I have read Redirecting… and I have a couple
questions
based on what I’ve seen in the code base.
If I have multiple workers configured using a single servers as an
endpoint.
I want to leverage the ngx_http_upstream_init_main_conf functions to do
the connection pooling for my service.
I see there is some mutex locking commented out in the code. What I want
to confirm is if I have 2 or more workers configured, each worker has
its own memory allocation for ngx_http_upstream_main_conf_t *umcf =
conf; Thus we don’t have to do any mutex locking between the separate
worker processes. Each worker will have a pooled set of connections to
the same endpoint.
Is my understanding correct?
Any help is greatly appreciated.
Charles