rpc server executor reverted to blocking

The threading executor consumes allot of memory, as the max thread count
is 5 x cpus, and this cannot be limited.
Blocking or threading doesnt make allot of difference for our needs,
as api parallelism is blocked by the GIL.

Change-Id: I10eab99204c779d93590552c2eb4b3ed207975d7
This commit is contained in:
Idan Hefetz 2018-05-23 08:35:08 +00:00
parent 9cc98e8e46
commit 1262783cca

View File

@ -107,6 +107,6 @@ def get_server(target, endpoints, transport, serializer=None):
return messaging.get_rpc_server(transport, return messaging.get_rpc_server(transport,
target, target,
endpoints, endpoints,
executor='threading', executor='blocking',
serializer=serializer, serializer=serializer,
access_policy=access_policy) access_policy=access_policy)