Opened 10 months ago
Last modified 10 months ago
#2617 new defect
nginx 1.22 - sending GOAWAY to client after 60s — at Initial Version
Reported by: | Owned by: | ||
---|---|---|---|
Priority: | minor | Milestone: | |
Component: | documentation | Version: | 1.22.x |
Keywords: | Cc: | ||
uname -a: | |||
nginx -V: | nginx-1.22 |
Description
Hi nginx community,
I wanted to about the changes happened from nginx 1.18 to 1.20
https://github.com/nginx/nginx/commit/0f5d0c5798eacb60407bcf0a76fc0b2c39e356bb
"With this change, behaviour of HTTP/2 becomes even closer to HTTP/1.x,
and client_header_timeout instead of keepalive_timeout is used before
the first request is received."
Usually 5G NFs establish two http2 connections, one used for requests and the other as standby that will be used when we reach max streams, so only keepalive http2 ping will be sent.
On older nginx say(1.18) above was achieved by using client_header_timeout and keepalive_timeout.
Now with the changes to 1.20,
could keep-alive timeout will help here to maintain the former behavior of nginx 1.18, where keepalive_timeout bring the connection if idle ( no control or data => no http2 ping)
Thanks