HTTP1.1 connection pool and Keep-Alive header?

[Eric Covener]

Hello, I am investigating some issues related to the roughly simultaneous chrome reuse and server-side close of a HTTP/1.1 keepalive connection on an XHR POST.

Does chromium parse the Keep-Alive response header to setup a max TTL on connections put into the pool to try to avoid these races?  Any hints about where i could see this in the chromium source if there is anything like this?

I am familiar with this (https://bugs.chromium.org/p/chromium/issues/detail?id=377581) change where a 408 response in place of a connection closure can be used to help this kind of race, but for the moment I am focusing on how we might be able to avoid the race.

Thanks.

[Eric Lawrence]

As far as I know, no, Chromium does not attempt to control connection pool duration via a Keep-Alive value.

https://source.chromium.org/chromium/chromium/src/+/main:services/network/public/cpp/header_util.cc;l=50;drc=267b0b74a1bf120473aa3957667f814a43801e06

• By default, an idle socket (not waiting on a response) will live for five minutes awaiting reuse

[Eric Lawrence]

https://bugs.chromium.org/p/chromium/issues/detail?id=131960 says "By-design" that this isn't supported.

[Eric Covener]

Great find, thanks for both responses!

--
Eric Covener
cov...@gmail.com