https://issues.apache.org/bugzilla/show_bug.cgi?id=57520

--- Comment #8 from Yann Ylavic <[email protected]> ---
(In reply to Yann Ylavic from comment #6)
> But in this scenario there shouldn't be any "Connection reset by peer:
> proxy: error reading status line from remote server", which indicates that
> an established connection was *reset* by jetty with no response at all, no
> 200 nor 503...
> This error won't put the bachend in recovery state (the connect()ion succeed
> and there is no status to fail on), but with proxy-initial-not-pooled and a
> normal browser as client, this one will resend the request without notifying
> the user.

This is wrong, proxy-initial-not-pooled does not help the browser here, it just
prevent mod_proxy from reusing a kept alive connection on the backend side when
the request comes from new client connection (first one on this connection),
precisely because the client does not expect an error is this case (and reusing
a connection whose ttl is above backend's KeepAliveTimeTimeout may produce the
error).
Particularly, proxy-initial-not-pooled should not be used when ttl is lower
than backend's KeepAliveTimeTimeout since reusing backend connections is not an
issue and may even help fault tolerance as seen is previous message.

I mixed up with mod_proxy's behaviour (hard coded) on the client side when this
error occurs.
Either the request comes from a kept alive connection and hence the browser is
expecting that kind of error, so mod_proxy will respond with a closed
connection (no HTTP response at all) so that the browser can recover (resend
the request) without notifying the user.
Or the request is the first one on the client connection and httpd will respond
with a 502 because the browser can't recover (unexpected error).
Unfortunately this is not something which is controllable on httpd, hence this
errors are not *always* transparent for the user.

-- 
You are receiving this mail because:
You are the assignee for the bug.

---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]

Reply via email to