On 2/20/20 8:36 AM, Pluem, Ruediger, Vodafone Group wrote: > > >> -----Ursprüngliche Nachricht----- >> Von: Eric Covener <cove...@gmail.com> >> Gesendet: Mittwoch, 19. Februar 2020 17:41 >> An: Apache HTTP Server Development List <dev@httpd.apache.org> >> Betreff: KeepAliveTimeout vs. event issue >> >> IIUC event MPM can close keepalive connections up to 100ms than the >> configured value. >> >> If other software parses the Keep-Alive response header and fudges the >> result by e.g. tens of milliseconds for its own TTL to avoid races, >> should we internally try to avoid these whole-second >> keepalivetimeouts? For example under event adding the 100ms back in >> so at worst the timeout occurs in the same timeout=X advertised? > > Just to be sure I get the issue correct. You say that with event a keepalive > connection might be kept open for x s + 100 ms if keepalivetimeout > is configured to x? > And the client thinks it is only kept open for x seconds as we announce > timeout=x in the keep-alive header? > > I was not able to find any current RFC that describes the keep-alive > response header. I only found what Mozilla documents here: > > https://developer.mozilla.org/en-US/docs/Web/HTTP/Headers/Keep-Alive > > And there they say that timeout is the *minimum* time the connection is > kept alive. So if we keep it open longer this would be fine. > Same thing is written in this draft: "A host MAY keep an idle connection open for longer than the time that it indicates, but it SHOULD attempt to retain a connection for at least as long as indicated." https://tools.ietf.org/id/draft-thomson-hybi-http-timeout-01.html#keep-alive
Regards Giovanni > Or is it exactly the other way around? Does event close it after > x s - 100 ms and hence earlier than announced? > > Regards > > Rüdiger >