Hi Chris,
> We are running 1.4.24 for an application that sees almost entirely small > http requests. We have the following timeouts: > timeout client 7s > timeout server 4s > timeout connect 4s > timeout http-request 7s > > There are a significant number of cR/http-408 responses in the logs. > The docs say that should represent timeouts from "timeout http-request". > My expectation was that in those cases the total request time would be > equal to "timeout http-request". However, a full logging trace shows > some are shorter, and many are longer, than the timeout. We had a similar report 2 months ago against 1.5-dev19 in the ML thread "http responses randomly getting RSTs" [1], Willy also posted 2 debug patches against v1.5 [2], but we never got any results back. If you are willing to do some troubleshooting, you could upgrade to 1.5-dev23, confirm that the issue is still there and try the patches. If you decide to do this I suggest your read that whole thread first. On the other hand a quick suggestion would be to upgrade to latest stable 1.4.25, which contains some important fixes, although I don't see anything related to what you are reporting. Regards, Lukas [1] http://marc.info/?t=139273066200005&r=2&w=2 [2] http://marc.info/?l=haproxy&m=139292555930452&w=2
