Hi Pierre-André, On Mon, Mar 24, 2025 at 08:51:51AM +0000, psavalle wrote: > > I'll have a look but need to be away from keyboard for an hour now, > > I'm sharing this just in case you have an idea. I'm running with > > HAPROXY_TEST_TIMEOUT=400 if that can help (the 504 above makes me > > think it could be related but don't have the time to test otherwise > > now). > > Ah, yes, to make sure that both requests are concurrent in the regression > test, > I used a 'delay 0.5s' statement in the server body which would fail with a > timeout > of 400ms.
Yeah I see. But we try hard to avoid delays everywhere because delay-based tests simply only work on a single machine in practice, randomly fail on others' machines and almost always on the CI :-/ > Here's an updated patch which uses 'barrier' instead. I've tested it a bunch > of times with `HAPROXY_TEST_TIMEOUT=400` instead of the 5s default. Unfortunately I couldn't get it to work even a single time :-( I tried hard to fiddle with barriers as well to try to improve the serialization but didn't manage to force it to proceed like we want. Pretty frustrating. What I tried was to make sure that c2 starts only once the two c1 were sent, and that the s1 server only starts to respond after c2 gets their response from s0. But for now I failed miserably :-( If I don't manage to get them to work, I propose you to merge everything but tag the VTC as "broken" (we already have a few such) so that they don't run by default. It's too bad to delay the inclusion of a feature just because we're unable to express it in vtc :-( Thanks, Willy
