Currently running 1.5-dev19, very tricky to get a packet capture as it only happened in production. When I try to do synthetic testing replicate it in a staging environment I cannot get it to happen. When a backend does go down the layer 7 check shows a valid status code for a down host. The other thought if the agent is marking it as down, I know in 1.4.24 and 1.5-dev19 there was some flap detection work done due to crashes. I was unable to find any reference to it in docs, however, is there any flap detection if a node flaps so many times it marks it down for a period of time even if layer 7 checks start to pass?
On Mon, Feb 2, 2015 at 4:18 PM, Lukas Tribus <luky...@hotmail.com> wrote: >> Hey, >> >> I have ran into a odd scenario, where the backend is DOWN however the >> layer 7 checks are passing. I have included the check which we >> received. The haproxy setup is fairly simple using proxy protocol. I >> could only find one example of this issue here, however, no follow up >> was done on the reason: >> http://comments.gmane.org/gmane.comp.web.haproxy/10454 > > What release are you running? I guess we will have to see the actual > health check as a traffic capture, like Willy suggested in that thread. > > > Lukas > >