On Tue, Jan 27, 2009 at 07:09:47PM -0800, Chris Charman wrote: > I have a backup server that shows as down -- the server is reachable (via > wget/telnet) from the haproxy host. I'm using a simple "check." > > Any thoughts?
nothing that comes to mind. Is there a "source" parameter in the configuration, which might be inappropriate for this server ? Is the application on the server somewhat different than the one on the other servers ? Eg: it could expect a "host:" header while active servers would not care. Does the server respond fast enough for the "inter" parameter (which defaults to 2 seconds) ? A tcpdump trace of the request/response would be nice. Do you know since when the server was marked down and if this was related to a real failure ? Regards, Willy
