Willy, Our agent currently replies with:
drain\n stop\n 50%\n etc. and its not in production yet (looking forward to it being merged), so no problem to change if you prefer a different method. http://blog.loadbalancer.org/open-source-windows-service-for-reporting-server-load-back-to-haproxy-load-balancer-feedback-agent/ Thanks again. On 25 November 2013 06:59, Willy Tarreau <[email protected]> wrote: > Hi Simon, > > On Mon, Nov 25, 2013 at 10:46:31AM +0900, Simon Horman wrote: >> Hi Willy, >> >> please consider applying this series which adds support for a secondary >> health check and removes support for the recently added lb-agent-chk. >> >> This series is a revised subset of >> "[PATCH v6 00/22] Agent Check Enhancements and External Check". >> It builds on patches from that series that have already been applied. > > That's all clean all perfect, I have merged them. > > I have one comment about the protocol used by the agent, which probably > is more for the lb.org guys. I'm seeing in the check parser that we have > a number of strncasecmp() and some difficulty to find the end of the > string. I suspect we specified the protocol as "returning string X or Y" > without clearly enforcing a line feed at the end (LF or CRLF). > > If compatible with deployed software, we should enforce a complete line > in the spec, otherwise it will be complicated to extend the protocol in > the future. In my opinion agents will mainly be scripts running in xinetd > doing "echo 100%" and I really doubt some people will use "echo -n 100%" ! > > Thanks for this great work! > Willy > -- Regards, Malcolm Turnbull. Loadbalancer.org Ltd. Phone: +44 (0)870 443 8779 http://www.loadbalancer.org/
