> I'm using service_loadbalancer from kubernetes > (https://github.com/kubernetes/contrib/tree/master/service-loadbalancer ) > . This program would re-spawn haproxy when it found a change of > upstream endpoints. > When service_loadbalancer starts, it runs haproxy -sf $(cat pidfile) > several times very quickly, and on that moment, haproxy -sf doesn't > take effect, there were many haproxy processes left.
Also see Lukas Lösche's reports and efforts: https://github.com/haproxy/haproxy/issues/48 http://thread.gmane.org/gmane.comp.web.haproxy/25105/ and downstream marathon-lb report: https://github.com/mesosphere/marathon-lb/issues/5 Workaround in marathon-lb with manual locking: https://github.com/mesosphere/marathon-lb/commit/83260fdf687c774064b54d3bb009f5b3a1d75c97 Regards, Lukas