Hi, I found out that if I take out one server from node_admin backend, it works much better. Can someone please help shed some lights on this issue? Is it related to some session/cache issue?
Wei On 1/23/13 10:34 AM, "Wei Kong" <[email protected]> wrote: >We are getting a lot of 502 errors from haproxy on a server that is not >too busy > >The logged message shows > >Jan 23 18:31:01 preview2-stratus-admin-1 haproxy[28352]: >50.196.128.9:54443 [23/Jan/2013:18:29:01.649] node_frontend >node_admin/server1 6/0/2/-1/+120043 502 +204 - - SH-- 2/2/0/0/0 0/0 "POST >/log/pixgrid/NOSERVERID/2013-01-23T18:07:08.066Z/NODATETME/1/100 HTTP/1.1" > >What does SH mean here? > >Our haproxy config: > >global > daemon > maxconn 10000 > stats socket /var/run/haproxy.stat mode 666 level admin > >defaults > mode http > retries 3 > timeout connect 6000ms > timeout http-request 60000ms > timeout client 600000ms > timeout server 600000ms > option httplog > option logasap > log global > log 127.0.0.1 local0 > log 127.0.0.1 local1 notice > option httpclose > no option checkcache > > >frontend node_frontend > bind *:7000 > bind *:80 > acl is_admin hdr_beg(host) -i <adminhost> >use_backend node_admin if is_admin >default_backend default > >backend default > mode http > server server0 127.0.0.1:7002 maxconn 10 > >backend node_admin > mode http > server server0 127.0.0.1:7001 maxconn 150 > server server1 10.171.99.144:7001 maxconn 200
