outage backend for haproxy

From: D Tiz <tizatron#gmail.com>
Date: Fri, 27 May 2011 12:29:52 -0700


Here is snippet from my haproxy config file.

Two backends, one for and outage page and one for the cdn.

listen http-pool 0.0.0.0:80

       mode http
       balance roundrobin
       option httpclose
       stats uri /haproxy-status
       acl is_outage always_true
       use_backend outage if is_outage
       default_backend cdn
backend outage
       mode http
       reqirep ^Host:\ haproxy-host-01 Host:\
foo.s3-website-us-east-1.amazon.com
       server s3 foo.s3-website-us-east-1.amazon.com
backend cdn
       mode http
       server s3 10.10.222.175:8555 check inter 3000 rise 2 fall 3 maxconn 10
       server s3 10.10.222.209:8555 check inter 3000 rise 2 fall 3 maxconn 10

In my test client if I enter haproxy-host -01, the client is served content from the s3 bucket in amazon. Yeah!

But it I enter, haproxy-host-01/foo/bar/gore, then I get an error.

Is there something I can add to to the backend to make sure I land on the outage page at the correct location? I am assuming that /foo/bar/gore is passed on to amazon.

Trying various incantations of reqirep, no joy. Received on 2011/05/27 21:29

This archive was generated by hypermail 2.2.0 : 2011/05/27 21:45 CEST