Reverse proxy setup on haproxy gives 500; wget on base server & nginx works

Question

I am trying to setup haproxy as a reverse proxy for a server. I am on Centos.

The config goes like this:

global
   #log /dev/log local0
   #log /dev/log local1 notice
   log 127.0.0.1 local2 info
   log 127.0.0.1 local2 notice
   log 127.0.0.1 local2 debug
   chroot /var/lib/haproxy
   stats timeout 30s
   user haproxy
   group haproxy
   daemon

defaults
mode http
log global
option httplog
option dontlognull
option http-server-close
option forwardfor except 127.0.0.0/8
option redispatch
retries 3
timeout http-request 10s
timeout queue 1m
timeout connect 10s
timeout client 1m
timeout server 1m
timeout http-keep-alive 10s
timeout check 10s
maxconn 3000

frontend http_front
   bind *:801
   option forwardfor
   stats enable
   default_backend http_back

backend http_back
   mode http
   option httpchk
   option forwardfor
   http-send-name-header Host
   balance roundrobin
   server server1 stg-hostserv.com:80

But, if I do a wget against it, I am getting the below error.

# wget http://0.0.0.0:801
--2018-07-16 14:26:24--  http://0.0.0.0:801/
Connecting to 0.0.0.0:801... connected.
HTTP request sent, awaiting response... 500 Internal Server Error
2018-07-16 14:26:24 ERROR 500: Internal Server Error.


     haproxy  -f /etc/haproxy/haproxy.cfg -d
[WARNING] 197/200148 (13833) : config : frontend 'GLOBAL' has no 'bind' directive. Please declare it as a backend if this was intended.
Available polling systems :
      epoll : pref=300,  test result OK
       poll : pref=200,  test result OK
     select : pref=150,  test result FAILED
Total: 3 (2 usable), will use epoll.
Using epoll() as the polling mechanism.
00000000:http_front.accept(0004)=0006 from [127.0.0.1:60696]
00000000:http_front.clireq[0006:ffffffff]: GET / HTTP/1.1
00000000:http_front.clihdr[0006:ffffffff]: User-Agent: Wget/1.14 (linux-gnu)
00000000:http_front.clihdr[0006:ffffffff]: Accept: */*
00000000:http_front.clihdr[0006:ffffffff]: Host: 0.0.0.0:801
00000000:http_front.clihdr[0006:ffffffff]: Connection: Keep-Alive
00000000:http_back.srvrep[0006:0007]: HTTP/1.1 500 Internal Server Error
00000000:http_back.srvhdr[0006:0007]: Content-Type: text/html
00000000:http_back.srvhdr[0006:0007]: Server: Microsoft-IIS/8.0
00000000:http_back.srvhdr[0006:0007]: X-Powered-By: ASP.NET
00000000:http_back.srvhdr[0006:0007]: Date: Tue, 17 Jul 2018 12:02:00 GMT
00000000:http_back.srvhdr[0006:0007]: Connection: close
00000000:http_back.srvhdr[0006:0007]: Content-Length: 1208
00000001:http_front.clicls[0006:ffffffff]
00000001:http_front.closed[0006:ffffffff]
^C
[root@izp0w3tkx2yr8zhes26ajqz ~]#

I tried different config for the server and consistently hit 500 error. Wget to the base server works without any issues. I setup nginix to the same thing and it works beautifully. Just haproxy does not seem to work. The customer wants it on haproxy. :)

Can you please advise where I can look at to further debug. Appreciate your assistance.

You're not showing us logs or configuration that appear to match the test you claim to be making. Where does the double `//` in `GET //app_themes/base...` come from? What's in the HAProxy log and the back-end log? — Michael - sqlbot, Jul 16 '18 at 10:40
Sorry. My bad. I did a few tests and must have mixed up the logs. Here are the matching ones. — Prak_Rum, Jul 17 '18 at 12:02
Does your backend Microsoft-IIS/8.0 check `host` header? as you set `http-send-name-header Host`, so request from HAProxy to stg-hostserv.com:80 looks like `GET / HTTP/1.1 Host: izp0w3tkx2yr8zhes26ajqz` — nuster cache server, Jul 19 '18 at 01:01
Hi, Thanks for the response. That option is one of the last that I added. Even without that option, I got the same error. — Prak_Rum, Jul 19 '18 at 03:16
Can you check `curl http://stg-hostserv.com:80/ -H "Host: izp0w3tkx2yr8zhes26ajqz"` ? — nuster cache server, Jul 20 '18 at 00:43
Wow!! I got the same error.
500 - Internal server error.

There is a problem with the resource you are looking for, and it cannot be displayed. — Prak_Rum, Jul 20 '18 at 01:52
I tried changing the hostname and it did not fix. :( http-send-name-header stg-hostserv.com — Prak_Rum, Jul 20 '18 at 05:03
curl -v http://stg-hostserv.com/ -H "Host: stg-hostserv.com" #Works curl -v http://stg-hostserv.com/ -H "Host: dummy" #Does not Work curl -v http://0.0.0.0:801/ -H "Host: stg-hostserv.com" #HAproxy endpoint Does not Work with correct hostname :( — Prak_Rum, Jul 20 '18 at 05:20
If I change it to the below and pass the header as parameter it works.. curl -v http://0.0.0.0:801/ -H "Host: stg-hostserv.com" http-request set-header X-Forwarded-Host %[req.hdr(Host)] Anyway I can avoid passing the host name and set it in the config? Please advise. — Prak_Rum, Jul 20 '18 at 05:31
It worked when I set it up like this. http-request set-header Host stg-hostserv.com Thanks a ton guys. You are great. — Prak_Rum, Jul 20 '18 at 05:36

score 0 · Accepted Answer · edited Jul 21 '18 at 05:15

0

This update from nuster cache server helped solve the problem:

Does your backend Microsoft-IIS/8.0 check host header? as you set http-send-name-header Host, so request from HAProxy to stg-hostserv.com:80 looks like GET / HTTP/1.1 Host: izp0w3tkx2yr8zhes26ajqz

HAProxy worked when I set:

http-request set-header Host stg-hostserv.com

edited Jul 21 '18 at 05:15

Nisse Engström

4,738
23
27
42

answered Jul 21 '18 at 01:59

Prak_Rum

25
1
11

Reverse proxy setup on haproxy gives 500; wget on base server & nginx works

500 - Internal server error.

There is a problem with the resource you are looking for, and it cannot be displayed.

1 Answers1