Hi
I am having a strange error when sending requests back to haproxy. The
configuration is to have an additional layer of frontend/backend to
generate a unique request id for every request then redirect back to
haproxy to the real frontend.
This works fine in the the beginning, but after several minutes I started
to see 502 errors, most of the query still works, 502 only happens to about
1% of the request but it is already causing problems. Any idea what could
cause this?
The haproxy.cfg looks like this
frontend logger-in
mode http
bind *:80,*:8080
no log
timeout http-keep-alive 5s
acl existing-uniq-id req.hdr(X-Unique-ID) -m found
http-request set-header X-Unique-ID %{+X}o\ci:%cp_%fi:%fp_%Ts_%rt:%pid
unless existing-uniq-id
default_backend logger
backend logger
mode http
server localhost localhost:81 send-proxy
frontend jetty-in
mode http
bind *:80,*:8080
option httplog
option dontlognull
...
-------------------------------------------------
502 errors:
Feb 16 23:45:57 localhost haproxy[73824]: 10.23.0.36:49839
[16/Feb/2017:23:45:57.938] logger-in logger/localhost 0/0/0/-1/0 502 204 -
- SH-- 49/15/2/2/0 0/0 "GET request url HTTP/1.1"
Feb 16 23:46:54 localhost haproxy[73824]: 10.23.0.32:53487
[16/Feb/2017:23:46:54.990] logger-in logger/localhost 0/0/0/-1/0 502 204 -
- SH-- 62/21/4/4/0 0/0 "GET request url HTTP/1.1"
Feb 16 23:47:27 localhost haproxy[73824]: 10.23.0.46:57830
[16/Feb/2017:23:47:27.306] logger-in logger/localhost 0/0/0/-1/0 502 204 -
- SH-- 62/21/6/6/0 0/0 "GET request url HTTP/1.1"
Feb 16 23:47:31 localhost haproxy[73824]: 10.23.0.38:64621
[16/Feb/2017:23:47:31.377] logger-in logger/localhost 0/0/0/-1/0 502 204 -
- SH-- 69/25/8/8/0 0/0 "GET request url HTTP/1.1"
Feb 16 23:47:33 localhost haproxy[73824]: 10.23.0.32:53630
[16/Feb/2017:23:47:33.706] logger-in logger/localhost 0/0/0/-1/0 502 204 -
- SH-- 77/29/11/11/0 0/0 "GET request url HTTP/1.1"
--
Chungwei Yen
