Comment 0 for bug 1884035

Hello!

We have a problem with default timeouts 50s at haproxy. When we create LoadBalancer neutron-lbaas-agent create an a process, for exmaple:

nobody 24826 0.0 0.0 62000 10808 ? Ss Jun17 0:50 | \_ haproxy -f /var/lib/neutron/lbaas/v2/6e10e798-1147-4edc-a502-2c92d410720f/haproxy.conf -p /var/lib/neutron/lbaas/v2/6e10e798-1147-4edc-a502-2c92d410720f/haproxy.pid -sf 972/v2/76145b16-6ebc-4b56-a74a-23ef2ae5af93/haproxy.pid -sf 217

When we go to container neutron-lbaas-agent and check file haproxy.conf:

$ cat /var/lib/neutron/lbaas/v2/6e10e798-1147-4edc-a502-2c92d410720f/haproxy.conf
# Configuration for 6e10e798-1147-4edc-a502-2c92d410720f
global
    daemon
    user nobody
    group haproxy
    log /dev/log local0 debug alert
    log /dev/log local1 notice alert
    maxconn 12000
    stats socket /var/lib/neutron/lbaas/v2/6e10e798-1147-4edc-a502-2c92d410720f/haproxy_stats.sock mode 0666 level user

defaults
    log global
    retries 3
    option redispatch
    timeout connect 5000
    timeout client 50000
    timeout server 50000

frontend ....

I don't understand which service or daemon generate this config, because changing configuration at /etc/kolla/haproxy/haproxy.cfg don't help generate new config with new timeouts.

I found temp solution: manually change timeouts (connect, client and server) and restart process.

But how I can automate this process?
I think, kolla-ansible has template for this, but I don't know where I can configure this, could everybody help me with it?