# DNS Server State cleanup

**URL:** <https://discourse.haproxy.org/t/dns-server-state-cleanup/7847>\
**Category:** Help!\
**Created:** [July 19, 2022, 10:26pm UTC](https://discourse.haproxy.org/t/dns-server-state-cleanup/7847 "2022-07-19T22:26:32Z")\
**Posts on this page:** 1\
**Page:** 1

<div class="post-metadata">

**Author:** ![FrancisL](https://avatars.discourse-cdn.com/v4/letter/f/919ad9/32.png) [@FrancisL](https://discourse.haproxy.org/u/FrancisL)\
**Post date:** [July 19, 2022, 10:26pm UTC](https://discourse.haproxy.org/t/dns-server-state-cleanup/7847/1 "2022-07-19T22:26:32Z")

</div>

Hello,

We use HAPROXY with SRV records, it works great but we have a small issue where it supposed to have only 2 Backend active (can scale up to 4) but we only had 2 at the current time.

We use haproxy 2.4.17 and this is the behavior we observed.

We have the following server-template:

```auto
  server-template domain-x 4 "_tcp_.entrypoint.my-custom-discovery-dns.local" resolvers dns resolve-opts allow-dup-ip resolve-prefer ipv4 fastinter 10s rise 2 fall 2 on-error fastinter check inter 10000

```

Somehow we started to get some ‘errors’ on one of our system where some traffic were serving wrong ‘Website’.

The show stat unix command gave us the following:

domain-x-4\_back,domain-x-41,0,0,0,0,0,0,0,0,0,0,0,0,UP,1,1,0,0,0,410,0,1,861,1,0,2,0,0,L4OK,0,0,0,0,0,0,0,0,0,0,-1,0,0,0,0,Layer4 check passed,2,2,3,10.10.12.203:49156,http,0,0,0,0,0,0,0,0,0,0,0,0,0,1,-,0,0,0,  
domain-x-4\_back,domain-x-42,0,0,0,0,0,0,0,0,0,0,0,0,UP,1,1,0,0,0,103995,0,1,861,2,0,2,0,0,L4OK,0,0,0,0,0,0,0,0,0,0,-1,0,0,0,0,Layer4 check passed,2,2,3,10.10.21.180:49157,http,0,0,0,0,0,0,0,0,0,0,0,0,0,1,-,0,0,0,  
domain-x-4\_back,domain-x-43,0,0,0,0,0,0,0,0,0,0,0,0,UP,1,1,0,0,0,410,0,1,861,3,0,2,0,0,L4OK,0,0,0,0,0,0,0,0,0,0,-1,0,0,0,0,Layer4 check passed,2,2,3,10.10.12.157:49295,http,0,0,0,0,0,0,0,0,0,0,0,0,0,1,-,0,0,0,  
domain-x-4\_back,domain-x-44,0,0,0,0,0,0,0,0,0,0,0,0,UP,1,1,0,0,0,103488,0,1,861,4,0,2,0,0,L4OK,0,0,0,0,0,0,0,0,0,0,-1,0,0,0,0,Layer4 check passed,2,2,3,10.10.12.44:49160,http,0,0,0,0,0,0,0,0,0,0,0,0,0,1,-,0,0,0,

As we can see somehow we have 4 Servers, but there should be only 2 (valid).

When we run an ‘haproxy reload’ the following occurs:

socat /run/haproxy/haproxy\_1\_.sock - \<\<\< “show servers state” \> /var/state/haproxy/state  
… reload haproxy

What we are thinking maybe we could “run and” sed command ( to remove the “IP” and “Port” to be “-”) after the saving of the haproxy state…

IF ( the srv\_check\_status srv\_check\_result srv\_check\_health srv\_check\_state ) is not healthy ? aka (in a bad state??)

Not sure what would be the condition yet …

Basically is there a way to tell haproxy to 'revalidate if the IP are still part of the SRV Record ?

Because when i was doing an dig _tcp_.entrypoint.my-custom-discovery-dns.local only two ip address are showing up… but the server-state is kept with the other IP’ / Port and these IP/Port may get re-used by the newer docker launches…