[dnsdist] Marking downstream Unexplained

[Alejandro Adroher Mellado]

I solved the issue ... I think.

Let me explain it, just to know if has any sense for you.
On my Abuse server I had a hard 20 QPS limit for all the traffic, and when I forwarded the 200 QPS coming from Amazon Web Servers from my recursor to the abuse pool, I got these messages about up & down.

Now I've changed the 20 QPS hard limit to 20 QPS by IP, and now I'm trying to reproduce the issue and I cannot. Seems to be a resolved issue, but I'm not sure.

Maybe the health checks can be affected by a configuration like I had? 

Thanks a lot Remi.


-----Original Message-----
From: Remi Gacogne [mailto:remi.gacogne+dnsdist at powerdns.com] 
Sent: jueves, 7 de abril de 2016 13:48
To: Alejandro Adroher Mellado <alejandro.adroher at omniaccess.com>; dnsdist at mailman.powerdns.com
Subject: Re: [dnsdist] Marking downstream Unexplained

Hi Alejandro,

On 04/07/2016 11:14 AM, Alejandro Adroher Mellado wrote:
> Following this thread, when I create a rule to forward queries from 
> Amazon servers (200 QPS)to my Abuse Pool makes my abuse server become 
> unstable.
> 
> I have enabled maxCheckFailures=5 but the issue persist.
> 
> Anyone knows how to deal with this?

You can get more information on what makes the health check fail by setting setVerboseHealthChecks(true)

It won't fix your issue, of course, but it might give you a hint on what is going on. You could also capture the traffic, for example with tcpdump, to see if and how the backend responds to the health check.


--
Remi Gacogne
PowerDNS.COM BV - https://www.powerdns.com/