forwarder cache

[Darren Ankney]

On Tue, Nov 29, 2022 at 5:27 PM Hamid Maadani <hamid at dexo.tech> wrote:
> If I comment out the stale config options, reload and query test.com, I just get this in logs:
> 29-Nov-2022 21:57:49.931 queries: info: client @0x7f325e5a2108 192.168.56.1#57660 (test.com): query: test.com IN A +E(0) (172.17.0.3)
> 29-Nov-2022 21:57:49.931 resolver: debug 1: fetch: test.com/A
> 29-Nov-2022 21:57:49.933 query-errors: info: client @0x7f325e5a2108 192.168.56.1#57660 (test.com): query failed (SERVFAIL) for test.com/IN/A at query.c:7375

That looks like, if the stale config options are removed, then NS1
can't get an answer from NS2 at all?  Or you are saying that's what
you get if NS2 isn't running and you query NS1 regarding test.com
without the stale config options?

> In my setup, I have separated the instances by:
> - using separate config directories (/etc/bind/ns1/ , /et/bind/ns2/)
> - using separate work directories (/var/cache/ns1/ , /var/cache/ns2/)
> - turning off PID file for NS2 (only one PID file exists, and it is for NS1)
> - separating ports the listen on (NS1 -> 53, NS2 -> 153)
> - separating control ports for rndc (953 for NS1 and 1953 for NS2)
> instead of chrooting them.
>
> For what I know, this should completely separate the instances apart, unless I'm missing something. Also, they are separate processes, so they do not share any memory unless there is something different in bind that I do not know about.
>

That seems reasonable to me.