I tried reproducing this issue with several modified versions of aardvark-dns to see if I could narrow down the cause.
I could not reproduce this with upstream source at head, nor with upstream source at v1.0.3. I did reproduce this issue both with the debian packaged version of 1.0.3, and with upstream source at head with the debian trust-dns-0.21.patch applied and the corresponding dependency bump. I was also able to get better logging by starting the container with: $ podman --log-level=trace run -it --rm --name test1 --net test-net alpine When I reproduce this, with an nslookup for test1 in this container with trace logging, I see ~128k messages in syslog of the form: 2022-12-11T11:24:35.222884-08:00 salmon aardvark-dns[0]: reading A The messages end when the query fails. The log messages are different if I do an nslookup for a non-existent name, but also show a tight loop; worse, the log messages continue after the query fails until the container exits. My guess is we're using trust-dns 0.21 wrong, and it's causing us to loop repeatedly instead of returning a response? Thanks Duncan
