mik-laj edited a comment on pull request #13424:
URL: https://github.com/apache/airflow/pull/13424#issuecomment-753386669
> It looks like there is a race condition with docker compose
that causes services that started fast enough (before DNS)
to get a different reverse-DNS IP lookup (usually it is
just <SERVICE> but sometimes it is
<DOCKER_COMPOSE_APP>_<SERVICE>_1_<NETWORK>)
```
DNS fwd/rev mismatch: rabbitmq !=
docker-compose_rabbitmq_1.docker-compose_default
rabbitmq [172.25.0.8] 5672 (amqp) : Connection refused
sent 0, rcvd 0
```
I think we have two separate messages here. The first is a warning that
always occurs if one host have multiple DNS names and run netcat in very
verbose mode (``-vv``). The second is a real error message that contains
information about the problemr that happened.
Take a look at an example.
```bash
$ nc -zvv kdc-server-example-com 88
DNS fwd/rev mismatch: kdc-server-example-com !=
docker-compose_kdc-server-example-com_1.example.com
kdc-server-example-com [10.5.0.2] 88 (kerberos) open
sent 0, rcvd 0
$ echo $?
0
$ nc -z kdc-server-example-com 88
$ echo $?
0
```
This warning does not affect communication in most cases and is only
relevant in very restrictive network configurations, e.g. in the case of
Kerberos, which tries to validate the message source a little more.
So if we want to improve CI stability, this change will not help us in
anything.
----------------------------------------------------------------
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
For queries about this service, please contact Infrastructure at:
[email protected]
