This is tracking the upstream effort needed to deliver the feature request described below.

What's the feature?

Currently, if OVN is configured to reply for DNS requests, it always replies with a hardcoded DNS_DEFAULT_RR_TTL, with is 3600 seconds.  It should be configurable instead.

Why is it needed?

In a dynamic cloud environment workloads can come and go quickly.  Services can be re-spawned on different IP addresses.  But if other workloads may cache the DNS reply for up to an hour, they may have trouble connecting to the service that was moved to a different IP.  For example, default DNS TTL in OpenShift is just 5 seconds, which allows a very fast service fail-over to a new pod.

Who will benefit? 

OpenStack allows to use OVN DNS for VMs and services running in it.  This may be a good thing to have TTL configurable, so the connection issues can be avoided on VM fail-over.  Other environments may also find this functionality useful.