-
Notifications
You must be signed in to change notification settings - Fork 0
Configuring a Registry
Daniel Lamando edited this page Feb 22, 2022
·
3 revisions
Registries are used to define "ownership" of the records managed on your DNS providers.
Your choice of registry can make it possible (or impossible) to share a DNS zone
with multiple kubernetes-dns-sync
installations.
There are two available registry types:
-
txt
defines thatkubernetes-dns-sync
will create its ownTXT
records to keep track of what other DNS records it's managing. You'll most likely want to use this when setting up an existing domain. Really, go fortxt
unless you have a reason not to. -
noop
defines that the DNS zones will be wholly owned bykubernetes-dns-sync
. Any and all unrecognized records that exist in the zone will be deleted. This can be useful for single-purpose domains.
If you will be connecting multiple DNS sync programs to one DNS zone,
make sure the txt_owner_id
field is unique for each instance!
See also: TXT Registry Behavior.
Annotated TOML block:
[registry]
type = "txt"
### An ID for our instance
txt_owner_id = "example"
### If you are migrating from external-dns, you can optionally "adopt" (aka steal)
### its records by adding its owner-id here:
# auto_adopt_from_owner_ids = [ "my-external-dns" ]
### To support wildcards, CNAME, or to hide the registry TXTs,
### set either a prefix *or* suffix:
# txt_prefix = "_registry."
# txt_suffix = "._registry"
### If you use one of these, set the same value on every attached dns-sync
There's no extra options, just add this block and forget it:
[registry]
type = "noop"
Don't forget to review the planned changes carefully when you add a your existing zones.