-
-
Notifications
You must be signed in to change notification settings - Fork 12
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
dns: overflowing header size #1350
Comments
Spinned up a coredns instance (here: If I query coredns with something that results in very large responses, the response begins with "Truncated, retrying in TCP mode.", for example with the
Querying my Gravity instance directly does not result in the truncation and TCP retry, same output just without the truncation.
Now, if I alter the path to swap out Gravityfor our old (and working) DNS forwarder:
So there appears to be something not working correctly when using Gravity that works when substituting Gravity for dnsmasq. Let me know if I can provide logs, pcaps or whatnot. |
Got two RKE2-clusters. Internally RKE2 uses a CoreDNS-base service (
rke2-coredns-rke2-coredns
, image:rancher/hardened-coredns:v1.11.1-build20240305
) to provide name resolution for the pods. These, in turn, pass to the hosts name resolution for upstream resolving.Yesterday I switched the hosts name resolution over to Gravity, and this morning we had quite some problems in most of the RKE2 pods. The
rke2-coredns-rke2-coredns
pods were logging a lot of errors of this exact formatplugin/errors: 2 login.microsoftonline.com. A: dns: overflowing header size
Unfortunately I had to switch back from Gravity to the old resolution service, restarting the pods the error messages immediately disappeared.
It seems this host (
login.microsoftonline.com
) gives quite many answers, maybe the overflowing issue is related to this?For the moment I can't replicate the issue, the clusters are semi-production and cannot experiment with them further. But I'd imagine setting up a chain like coredns -> gravity -> upstream and then querying the coredns for
login.microsoftonline.com
might replicate the issue.The text was updated successfully, but these errors were encountered: