Skip to content

OCPBUGS-66104: Fine tune CoreDNS pod configuration to improve performance #5695

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.

Sign up for GitHub

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

Merged
openshift-merge-bot merged 1 commit into from
Mar 20, 2026
+1 −1
Merged

OCPBUGS-66104: Fine tune CoreDNS pod configuration to improve performance #5695

Changes from all commits
Commits
Show all changes
1 commit
Select commit
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
2 changes: 1 addition & 1 deletion templates/common/cloud-platform-alt-dns/files/coredns-corefile.yaml
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -4,10 +4,10 @@ contents:
inline: |
. {
errors
bufsize 512
health :18080
forward . {{`{{- range $upstream := .DNSUpstreams}} {{$upstream}}{{- end}}`}} {
policy sequential
force_tcp
Copy link
Contributor

@patrickdillon patrickdillon Mar 17, 2026

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

@sadasu Your PR description says "prefer udp", is that out of date and you ended up moving to force_tcp? Or should this be prefer_udp?

prefer_udp makes more sense to me (and gemini mentioned it is recommended on aws eks!) but I haven't been following this pr closely

Copy link
Contributor Author

@sadasu sadasu Mar 19, 2026

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Yes, prefer_udp was one of the first options we tried because that was the preferred solution during high load conditions. Testing revealed that the upstream servers reduced load by not responding to UDP requests.
We tried setting the max_concurrent to a value < 1000 which is the default value as a way to circumvent i/o timeouts due to UDP port exhaustion. But, during our testing we found that udp port exhaustion was not the cause of our i/o timeouts.
force_tcp was providing us with the best results.

}
cache 30
reload
Expand Down