-
Bug
-
Resolution: Unresolved
-
Normal
-
None
-
4.17.z, 4.16.z, 4.18.z, 4.19.z, 4.20.z
Description of problem:
The Cluster DNS operator provides since 4.14 cache overwrites which are not working as expected.
Version-Release number of selected component (if applicable):
OCP 4.16.46 OCP 4.17.37 OCP 4.18.25 OCP 4.19.12
How reproducible:
everytime
Steps to Reproduce:
1. edit the dns.operator with cache setting 2. trace upstream dns logs for a particular query 3. query inside the cluster the dns forwarded resource record
Actual results:
every 30 seconds, an upstream DNS Server get a request for the resource that should be cached resource record TTL is always 30 seconds which indicates the seen behavior
Expected results:
the timeout settings from the Operator shall be honored. resource record TTL shall be adjusted accordingly to streamline the configured timeout setting
Additional info:
Upstream patch is already created https://github.com/coredns/coredns/pull/7602
The Cluster DNS Operator can wait for the PR to merged and backported or alternatively implement response type specific cache settings by utilizing TTL and minimum TTL
# current dns.Operator overwrites
cache <ttl> {
denial <maxrec> <ttl>
}
# propsed dns.Operator overwrites
cache {
success <maxrec> <ttl> <minttl>
denial <maxrec> <ttl> <minttl>
}
The resource specific overwrites can take the minimum TTL which will end in resource record responses as increased/lowered TTL for the record and remove the burden of overloading upstream DNS queries.