Description of problem:
In our DNS server, we have the entry for *.svc which points to an IP address which is not the part of OCP cluster and it is used for different application.
We have all the configurations of DNS in network config and DNSMASQ properly configured but route docker-registry.default.svc is not getting resolved to its svc's ClusterIP instead it is getting resolved to IP from upstream DNS. Due to this, the Upgrade is failing while checking the component health. for this we have to manually add the entry "server=/.svc/127.0.0.1/" in the dnsmasq config to make it work. Also using the complete route like: docker-registry.default.svc.cluster.local resolves to ClusterIP of svc.
Our query: why the route is going to upstream server instead of getting it searched by the search parameter in resolv.conf?
Is there any alternative fix for this other than adding the line to dnsmasq config?
Version-Release number of selected component (if applicable):
add the *.svc entry to upstream DNS
The OCP internal svc routefor ex: docker-registry.default.svc is not getting resolved to svc's ClusterIP
The route should be resolved by the search parameter and resolved to svc's ClusterIP
# curl -kv https://apiserver.openshift-template-service-broker.svc/healthz
HTTP/1.1 404 Not Found
<h2> 404 - File or directory not found. </h2>
# nslookup docker-registry.default.svc
The route is getting resolved to the upstream IP x.x.x.193 which is not the part of OCP cluster.