Description of problem: Checked the TTL for *.apps DNS record on different Platform and found that it is 0 second on Azure, but 60 seconds on AWS platform and 600 seconds on GCP platform. Using 0 second for TTL may cause potentially performance issue. I have found sometimes it failed to resolve name when curling a route repeatedly on Azure platform. Version-Release number of selected component (if applicable): 4.2.0-0.nightly-2019-08-31-102058 How reproducible: 100% Steps to Reproduce: 1. install 4.2 cluster 2. dig console-openshift-console.apps.hongli-az058.qe.azure.devcluster.openshift.com Actual results: $ dig console-openshift-console.apps.hongli-az058.qe.azure.devcluster.openshift.com ; <<>> DiG 9.11.9-RedHat-9.11.9-1.fc30 <<>> console-openshift-console.apps.hongli-az058.qe.azure.devcluster.openshift.com ;; global options: +cmd ;; Got answer: ;; ->>HEADER<<- opcode: QUERY, status: NOERROR, id: 35106 ;; flags: qr rd ra; QUERY: 1, ANSWER: 1, AUTHORITY: 0, ADDITIONAL: 1 ;; OPT PSEUDOSECTION: ; EDNS: version: 0, flags:; udp: 4096 ;; QUESTION SECTION: ;console-openshift-console.apps.hongli-az058.qe.azure.devcluster.openshift.com. IN A ;; ANSWER SECTION: console-openshift-console.apps.hongli-az058.qe.azure.devcluster.openshift.com. 0 IN A 52.141.217.141 ;; Query time: 52 msec ;; SERVER: 10.72.17.5#53(10.72.17.5) ;; WHEN: Mon Sep 02 11:03:25 CST 2019 ;; MSG SIZE rcvd: 122 Expected results: Use 60 or 300 seconds TTL for *.apps record. Additional info: Not sure if 60s or 300s is better, but we should keep the same default value in all platform if possible.
Great catch, thanks for the report. We've chosen 30s as the new default, and existing records should get migrated during an upgrade (except for AWS, for which we don't control the TTL in this case — TTL for alias records in Route53 are not configurable).
verified with 4.2.0-0.nightly-2019-09-08-180038 and issue has been fixed. $ dig console-openshift-console.apps.hongli-az038.qe.azure.devcluster.openshift.com ; <<>> DiG 9.11.9-RedHat-9.11.9-1.fc30 <<>> console-openshift-console.apps.hongli-az038.qe.azure.devcluster.openshift.com ;; global options: +cmd ;; Got answer: ;; ->>HEADER<<- opcode: QUERY, status: NOERROR, id: 37063 ;; flags: qr rd ra; QUERY: 1, ANSWER: 1, AUTHORITY: 0, ADDITIONAL: 1 ;; OPT PSEUDOSECTION: ; EDNS: version: 0, flags:; udp: 4096 ;; QUESTION SECTION: ;console-openshift-console.apps.hongli-az038.qe.azure.devcluster.openshift.com. IN A ;; ANSWER SECTION: console-openshift-console.apps.hongli-az038.qe.azure.devcluster.openshift.com. 30 IN A 13.86.100.124 And TTL also has been changed to 30s on GCP platform: $ dig console-openshift-console.apps.hongli-gcp038.qe.gcp.devcluster.openshift.com ; <<>> DiG 9.11.9-RedHat-9.11.9-1.fc30 <<>> console-openshift-console.apps.hongli-gcp038.qe.gcp.devcluster.openshift.com ;; global options: +cmd ;; Got answer: ;; ->>HEADER<<- opcode: QUERY, status: NOERROR, id: 293 ;; flags: qr rd ra; QUERY: 1, ANSWER: 1, AUTHORITY: 0, ADDITIONAL: 1 ;; OPT PSEUDOSECTION: ; EDNS: version: 0, flags:; udp: 4096 ;; QUESTION SECTION: ;console-openshift-console.apps.hongli-gcp038.qe.gcp.devcluster.openshift.com. IN A ;; ANSWER SECTION: console-openshift-console.apps.hongli-gcp038.qe.gcp.devcluster.openshift.com. 30 IN A 35.225.130.254
Since the problem described in this bug report should be resolved in a recent advisory, it has been closed with a resolution of ERRATA. For information on the advisory, and where to find the updated files, follow the link below. If the solution does not work for you, open a new bug report. https://access.redhat.com/errata/RHBA-2019:2922