Description of problem: In Azure cloud cluster, EVEN in fresh env without any user operation, KAS pod logs spam with many "http: TLS handshake error from 168.63.129.16:<port>: EOF" Version-Release number of selected component (if applicable): 4.4.0-0.nightly-2020-04-02-130551 How reproducible: Always Steps to Reproduce: 1. Launch fresh Azure IPI env successfully. 2. Without any user operation, check KAS pod logs $ oc logs -n openshift-kube-apiserver -c kube-apiserver kube-apiserver-xxia1-0403-tkfs8-master-0 --tail 1 -f Actual results: 2. EVEN in fresh env given NO any user operation, the logs spam frequently with many below errors: I0403 10:27:08.835307 1 log.go:172] http: TLS handshake error from 168.63.129.16:55477: EOF I0403 10:27:18.851322 1 log.go:172] http: TLS handshake error from 168.63.129.16:55548: EOF I0403 10:27:28.866113 1 log.go:172] http: TLS handshake error from 168.63.129.16:55623: EOF I0403 10:27:38.866170 1 log.go:172] http: TLS handshake error from 168.63.129.16:55712: EOF I0403 10:27:48.866219 1 log.go:172] http: TLS handshake error from 168.63.129.16:55811: EOF I0403 10:27:58.866360 1 log.go:172] http: TLS handshake error from 168.63.129.16:55897: EOF ... Expected results: 2. Because we announce supporting Azure cloud, we should do this as better as possible, thus above errors in FRESH Azure env should not be seen for customer user experience. Though there is bug 1753443#c3 closed, it suggested others to do something to fix the issue. Here we suggest the cluster itself can do something to fix the issue :) Additional info:
This is very probably a non-https LB health check set up by the Azure installer. Use the http port for that like gcp does.
*** This bug has been marked as a duplicate of bug 1828382 ***
This bug was reproduced with OCP 4.5.0-0.nightly-2020-05-20-235311 which installed on Disconnected UPI on Azure with RHCOS & Private Cluster. $ oc logs -n openshift-kube-apiserver -c kube-apiserver kube-apiserver-pdazqeci-0521-05210324-master-1 --tail 1 -f I0521 07:56:19.059098 1 log.go:172] http: TLS handshake error from 168.63.129.16:57818: EOF I0521 07:56:29.074520 1 log.go:172] http: TLS handshake error from 168.63.129.16:58155: EOF I0521 07:56:39.074489 1 log.go:172] http: TLS handshake error from 168.63.129.16:58527: EOF ...
> (In reply to Ke Wang from comment #3) > This bug was reproduced with OCP 4.5.0-0.nightly-2020-05-20-235311 which > installed on Disconnected UPI on Azure with RHCOS & Private Cluster. The UPI azure bug is separate > https://bugzilla.redhat.com/show_bug.cgi?id=1836016 > > > $ oc logs -n openshift-kube-apiserver -c kube-apiserver > kube-apiserver-pdazqeci-0521-05210324-master-1 --tail 1 -f > I0521 07:56:19.059098 1 log.go:172] http: TLS handshake error from > 168.63.129.16:57818: EOF > I0521 07:56:29.074520 1 log.go:172] http: TLS handshake error from > 168.63.129.16:58155: EOF > I0521 07:56:39.074489 1 log.go:172] http: TLS handshake error from > 168.63.129.16:58527: EOF > ... so closing this bug. *** This bug has been marked as a duplicate of bug 1828382 ***