Description of problem: kubernetes: Only set websocket protocols if we actually have them Reported upstream in https://github.com/cockpit-project/cockpit/issues/5347 Version-Release number of selected component (if applicable): 122-1 and kubernetes 1.4 How reproducible: Always Additional info: Fixed upstream in a5f781c6bd9ab1256a19648a792a47be79b28e71
Upstream master commit https://github.com/cockpit-project/cockpit/commit/d220e21d0eb75ca2361d77393d6840167001298b
Changing to OCP product. This doesn't affect Cockpit as released in RHEL ... only the registry container and its use of WebSockets.
Fixed in cockpit 122-3 registry container image: openshift3/registry-console:3.3-5
Hi Dominik, Any instructions on how to check this bug on registry console? The origin issue happens on getting pod container logs but I have no idea how to check on registry console.
I should check it could get container logs successfully on cockpit web console through https://<master>:9090, right?
Should be an issue with a newer kubernetes version (1.4 as reported upstream), but I'm not sure if this is testable right now. We included the patch because this Cockpit version might be used in newer environments later on. If the container logs are showing up well for current and maybe newer versions, then this should be good. If you can only test the current version, then it's at least not a regression.
Target release is set as 3.3.1 but the kubernetes version for 3.3.1 is still in version 1.3(kubernetes v1.3.0+52492b4) For OCP 3.4, kubernetes version is upgraded to 1.4(kubernetes v1.4.0+776c994), however in OCP 3.4,latest cockpit version is still 122-1 So I guess this could not be tested yet, right?
Please change to ON_QA when it could be tested
Troy keeps track of which images are rebuilt for testing, but you are correct: This can only be tested with a patched cockpit build (>= 122-3) and kubernetes 1.4.
Since this needs kubernetes 1.4, and a newer cockpit, I have dropped it from the errata it was on, and changed the release targets. I am also moving it back to assigned. When we have an updated cockpit for release Dominik can put it onto Modified and I'll get the image and errata setup.
Thanks Troy and Dominik
Once we have a 3.4 build with the latest Cockpit, this should be good to be tested.
I mis-read comment 12. I'll do a 3.4 build.
This should now be fixed in image openshift3/registry-console:3.4-1
Checked openshift3/registry-console:3.4-1, it has 122-3 cockpit build
Will you prepare new OCP 3.4 build with latest cockpit-kubernetes-122-3.el7.x86_64.rpm ?
The OCP 3.4 build is openshift3/registry-console:3.4-1
openshift v3.4.0.34+87d9d8d kubernetes v1.4.0+776c994 etcd 3.1.0-rc.0 openshift3/registry-console:3.4-1 cockpit-shell-122-3.el7.noarch cockpit-kubernetes-122-1.el7.x86_64 cockpit-bridge-122-3.el7.x86_64 cockpit-docker-122-3.el7.x86_64 cockpit-ws-122-3.el7.x86_64 Test on OCP 3.4 with above version, login on cockpit web console, check containers' logs, no matter proxy is used for browser or not, the containers' logs can be shown normally, attach screenshots for containers' logs. Didn't find issue during the check, so move the bug to Verified.
Created attachment 1230730 [details] log-1
Created attachment 1230731 [details] log-2
Created attachment 1230732 [details] log-3
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-2017:0067