Description of problem: registry-console is not needed in all cases (e.g. when embedding OpenShift), please consider making registry-console installation optional. Also, registry-console should have node selector (or at least it should go by default to the nodes where registries are being scheduled).
As a workaround, in a post-install task you can run oc patch as: oc patch dc registry-console -p '{"spec":{"template":{"spec":{"nodeSelector":{"role":"infra"}}}}}'
I'm reopening this for re-evaluation on the basis that a) we have most of the other components optional already, b) implementing this should be relatively straightforward, c) having this kind of one-off case is not consistent with the overall installer experience and might even create a bad precedence. However, if still deemed WONTFIX I won't insist keeping this open after that. Thanks.
Thanks Marko. I saw that Scott already reached to product management to confirm we can make it optional.
(In reply to Brenton Leanhardt from comment #11) > Thanks Marko. I saw that Scott already reached to product management to > confirm we can make it optional. Brenton, was there any conclusion about it that you could share with us? Thanks.
PM has said that we're fine to make it optional. It's a trivial change that we'd welcome a PR for. :-)
In 3.10 registry console will go to the master nodes in all scenarios. For making it optional, please submit a PR and I'll help review it and get it in.
*** Bug 1548984 has been marked as a duplicate of this bug. ***
Created https://github.com/openshift/openshift-ansible/pull/8766
registry-console deployment is on by default, but can be disabled by setting 'openshift_hosted_manage_registry_console=false'
Verified this bug with openshift-ansible-3.10.1-1.git.157.2bb6250.el7.noarch, and PASS. Set openshift_hosted_manage_registry_console=false, run installation, registry-console is not deployed now. PLAY [Create Hosted Resources - cockpit-ui] ************************************ TASK [Gathering Facts] ********************************************************* Tuesday 19 June 2018 04:28:17 -0400 (0:00:00.044) 0:15:03.655 ********** ok: [qe-jialiu310-master-etcd-1.0619-fv4.qe.rhcloud.com] TASK [cockpit-ui : fetch the docker-registry route] **************************** Tuesday 19 June 2018 04:28:19 -0400 (0:00:01.168) 0:15:04.824 ********** skipping: [qe-jialiu310-master-etcd-1.0619-fv4.qe.rhcloud.com] => {"changed": false, "skip_reason": "Conditional result was False", "skipped": true} TASK [cockpit-ui : Create passthrough route for registry-console] ************** Tuesday 19 June 2018 04:28:19 -0400 (0:00:00.031) 0:15:04.855 ********** skipping: [qe-jialiu310-master-etcd-1.0619-fv4.qe.rhcloud.com] => {"changed": false, "skip_reason": "Conditional result was False", "skipped": true} TASK [cockpit-ui : Create temp directory for kubeconfig] *********************** Tuesday 19 June 2018 04:28:19 -0400 (0:00:00.032) 0:15:04.888 ********** skipping: [qe-jialiu310-master-etcd-1.0619-fv4.qe.rhcloud.com] => {"changed": false, "skip_reason": "Conditional result was False", "skipped": true} TASK [cockpit-ui : set_fact] *************************************************** Tuesday 19 June 2018 04:28:19 -0400 (0:00:00.033) 0:15:04.921 ********** skipping: [qe-jialiu310-master-etcd-1.0619-fv4.qe.rhcloud.com] => {"changed": false, "skip_reason": "Conditional result was False", "skipped": true} TASK [cockpit-ui : Copy the admin client config(s)] **************************** Tuesday 19 June 2018 04:28:19 -0400 (0:00:00.033) 0:15:04.955 ********** skipping: [qe-jialiu310-master-etcd-1.0619-fv4.qe.rhcloud.com] => {"changed": false, "skip_reason": "Conditional result was False", "skipped": true} TASK [cockpit-ui : Deploy registry-console] ************************************ Tuesday 19 June 2018 04:28:19 -0400 (0:00:00.031) 0:15:04.987 ********** skipping: [qe-jialiu310-master-etcd-1.0619-fv4.qe.rhcloud.com] => {"changed": false, "skip_reason": "Conditional result was False", "skipped": true} TASK [cockpit-ui : Delete temp directory] ************************************** Tuesday 19 June 2018 04:28:19 -0400 (0:00:00.032) 0:15:05.019 ********** skipping: [qe-jialiu310-master-etcd-1.0619-fv4.qe.rhcloud.com] => {"changed": false, "skip_reason": "Conditional result was False", "skipped": true} [root@qe-jialiu310-master-etcd-1 ~]# oc get dc NAME REVISION DESIRED CURRENT TRIGGERED BY docker-registry 1 1 1 config router 1 1 1 config
(In reply to Clayton Coleman from comment #16) > In 3.10 registry console will go to the master nodes in all scenarios. For > making it optional, please submit a PR and I'll help review it and get it in. Why the regsity console must be on master nodes? Shouldn't it go by default to the nodes where registries are being scheduled?
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-2018:1816