Description of problem: https://prow.ci.openshift.org/job-history/origin-ci-test/pr-logs/directory/pull-ci-openshift-console-master-e2e-gcp-console is consistently broken. Latests build shows different errors: displays subscription creation form for selected Operator displays Operator as subscribed in OperatorHub displays subscription creation form for selected Operator
Sam Padgett wrote: "There’s a confluence of things. The olm custom catalog isn’t loading before the tests run. We updated the olm tests around the same time we added cypress. Then I suspect an auth change might have affected us yesterday [8/12]" irt. the Cypress test failures, PR (https://github.com/openshift/console/pull/6319) was put in on 8/12 which seemed to fix the cypress flakes There does seem to be olm flake when the catalog does load: "[SEVERE] https://console-openshift-console.apps.ci-op-gnx2w8cv-75d12.origin-ci-int-gce.dev.openshift.com/static/main-chunk-737049ed273daaf189aa.min.js 0:1219704 "Uncaught error" "Uncaught TypeError: Cannot read property 'style' of null" Investigating....
The full olm error I saw was: 2) Interacting with an `AllNamespaces` install mode Operator (Jaeger) : displays subscription creation form for selected Operator Error: Failed: omg js error: Uncaught TypeError: Cannot read property 'style' of null Just ran it twice locally and works, so seems to be a flake caused by catalog not loading: 2) Interacting with an `AllNamespaces` install mode Operator (Jaeger) ✔ displays subscription creation form for selected Operator ✔ selects all namespaces for Operator subscription ... So does seem to be a true flake, most likely caused by olm custom catalog not initially loading -not an error in the test.
https://github.com/openshift/console/pull/6327 updates the junit report filename so it gets indexed by ci-search, which will help us understand how often each flake is occurring.
PR was merged a couple of days ago [8/9] which was hoped to improve OLM scenario reliability. It looks like the olm catalog is loading but the tests aren't waiting long enough: https://github.com/openshift/console/pull/6253
I believe the linked PR will fix the 3 errors mentioned in the description. If we see other problems, we should track under separate Bugzilla bugs.
*** Bug 1865998 has been marked as a duplicate of this bug. ***
Checked the recent logs, there is no such flake as describe or anything related to olm operators/subscriptions which broke the integration testing. This could be Verified.
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 (OpenShift Container Platform 4.6 GA Images), 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-2020:4196