Bug 1857211

Summary: install-config is not a reliable way to determine the cluster version, can cause operator failure
Product: Migration Toolkit for Containers Reporter: Erik Nelson <ernelson>
Component: GeneralAssignee: John Matthews <jmatthew>
Status: CLOSED ERRATA QA Contact: Xin jiang <xjiang>
Severity: medium Docs Contact: Avital Pinnick <apinnick>
Priority: unspecified    
Version: 1.3.0CC: alpatel, ernelson, jmatthew, jmontleo, mberube, pvauter, rjohnson, vpagar
Target Milestone: ---   
Target Release: 1.3.0   
Hardware: Unspecified   
OS: Unspecified   
Whiteboard:
Fixed In Version: Doc Type: If docs needed, set a value
Doc Text:
Story Points: ---
Clone Of:
: 1857212 (view as bug list) Environment:
Last Closed: 2020-09-30 18:42:32 UTC Type: Bug
Regression: --- Mount Type: ---
Documentation: --- CRM:
Verified Versions: Category: ---
oVirt Team: --- RHEL 7.3 requirements from Atomic Host:
Cloudforms Team: --- Target Upstream Version:
Embargoed:
Bug Depends On:    
Bug Blocks: 1857212    

Description Erik Nelson 2020-07-15 13:08:53 UTC
Description of problem:
We currently use the "install-config" section of the kube-system/cluster-config-v1 to determine the cluster version (https://github.com/konveyor/mig-operator/blob/release-1.2.3/roles/migrationcontroller/tasks/main.yml#L62).

This is not reliable, and there's no guarantee that the install-config key will exist. We have had reported operator failures in customer clusters with the following error:

u001b[0m\n\r\nTASK [migrationcontroller : set_fact] ******************************************\r\n\u001b[1;30mtask path: /opt/ansible/roles/migrationcontroller/tasks/main.yml:62\u001b[0m\n\u001b[0;31mfatal: [localhost]: FAILED! => {\"msg\": \"The task includes an option with an undefined variable. The error was: 'dict object' has no attribute 'metadata'\\n\\nThe error appears to be in '/opt/ansible/roles/migrationcontroller/tasks/main.yml': line 62, column 7, but may\\nbe elsewhere in the file depending on the exact syntax problem.\\n\\nThe offending line appears to be:\\n\\n\\n    - set_fact:\\n      ^ here\\n\"}\u001b[0m\n\r\nTASK [migrationcontroller : k8s_status]

Also see the failed lookup: https://github.com/konveyor/mig-operator/blob/release-1.2.3/roles/migrationcontroller/tasks/main.yml#L68

We have a more reliable source proposed here: https://github.com/konveyor/mig-operator/issues/380

Version-Release number of selected component (if applicable):
1.2.3

How reproducible:
It's undetermined what would cause the install-config to be missing, but if it is, the operator will fail on this task every time.

Comment 1 Jason Montleon 2020-07-15 13:14:30 UTC
We need to confirm the infrastructure resource is available for all 4.x releases and wasn't introduced at a later time. We can still use it for supported releases if only available later; we just need to be sure not to break older 4.x releases.

Comment 8 Jason Montleon 2020-08-27 21:28:28 UTC
Moving back to assigned. I had to revert the PR since it was failing. More details @ https://github.com/konveyor/mig-operator/issues/380

Comment 15 Xin jiang 2020-09-23 12:26:18 UTC
verified with MTC 1.3.0 on OCP 4.5

Comment 19 errata-xmlrpc 2020-09-30 18:42:32 UTC
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 (Migration Toolkit for Containers (MTC) Tool image release advisory 1.3.0), 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-2020:4148