Bug 2078567 - Adding option that is not understood by legacy cloud provider to user-managed cloud.conf causes KCM to crashloop
Summary: Adding option that is not understood by legacy cloud provider to user-managed...
Keywords:
Status: CLOSED DUPLICATE of bug 2077933
Alias: None
Product: OpenShift Container Platform
Classification: Red Hat
Component: Cloud Compute
Version: 4.11
Hardware: Unspecified
OS: Unspecified
high
medium
Target Milestone: ---
: 4.11.0
Assignee: ShiftStack Bugwatcher
QA Contact: Jon Uriarte
URL:
Whiteboard:
Depends On:
Blocks:
TreeView+ depends on / blocked
 
Reported: 2022-04-25 16:06 UTC by Martin André
Modified: 2022-04-26 06:49 UTC (History)
4 users (show)

Fixed In Version:
Doc Type: If docs needed, set a value
Doc Text:
Clone Of:
Environment:
Last Closed: 2022-04-26 06:49:13 UTC
Target Upstream Version:
Embargoed:


Attachments (Terms of Use)

Description Martin André 2022-04-25 16:06:03 UTC
As the openstack platform is migrating to the external cloud provider, the user-managed cloud.conf in `openshift-config/cloud-provider-config` is now the source for both the in-tree cloud.conf (in `openshift-config-managed/kube-cloud-config`) and the CCM cloud.conf (in `openshift-cloud-controller-manager/cloud-conf`).

However, even after switching to CCM we need to keep initializing the in-tree cloud provider in order to be able to detach volumes created with legacy cloud provider.

Moreover, the legacy cloud provider errors out when it finds an option it does not understand, causing KCM to enters a crashloopback.

We can't legitimately ask customer to limit themselves to the intersection of in-tree and external cloud provider options. We should make the legacy cloud provider option parsing more tolerant and ignore unknown options.

Comment 1 Martin André 2022-04-26 06:49:13 UTC

*** This bug has been marked as a duplicate of bug 2077933 ***


Note You need to log in before you can comment on or make changes to this bug.