Bug 1998571 - Catalog operator reports multiple heads for several channels
Summary: Catalog operator reports multiple heads for several channels
Keywords:
Status: CLOSED INSUFFICIENT_DATA
Alias: None
Product: OpenShift Container Platform
Classification: Red Hat
Component: OLM
Version: 4.8
Hardware: Unspecified
OS: Unspecified
medium
medium
Target Milestone: ---
: ---
Assignee: Ankita Thomas
QA Contact: Jian Zhang
URL:
Whiteboard:
Depends On:
Blocks:
TreeView+ depends on / blocked
 
Reported: 2021-08-27 15:47 UTC by Ben Luddy
Modified: 2024-12-20 20:51 UTC (History)
6 users (show)

Fixed In Version:
Doc Type: If docs needed, set a value
Doc Text:
Clone Of:
Environment:
Last Closed: 2022-01-17 15:57:45 UTC
Target Upstream Version:
Embargoed:


Attachments (Terms of Use)

Description Ben Luddy 2021-08-27 15:47:50 UTC
Description of problem:

Filed based on comment in https://bugzilla.redhat.com/show_bug.cgi?id=1982294#c14. Of the reported packages from the index "registry.redhat.io/redhat/redhat-operator-index:v4.8" today:

- codeready-workspaces
- kubevirt-hyperconverged
- openshift-gitops-operator
- openshift-pipelines-operator-rh
- web-terminal

Only openshift-pipelines-operator-rh and kubevirt-hyperconverged contain any deprecated entries, and in both cases the deprecated entries are channel tails, not inner channel entries. This indicates that there is a separate/distinct cause other than the one tracked in BZ 1982294.

Version-Release number of selected component (if applicable): 4.8.5

Comment 1 Ankita Thomas 2021-09-03 17:49:37 UTC
I tried to reproduce this with the 4.8 index, but I haven't been able to. Looking at the redhat-operators index and the community-operators index, it doesn't looks like they have any upgrade graph issues.

This could be a transient error with the index at the time, can you provide the subscriptions and catalogs or a must gather?

Comment 2 Robert Baumgartner 2021-09-04 11:26:45 UTC
must-gather uploaded.
https://drive.google.com/file/d/1vOLG4DpOryWigbhaAcvOjnJYYq3OO6K1/view?usp=sharing

ClusterID: d723aa91-ffd9-4578-8ef3-153e92eaee7b
ClusterVersion: Stable at "4.8.5"
ClusterOperators:
	clusteroperator/operator-lifecycle-manager is not upgradeable because ClusterServiceVersions blocking cluster upgrade: openshift-operators/openshift-gitops-operator.v1.2.0 is incompatible with OpenShift versions greater than 4.8.0

Comment 3 Simon Reber 2021-09-28 15:47:02 UTC
(In reply to Robert Baumgartner from comment #2)
> ClusterID: d723aa91-ffd9-4578-8ef3-153e92eaee7b
> ClusterVersion: Stable at "4.8.5"
> ClusterOperators:
> 	clusteroperator/operator-lifecycle-manager is not upgradeable because
> ClusterServiceVersions blocking cluster upgrade:
> openshift-operators/openshift-gitops-operator.v1.2.0 is incompatible with
> OpenShift versions greater than 4.8.0
This is expected as OpenShift GitOps 1.2 is not supported on OpenShift Container Platform 4.9. See https://access.redhat.com/solutions/6338121

Comment 6 Kedar Bidarkar 2021-10-19 08:11:41 UTC
This is consistently reproducible on a cluster  upgraded from 4.7.34 to 4.8.15

And then, when trying to upgrade CNV via CNV_SOURCE=osbs

From fresh cluster today, 

]$ oc logs -f catalog-operator-6b54d9586d-zz2pj -n openshift-operator-lifecycle-manager

E1019 05:39:51.700308       1 queueinformer_operator.go:290] sync "openshift-cnv" failed: found multiple channel heads: [kubevirt-hyperconverged-operator.v2.6.4 kubevirt-hyperconverged-operator.v2.6.0 kubevirt-hyperconverged-operator.v4.8.2], please check the `replaces`/`skipRange` fields of the operator bundles
I1019 05:39:51.700475       1 event.go:282] Event(v1.ObjectReference{Kind:"Namespace", Namespace:"", Name:"openshift-cnv", UID:"765c8d6a-db3b-4cb9-8953-88f9113c6913", APIVersion:"v1", ResourceVersion:"131585", FieldPath:""}): type: 'Warning' reason: 'ResolutionFailed' found multiple channel heads: [kubevirt-hyperconverged-operator.v2.6.4 kubevirt-hyperconverged-operator.v2.6.0 kubevirt-hyperconverged-operator.v4.8.2], please check the `replaces`/`skipRange` fields of the operator bundles
I1019 05:39:52.756466       1 event.go:282] Event(v1.ObjectReference{Kind:"Namespace", Namespace:"", Name:"openshift-cnv", UID:"765c8d6a-db3b-4cb9-8953-88f9113c6913", APIVersion:"v1", ResourceVersion:"131585", FieldPath:""}): type: 'Warning' reason: 'ResolutionFailed' found multiple channel heads: [kubevirt-hyperconverged-operator.v2.6.4 kubevirt-hyperconverged-operator.v2.6.0 kubevirt-hyperconverged-operator.v4.8.2], please check the `replaces`/`skipRange` fields of the operator bundles
I1019 05:39:53.069015       1 event.go:282] Event(v1.ObjectReference{Kind:"Namespace", Namespace:"", Name:"openshift-cnv", UID:"765c8d6a-db3b-4cb9-8953-88f9113c6913", APIVersion:"v1", ResourceVersion:"131585", FieldPath:""}): type: 'Warning' reason: 'ResolutionFailed' found multiple channel heads: [kubevirt-hyperconverged-operator.v2.6.4 kubevirt-hyperconverged-operator.v2.6.0 kubevirt-hyperconverged-operator.v4.8.2], please check the `replaces`/`skipRange` fields of the operator bundles


Was expecting a installplan for CNV v4.8.0, but no luck, due to the above issue,

[kbidarka@localhost manifests-iib-1634616320]$ oc get ip -A
NAMESPACE                 NAME            CSV                                         APPROVAL    APPROVED
openshift-cnv             install-sr8sh   kubevirt-hyperconverged-operator.v2.6.7     Manual      true
openshift-cnv             install-wrknk   kubevirt-hyperconverged-operator.v2.6.7     Manual      true
openshift-local-storage   install-jhg8d   local-storage-operator.4.7.0-202110051546   Automatic   true
openshift-local-storage   install-rc2sk   local-storage-operator.4.8.0-202110011559   Automatic   true
openshift-storage         install-9xdv4   ocs-operator.v4.7.4                         Automatic   true
openshift-storage         install-jmhs4   ocs-operator.v4.7.5                         Automatic   true
openshift-storage         install-k6ntn   ocs-operator.v4.8.3                         Automatic   true

Comment 7 Kedar Bidarkar 2021-10-19 08:21:29 UTC
I had to reinstall the setup, as I need to verify a CNV blocker bug.

I can reinstall and share the setup, on which this was seen/reproducible.
As this is seen with upgrades and upgrades being very time consuming, feel it would be great to have this issue resolved or
a at-least we have a workaround.

Comment 9 Kedar Bidarkar 2021-11-13 12:18:52 UTC
I have hit this issue again today, 

The easiest way to reproduce this issue is to install OCP4.7 and also install OCP Virt 2.6.7 from our Jenkins job (which is fully automated. )
I have hit this more consistently on a Bare-Metal setup compared to a PSI Setup, IIRC.

E1113 12:05:17.405753       1 queueinformer_operator.go:290] sync "openshift-cnv" failed: found multiple channel heads: [kubevirt-hyperconverged-operator.v2.6.0 kubevirt-hyperconverged-operator.v2.6.8 kubevirt-hyperconverged-operator.v2.6.4], please check the `replaces`/`skipRange` fields of the operator bundles
I1113 12:05:17.405901       1 event.go:282] Event(v1.ObjectReference{Kind:"Namespace", Namespace:"", Name:"openshift-cnv", UID:"49e63ff0-6e94-45f0-a2ec-af26d71940d7", APIVersion:"v1", ResourceVersion:"47714", FieldPath:""}): type: 'Warning' reason: 'ResolutionFailed' found multiple channel heads: [kubevirt-hyperconverged-operator.v2.6.0 kubevirt-hyperconverged-operator.v2.6.8 kubevirt-hyperconverged-operator.v2.6.4], please check the `replaces`/`skipRange` fields of the operator bundles
E1113 12:05:17.905954       1 queueinformer_operator.go:290] sync "openshift-cnv" failed: found multiple channel heads: [kubevirt-hyperconverged-operator.v2.6.0 kubevirt-hyperconverged-operator.v2.6.8 kubevirt-hyperconverged-operator.v2.6.4], please check the `replaces`/`skipRange` fields of the operator bundles
I1113 12:05:17.906128       1 event.go:282] Event(v1.ObjectReference{Kind:"Namespace", Namespace:"", Name:"openshift-cnv", UID:"49e63ff0-6e94-45f0-a2ec-af26d71940d7", APIVersion:"v1", ResourceVersion:"47714", FieldPath:""}): type: 'Warning' reason: 'ResolutionFailed' found multiple channel heads: [kubevirt-hyperconverged-operator.v2.6.0 kubevirt-hyperconverged-operator.v2.6.8 kubevirt-hyperconverged-operator.v2.6.4], please check the `replaces`/`skipRange` fields of the operator bundles

Comment 11 Ankita Thomas 2021-12-06 18:13:53 UTC
There was an issue with identifying skipped channel head candidates that was fixed in 4.9, backporting that fix to 4.8, 4.7

Comment 16 Red Hat Bugzilla 2023-09-15 01:14:20 UTC
The needinfo request[s] on this closed bug have been removed as they have been unresolved for 500 days


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