+++ This bug was initially created as a clone of Bug #1845165 +++ Description of problem: ocs-must-gather image is missing in relatedImages section of OCS operator CSV Version of all relevant components (if applicable): ocs-olm-operator:4.4.0-437.ci Does this issue impact your ability to continue to work with the product? This is necessary for working in restricted network environments. Is there any workaround available to the best of your knowledge? It is possible to mirror the required image manually. Rate from 1 - 5 the complexity of the scenario you performed that caused this bug (1 - very simple, 5 - very complex)? 1 Can this issue reproducible? Yes Can this issue reproduce from the UI? No If this is a regression, please provide more details to justify this: No Steps to Reproduce: 1. Get ocs-operator.v4.4.0-*.ci.clusterserviceversion.yaml file: $ podman run --rm --entrypoint /usr/bin/cat \ quay.io/rhceph-dev/ocs-olm-operator:4.4.0-437.ci \ /manifests/4.4.0-437.ci/ocs-operator.v4.4.0-437.ci.clusterserviceversion.yaml \ > ocs-operator.v4.4.0-437.ci.clusterserviceversion.yaml 2. Look for ocs-must-gather image in the relatedImages section. $ grep ocs-must-gather ocs-operator.v4.4.0-437.ci.clusterserviceversion.yaml $ less ocs-operator.v4.4.0-437.ci.clusterserviceversion.yaml Actual results: ocs-must-gather is not listed in the relatedImages section of OCS operator CSV Expected results: ocs-must-gather is listed in the relatedImages section of OCS operator CSV Additional info: See also bug 1840708 --- Additional comment from Jose A. Rivera on 2020-06-17 08:38:08 CDT --- This is already fixed in release-4.5. I believe we dropped disconnected install from OCS 4.4. The presence of lib-bucket-provisioner doesn't make it impossible to install in a disconnected environment, Annette has found out how to work around it just, but it is a manual workaround so we still don't get one-button installation for OCS in a disconnected environment. Still, backporting this would make it even easier to do as a support exception or something, so I'm fine with it. I'm giving devel_ack+. --- Additional comment from Jose A. Rivera on 2020-06-17 08:39:13 CDT --- I was wrong! This is not already fixed for release-4.5. So we'll need a separate backport and BZ for that one.
must-gather is not one of the operands of ocs-operator. It doesn't belong in related-images. OCS Operator doesn't pull must-gather for any purpose. It's used separately and should be fetched separately. maybe this bug needs an alternative approach.
I talked to Trevor King from the OCP side and basically we agreed that while this is "technically" not part of the operator(s), there is no other mechanism for copying images to a disconnected environment that aren't shipped with the release payload (e.g. like the ocp must-gather is). I think this is something we can live with, as such I'm going to say it's valid to just ship it as part of relatedImages.
The code is now in downstream release-4.5 There is no PR as the patches were accidentally pushed directly to the branch instead of via a pull request.
Why isn't the bug ON_QA then?
(In reply to Yaniv Kaul from comment #9) > Why isn't the bug ON_QA then? I don't know, maybe because (due to the accident), there is no backport PR. I intentionally did not move it manually to see what errata automation does, but it has done nothing... Moving manually now.
First build containing this fix and passed CI is: https://ceph-jenkins.rhev-ci-vms.eng.rdu2.redhat.com/job/OCS%20Build%20Pipeline%204.5/66/
I've tried to check the ocs-must-gather image in the relatedImages section using the steps from description and I don't see it there: ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ # export VERSION=4.5.0-487.ci # podman run --rm --entrypoint /usr/bin/cat \ quay.io/rhceph-dev/ocs-olm-operator:${VERSION} \ /manifests/${VERSION}/ocs-operator.v${VERSION}.clusterserviceversion.yaml \ > ocs-operator.v${VERSION}.clusterserviceversion.yaml # grep ocs-must-gather ocs-operator.v${VERSION}.clusterserviceversion.yaml # echo $? 1 # cat ocs-operator.v${VERSION}.clusterserviceversion.yaml << truncated >> relatedImages: - image: quay.io/rhceph-dev/rook-ceph@sha256:d111b35caa3611b84cb164b8a79eb9a90451794afb2fc828ee2db26276a06fc1 name: rook-container - image: quay.io/rhceph-dev/cephcsi@sha256:b4e7caf299762bd78f40f174a166d0d8399eef00593e6afcb9696b241cd3ceb0 name: rook-csi - image: registry.redhat.io/openshift4/ose-csi-driver-registrar@sha256:5ed44407414538adcba19f3780be37193f1b5a8789fbf3a4e7ed65c7693f1b66 name: rook-csi-registrar - image: registry.redhat.io/openshift4/ose-csi-external-resizer-rhel7@sha256:bc1b07743216f68f0d292e7ed7607b2ce04c7f81d453a09cecc4cc3571eb19e5 name: rook-csi-resizer - image: registry.redhat.io/openshift4/ose-csi-external-provisioner-rhel7@sha256:f04a827f92c355f3f486957518bcf282fd7b9546ecdc275d04e8389f6290e712 name: rook-csi-provisioner - image: registry.redhat.io/openshift4/ose-csi-external-attacher@sha256:fc97bc7790703299ee19d9c4353699e049773babbfe74d0b7691ad8536cf250a name: rook-csi-attacher - image: quay.io/rhceph-dev/rhceph@sha256:45e6f28bb67c81b826acb64fad5c0da1cac3dffb41a88992fe4ca2be79575fa6 name: ceph-container - image: quay.io/rhceph-dev/mcg-operator@sha256:60b796c712be8dd01ab57216d92b2c2b1f3b98875e7cb6016bcbe989856a4435 name: noobaa-operator - image: quay.io/rhceph-dev/mcg-core@sha256:049ba73bee755ebb4503c4beab834f0b95f67df3376257e875a24584866eca0e name: noobaa-core - image: registry.redhat.io/rhscl/mongodb-36-rhel7@sha256:3292de73cb0cd935cb20118bd86e9ddbe8ff8c0a8d171cf712b46a0dbd54e169 name: noobaa-db version: 4.5.0-487.ci ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ Am I missing something?
Fix is definitely there in the used build, Pulkit needs to check why the image is not appearing in related image section.
Hi Pulkit, I was thinking about something like that. The questions what I have (and what I don't fully understand) are: * when the csv-merger is executed or who can decide/affect if the image will be listed in the relatedImages section or not? * where is the correct place to check that the must-gather image is in the relatedImages section? Is it ok, to check it in the /manifests/${VERSION}/ocs-operator.v${VERSION}.clusterserviceversion.yaml file in the ocs-olm-operator image? Thanks
(accidentally removed second needinfo)
Tested and verified on version 4.5.0-515.ci, ocs-must-gather image is listed in the relatedImages section of the ocs-olm-operator CSV. ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ # export VERSION=4.5.0-515.ci # podman run --rm --entrypoint /usr/bin/cat \ quay.io/rhceph-dev/ocs-olm-operator:${VERSION} \ /manifests/${VERSION}/ocs-operator.v${VERSION}.clusterserviceversion.yaml \ > ocs-operator.v${VERSION}.clusterserviceversion.yaml # cat ocs-operator.v${VERSION}.clusterserviceversion.yaml --- apiVersion: operators.coreos.com/v1alpha1 kind: ClusterServiceVersion metadata: annotations: <<truncated>> provider: name: Red Hat, Inc relatedImages: - image: quay.io/rhceph-dev/rook-ceph@sha256:6aaf689232cb7fcb44e37dc1c34b17c7cc81d5fe244cfb4277fafdb5a3865ee4 name: rook-container - image: quay.io/rhceph-dev/cephcsi@sha256:244099ffc77fe965cd258e105aeff127de08673830a679ecb2525d9220e161fb name: rook-csi - image: registry.redhat.io/openshift4/ose-csi-driver-registrar@sha256:5ed44407414538adcba19f3780be37193f1b5a8789fbf3a4e7ed65c7693f1b66 name: rook-csi-registrar - image: registry.redhat.io/openshift4/ose-csi-external-resizer-rhel7@sha256:bc1b07743216f68f0d292e7ed7607b2ce04c7f81d453a09cecc4cc3571eb19e5 name: rook-csi-resizer - image: registry.redhat.io/openshift4/ose-csi-external-provisioner-rhel7@sha256:f04a827f92c355f3f486957518bcf282fd7b9546ecdc275d04e8389f6290e712 name: rook-csi-provisioner - image: registry.redhat.io/openshift4/ose-csi-external-attacher@sha256:fc97bc7790703299ee19d9c4353699e049773babbfe74d0b7691ad8536cf250a name: rook-csi-attacher - image: quay.io/rhceph-dev/rhceph@sha256:df67e134c9037707118a8533670b2f77647355a4af2fdc1247d79679cb6bb676 name: ceph-container - image: quay.io/rhceph-dev/mcg-operator@sha256:c0f05d916cd630c318a47958d11aa4249e849192fe5c905fdbba077486e06548 name: noobaa-operator - image: quay.io/rhceph-dev/mcg-core@sha256:241d9693c1166273ae386e51c317035785eaa24c31d0341e2a72a55d089d2558 name: noobaa-core - image: registry.redhat.io/rhscl/mongodb-36-rhel7@sha256:ba74027bb4b244df0b0823ee29aa927d729da33edaa20ebdf51a2430cc6b4e95 name: noobaa-db - image: quay.io/rhceph-dev/ocs-must-gather@sha256:2768a5eb6162307681d8dc35fc431c5a59a71cac1b7de98e7bdef851431a9ba3 name: ocs-must-gather version: 4.5.0-515.ci ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ >> 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 (Red Hat OpenShift Container Storage 4.5.0 bug fix and enhancement update), 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-2020:3754