-
Notifications
You must be signed in to change notification settings - Fork 32
*: Pivot from v1beta1 Cincinnati to v1 UpdateService #78
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Conversation
Following [1]: $ git checkout --orphan 0.19-scaffolding $ git rm -rf . $ operator-sdk-v0.19.3 init --domain openshift.io --repo github.com/openshift/cincinnati-operator $ operator-sdk-v0.19.3 create api --group cincinnati --version v1beta1 --kind Cincinnati --resource --controller $ git add -A .gitignore * The following paths are ignored by one of your .gitignore files: bin hint: Use -f if you really want to add them. hint: Turn this message off by running hint: "git config advice.addIgnoredFile false" We don't want to commit bin, so the warning is fine.
|
[APPROVALNOTIFIER] This PR is APPROVED This pull-request has been approved by: wking The full list of commands accepted by this bot can be found here. The pull request process is described here DetailsNeeds approval from an approver in each of these files:
Approvers can indicate their approval by writing |
7eb83af to
09d48de
Compare
Generated with: $ git rm -rf . $ operator-sdk-v0.19.3 init --domain openshift.io --repo github.com/openshift/cincinnati-operator $ operator-sdk-v0.19.3 create api --group updateservice --version v1 --kind UpdateService --resource --controller $ sed -i 's/cincinnati-operator/update-service-operator/' config/default/kustomization.yaml $ git add -A .gitignore * The following paths are ignored by one of your .gitignore files: bin cover.out hint: Use -f if you really want to add them. hint: Turn this message off by running hint: "git config advice.addIgnoredFile false" We don't want to commit bin or cover.out, so the warning is fine. This regenerates the file with the scaffolding for the new API name, and we can merge it into our main branch to update the type there. From [1], regenerating a new API is the prefered migration approach, and running the '... create api ...' call directly in the main branch gave: ... vet: functests/utils.go:79:53: SchemeGroupVersion not declared by package v1beta1 make: *** [vet] Error 1 Error: failed to create API with version "3-alpha": exit status 2 ... FATA[0006] failed to create API with version "3-alpha": exit status 2 [1]: operator-framework/operator-sdk#560 (comment)
Generated with: $ git checkout -b sdk-0.19 origin/master $ git merge --allow-unrelated-histories 0.19-scaffolding $ git checkout --theirs .gitignore # drop lots of user-editor cruft, see "Which file to place a pattern in..." in https://git-scm.com/docs/gitignore#_description $ emacs api/v1beta1/cincinnati_types.go # keep some of the outgoing pkg/apis/cincinnati/v1beta1/cincinnati_types.go content $ emacs controllers # keep some of the outgoing pkg/controller content $ emacs main.go # incorperate some of the outgoing cmd/manager/main.go content $ emacs Dockerfile # keep most of the pre-0.19 content. We will still vendor deps, and don't want to have GOARCH, etc., opinions $ emacs Makefile # keep most of the 0.19 content, with a few pre-0.19 rules for openshift/release compatability $ emacs README.md # bump SDK references to 0.19 $ make manifests $ git add .gitignore config Dockerfile main.go Makefile README.md $ git rm -rf cmd/manager pkg/apis pkg/controller $ git checkout HEAD -- go.sum # we'll auto-generate this in a future commit $ emacs go.mod # basically a union of the two sets. Drop the hopefully-obsolete 'replace' block. $ make bundle $ emacs config/manifests/bases/cincinnati-operator.clusterserviceversion.yaml # keep some of the outgoing deploy/olm-catalog content $ make bundle $ go mod tidy $ go mod vendor $ git add -A go.* vendor using: $ go version go version go1.15.2 linux/amd64 $ kustomize version {Version:kustomize/v3.8.6 GitCommit:c1747439cd8bc956028ad483cdb30d9273c18b24 BuildDate:2020-10-29T23:07:50Z GoOs:linux GoArch:amd64} $ controller-gen --version Version: v0.3.0 'oc' has a kustomize subcommand, but even in the master oc branch, that is currently v2 [1], while 'kustomize edit ...' needs v3 to avoid [2]: cd config/manager && /cli/oc kustomize edit set image controller=controller:latest Error: specify one path to a kustomization directory The 'docker build . -t ${IMG}' -> 'docker build -t ${IMG} .' change is for Podman compatibility [3]. Commenting out the 'test' prerequisite allows us to run 'make docker-build' without having functests try to run tests in whatever cluster you may be pointed at. The increased memory limits are based on peak measurements of ~11m CPU and ~94MiB in a 4.6.4 amd64 cluster with a single Cincinnati object. [1]: https://github.com/openshift/oc/blob/1bd1f5ff4aa98d6e4184c2948928b6111fa99191/vendor/modules.txt#L1293 [2]: https://prow.ci.openshift.org/view/gs/origin-ci-test/pr-logs/pull/openshift_cincinnati-operator/76/pull-ci-openshift-cincinnati-operator-master-operator-e2e/1326532618765733888 [3]: operator-framework/operator-sdk#4226
Generated with: $ git merge 0.19-scaffolding-v1 $ emacs main.go api/v1/updateservice_types.go controllers/updateservice_controller.go # resolve conflicts $ emacs README.md Dockerfile controllers/*.go functests/*.go config/manifests/bases/cincinnati-operator.clusterserviceversion.yaml # remaining api/v1beta1 -> api/v1 migration and renames $ emacs config/samples/updateservice_v1_updateservice.yaml # copy some content from the outgoing Cincinnati sample $ git rm -f config/samples/cincinnati_v1beta1_cincinnati.yaml controllers/suite_test.go $ rm -rf bundle $ make generate manifests bundle $ git add -A README.md main.go api bundle config controllers functests
|
@wking: The following test failed, say
Full PR test history. Your PR dashboard. DetailsInstructions for interacting with me using PR comments are available here. If you have questions or suggestions related to my behavior, please file an issue against the kubernetes/test-infra repository. I understand the commands that are listed here. |
|
#65 landed the bulk of this. I'll file separate PRs for anything unique to this PR that I want to keep. /close |
|
@wking: Closed this PR. DetailsIn response to this:
Instructions for interacting with me using PR comments are available here. If you have questions or suggestions related to my behavior, please file an issue against the kubernetes/test-infra repository. |
Baked in edges:
$ oc adm release info quay.io/openshift-release-dev/ocp-release:4.3.0-rc.0-x86_64 | grep Upgrades
Upgrades: 4.2.13
$ oc adm release info quay.io/openshift-release-dev/ocp-release:4.3.0-rc.3-x86_64 | grep Upgrades
Upgrades: 4.2.16, 4.3.0-rc.0, 4.3.0-rc.1, 4.3.0-rc.2
The wide 'from' regexp was appropriate for 4.3.0-rc.0, which had no
4.3 update sources. But rc.3 does have update sources, and we want to
allow 4.3.0-rc.0 -> 4.3.0-rc.3, because it is not impacted by the
4.2->4.3 GCP update bug. The overly-strict regexp was from 6d3db09
(Blocking edges to candidate 4.3.0-rc.3, 2020-01-23, openshift#34).
Also expand the referenced bugs to for the blocked 4.2 -> 4.3 edges:
* Update hangs with [1]:
Working towards 4.3.0...: 13% complete
and machine-config going Degraded=True with RequiredPoolsFailed:
Unable to apply 4.3.0-...: timed out waiting for the condition
during syncRequiredMachineConfigPools: pool master has not
progressed to latest configuration: controller version mismatch
for rendered-master-6c22... expected 23a6... has d780... retrying
Fixed in 4.2 with MCO 31fed93 [2] and in 4.2 with MCO 25bb6ae [3].
$ oc adm release info --commits registry.svc.ci.openshift.org/ocp/release:4.2.14 | grep machine-config
machine-config-operator https://github.com/openshift/machine-config-operator d780d197a9c5848ba786982c0c4aaa7487297046
$ oc adm release info --commits registry.svc.ci.openshift.org/ocp/release:4.2.16 | grep machine-config
machine-config-operator https://github.com/openshift/machine-config-operator 31fed93186c9f84708f5cdfd0227ffe4f79b31cd
So the 4.2 fix was in 4.2.16.
$ oc adm release info --commits registry.svc.ci.openshift.org/ocp/release:4.3.0-rc.0 | grep machine-config
machine-config-operator https://github.com/openshift/machine-config-operator 23a6e6fb37e73501bc3216183ef5e6ebb15efc7a
$ oc adm release info --commits registry.svc.ci.openshift.org/ocp/release:4.3.0-rc.3 | grep machine-config
machine-config-operator https://github.com/openshift/machine-config-operator 25bb6aeb58135c38a667e849edf5244871be4992
So the 4.3 fix was new in rc.3.
* Updates hang with FailedCreatePodSandBox events in the
openshift-ingress namespace like [4]:
pod/router-default-...: Failed create pod sandbox: rpc error: code
= Unknown desc = failed to create pod network sandbox
k8s_router-default-..._openshift-ingress_...(...): Multus: error
adding pod to network "openshift-sdn": delegateAdd: error invoking
DelegateAdd - "openshift-sdn": error in getting result from
AddNetwork: CNI request failed with status 400: 'failed to run
IPAM for ...: failed to run CNI IPAM ADD: failed to allocate for
range 0: no IP addresses available in range set: <ip1>-<ip2>
Fixed in 4.2 with MCO 9366460 [5] and in 4.3 with MCO 311a01e [6].
$ git --no-pager log --first-parent --oneline -4 origin/release-4.2
6e0df82c (origin/release-4.2) Merge pull request #1347 from openshift-cherrypick-robot/cherry-pick-1285-to-release-4.2
93664600 Merge pull request #1362 from rphillips/fixes/1787581_4.2
bd358bb7 Merge pull request #1323 from openshift-cherrypick-robot/cherry-pick-1320-to-release-4.2
31fed931 Merge pull request #1358 from runcom/osimageurl-race-42
so the 4.2 fix was after 4.2.16's 31fed93186.
$ git --no-pager log --first-parent --oneline -8 origin/release-4.3
3ad3a836 (origin/release-4.3) Merge pull request #1399 from celebdor/haproxy-v4v6
25503eee Merge pull request #1353 from russellb/1211-4.3-backport
67ab306b Merge pull request #1426 from mandre/ssc43
d74f56fe Merge pull request #1410 from retroflexer/manual-cherry-pick-from-master
207cc171 Merge pull request #1406 from openshift-cherrypick-robot/cherry-pick-1396-to-release-4.3
25bb6aeb Merge pull request #1359 from runcom/osimageurl-race-43
311a01e8 Merge pull request #1361 from rphillips/fixes/1787581_4.3
23a6e6fb Merge pull request #1348 from openshift-cherrypick-robot/cherry-pick-1285-to-release-4.3
So the 4.3 fix was between rc.0's 23a6e6fb37 and rc.3's 25bb6aeb58
(see 'release info' calls in the previous list entry for those
commit hashes).
* Update CI fails with [7,8]:
Could not reach HTTP service through <ip>:80 after 2m0s
and authentication going Degraded=True with RouteHealthDegradedFailedGet:
RouteHealthDegraded: failed to GET route: dial tcp <ip>:443:
connect: connection refused
Fixed in 4.2 with SDN 677b3a8 [9] and in 4.3 with SDN 74a8aee [10].
$ oc adm release info --commits registry.svc.ci.openshift.org/ocp/release:4.2.16 | grep ' node '
node https://github.com/openshift/sdn 770cb7bf922a721bc6c62af5490439d6174036fe
$ oc adm release info --commits registry.svc.ci.openshift.org/ocp/release:4.2.14 | grep ' node '
node https://github.com/openshift/sdn 770cb7bf922a721bc6c62af5490439d6174036fe
$ git --no-pager log --first-parent --oneline -4 origin/release-4.2
098a6410 (origin/release-4.2) Merge pull request openshift#95 from danwinship/fork-k8s-client-go-4.2
9955a65b Merge pull request openshift#72 from juanluisvaladas/too_many_dns_queries_42
677b3a80 Merge pull request openshift#90 from openshift-cherrypick-robot/cherry-pick-81-to-release-4.2
770cb7bf Merge pull request openshift#73 from danwinship/egressip-cleanup-4.2
So the fix landed after 4.2.16's 770cb7bf.
$ oc adm release info --commits registry.svc.ci.openshift.org/ocp/release:4.3.0-rc.0 | grep ' sdn '
sdn https://github.com/openshift/sdn d4e36d5019ef0e130e0d246581508821a7322753
$ git --no-pager log --first-parent --oneline -5 origin/release-4.3
490a574e (origin/release-4.3) Merge pull request openshift#98 from openshift-cherrypick-robot/cherry-pick-96-to-release-4.3
85ab1033 Merge pull request openshift#78 from openshift-cherrypick-robot/cherry-pick-57-to-release-4.3
d4e36d50 Merge pull request openshift#85 from openshift-cherrypick-robot/cherry-pick-84-to-release-4.3
dabc4ef5 Merge pull request openshift#83 from dougbtv/backport-build-use-host-local
74a8aee3 Merge pull request openshift#81 from openshift-cherrypick-robot/cherry-pick-79-to-release-4.3
So the fix landed before rc.0's d4e36d50.
* GCP update CI fails with [11]:
Could not reach HTTP service through <ip>:80 after 2m0s
in 4.2.16 -> 4.3.0-rc.0 [12], 4.2.16 -> 4.3.0-rc.3 [13,14,15], and
4.2.18 -> 4.3.1 [16]. This doesn't happen every time though; at
least one 4.2.16 -> 4.3.0-rc.3 has passed on GCP [17]. We don't
have a root-cause yet, but the final failure matches [8] discussed
above.
[1]: https://bugzilla.redhat.com/show_bug.cgi?id=1786993
[2]: openshift/machine-config-operator#1358 (comment)
[3]: openshift/machine-config-operator#1359 (comment)
[4]: https://bugzilla.redhat.com/show_bug.cgi?id=1787635
[5]: openshift/machine-config-operator#1362 (comment)
[6]: openshift/machine-config-operator#1361 (comment)
[7]: https://prow.svc.ci.openshift.org/view/gcs/origin-ci-test/logs/release-openshift-origin-installer-e2e-gcp-upgrade/214#1:build-log.txt%3A414
[8]: https://bugzilla.redhat.com/show_bug.cgi?id=1781763
[9]: openshift/sdn#90 (comment)
[10]: openshift/sdn#81 (comment)
[11]: https://bugzilla.redhat.com/show_bug.cgi?id=1785457
[12]: https://prow.svc.ci.openshift.org/view/gcs/origin-ci-test/logs/release-openshift-origin-installer-e2e-gcp-upgrade/216
[13]: https://prow.svc.ci.openshift.org/view/gcs/origin-ci-test/logs/release-openshift-origin-installer-e2e-gcp-upgrade/232
[14]: https://prow.svc.ci.openshift.org/view/gcs/origin-ci-test/logs/release-openshift-origin-installer-e2e-gcp-upgrade/233
[15]: https://prow.svc.ci.openshift.org/view/gcs/origin-ci-test/logs/release-openshift-origin-installer-e2e-gcp-upgrade/234
[16]: https://prow.svc.ci.openshift.org/view/gcs/origin-ci-test/logs/release-openshift-origin-installer-e2e-gcp-upgrade/286
[17]: https://prow.svc.ci.openshift.org/view/gcs/origin-ci-test/logs/release-openshift-origin-installer-e2e-gcp-upgrade/230
Builds on #76.
Similar to #65's d31969e, but via re-generating with the SDK tooling to distinguish those changes from the ones we manually apply afterwards.