Installation documentation broken at different points

Question

Installation documentation broken at different points

akram opened this issue 3 years ago · comments

While following the installation doc and after the introduction of new resources types, installation breaks at some points:

using 0.4.8-rc is suffering breaking changes
daemonset's hostpath container image requires quay.io/openshift/origin-csi-driver-shared-resource:4.10.0
my-csi-volume has a missing share attribute

Gabe Montero commented 3 years ago

/reopen

Akram Ben Aissi commented 3 years ago

/close

Gabe Montero · Answer 1 · Wed Oct 06 2021 00:05:50 GMT+0800 (China Standard Time)

While following the installation doc and after the introduction of new resources types, installation breaks at some points:
* using 0.4.8-rc is suffering breaking changes

* daemonset's `hostpath` container image requires `quay.io/openshift/origin-csi-driver-shared-resource:4.10.0`

* `my-csi-volume`  has a missing `share` attribute

share attribute is no longer valid @akram

it is sharedConfigMap and sharedSecret which map to the new CRDs

openshift-ci · Answer 2 · Wed Oct 06 2021 22:35:33 GMT+0800 (China Standard Time)

@gabemontero: Reopened this issue.

In response to this:

/reopen

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.

Gabe Montero · Answer 3 · Wed Oct 06 2021 22:39:40 GMT+0800 (China Standard Time)

@akram to summarize my posts in https://coreos.slack.com/archives/C02258G4S79/p1633497539444000

#59 was an incomplete fix ... I'll submit a new PR to address
possibly because we are not fully integrated in the payload yet, but something to ask about in #forum-testplaform, our image on quay.io is not getting updated whenever we merge a PR. Note the ages of the image:

quay.io/openshift/origin-csi-driver-shared-resource       4.10.0                              56ba55299526   11 days ago    383MB
quay.io/openshift/origin-csi-driver-shared-resource       latest                              56ba55299526   11 days ago    383MB

we have 6 commits in October alone, and today is Oct 6 ... much earlier than 11 days

so until we sort out the quay.io promotion frequency thing, you'll need to build master branch locally and use that image to run the correct version of code. Something like:

make build-image -e REGISTRY=docker.io -e REPOSITORY=gmontero
make deploy -e DRIVER_IMAGE=docker.io/gmontero/origin-csi-driver-shared-resource:latest

Gabe Montero · Answer 4 · Wed Oct 06 2021 22:44:51 GMT+0800 (China Standard Time)

@akram to summarize my posts in https://coreos.slack.com/archives/C02258G4S79/p1633497539444000
1. [Update daemonset registrar and hostpath images #59](https://github.com/openshift/csi-driver-shared-resource/pull/59) was an incomplete fix ... I'll submit a new PR to address

I was incorrect with ^^ ... simply misread my grep. We were in fact correct with #59 ... that said, what follows, with it not getting updated in 11 days, still explains the discrepancy @akram has seen in master branch

Gabe Montero · Answer 5 · Wed Oct 13 2021 20:22:14 GMT+0800 (China Standard Time)

the quay.io mirroring / promotion delay has been fixed for master/latest and whatever current release branch we are on.

There are some useful tips in https://issues.redhat.com/browse/DPTP-2567 for tracking the status of promotion and mirroring (sometimes CI does have short term outages). But definitely, the end to end flow that ultimately lands an image up at quay.io/openshift/origin-csi-driver-shared-resource is NOT instantaneous.

Of course, once we fully get in the payload, this concern is largely negated.

I'm leaving this issue open again to remind us to clean up some of the "dev install" documentation, removing some of the current "placeholders" we have, once the payload integration is complete.

Akram Ben Aissi · Answer 6 · Fri Oct 15 2021 15:59:05 GMT+0800 (China Standard Time)

@akram to summarize my posts in https://coreos.slack.com/archives/C02258G4S79/p1633497539444000
1. [Update daemonset registrar and hostpath images #59](https://github.com/openshift/csi-driver-shared-resource/pull/59) was an incomplete fix ... I'll submit a new PR to address
I was incorrect with ^^ ... simply misread my grep. We were in fact correct with #59 ... that said, what follows, with it not getting updated in 11 days, still explains the discrepancy @akram has seen in master branch

so until we sort out the quay.io promotion frequency thing, you'll need to build master branch locally and use that image to run the correct version of code. Something like:
make build-image -e REGISTRY=docker.io -e REPOSITORY=gmontero
make deploy -e DRIVER_IMAGE=docker.io/gmontero/origin-csi-driver-shared-resource:latest

Building the image locally needs also to have proper credentials for registry.ci.openshift.org ; we should document that somewhere. (Login to ci.apps aka openshift ci is needed).

oc login https://api.ci.l2s4.p1.openshiftapps.com:6443
oc registry login

Gabe Montero · Answer 7 · Wed Nov 10 2021 00:42:24 GMT+0800 (China Standard Time)

all that doc has been removed and replaced with in playload on tech preview clusters

/close

Gabe Montero · Answer 8 · Wed Nov 10 2021 00:47:49 GMT+0800 (China Standard Time)

weird I do not have authority to close this

@akram can you close this - it is no longer relevant

thanks

openshift-ci · Answer 9 · Wed Nov 10 2021 00:50:43 GMT+0800 (China Standard Time)

@gabemontero: You can't close an active issue/PR unless you authored it or you are a collaborator.

In response to this:

all that doc has been removed and replaced with in playload on tech preview clusters

/close

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.

openshift-ci · Answer 10 · Tue Nov 16 2021 17:16:19 GMT+0800 (China Standard Time)

@akram: Closing this issue.

In response to this:

/close

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.