Rebase onto Kubernetes 6241a21 by smarterclayton · Pull Request #1136 · openshift/origin

smarterclayton · 2015-02-25T06:53:46Z

Includes remote command execution and Docker caching

smarterclayton · 2015-02-25T06:53:58Z

[test]

smarterclayton · 2015-02-25T06:54:17Z

@pweil- @deads2k review please.

smarterclayton · 2015-02-25T06:55:13Z

Needs #1135 fixed upstream (although patched here).

smarterclayton · 2015-02-25T06:55:33Z

Might fix #1115

openshift-bot · 2015-02-25T06:57:01Z

continuous-integration/openshift-jenkins/test SUCCESS (https://ci.openshift.redhat.com/jenkins/job/test_pull_requests_openshift3/1219/)

smarterclayton · 2015-02-25T07:02:37Z

I cherry-picked Vish's change in as well.

pweil- · 2015-02-25T14:16:26Z

Godeps/_workspace/src/github.com/GoogleCloudPlatform/kubernetes/pkg/api/rest/create_test.go

Is it correct to be using the services strategy when validating pod name generation? In controllers/types.go they're technically the same strategy, just struck me as strange when reading it.

Disregard, the next file removes the Pods strategy....

I dunno, ask upstream.

On Feb 25, 2015, at 6:16 AM, Paul notifications@github.com wrote:

In Godeps/_workspace/src/github.com/GoogleCloudPlatform/kubernetes/pkg/api/rest/create_test.go:

@@ -25,17 +25,17 @@ import (

func TestCheckGeneratedNameError(t *testing.T) {
expect := errors.NewNotFound("foo", "bar")

if err := CheckGeneratedNameError(Pods, expect, &api.Pod{}); err != expect {

if err := CheckGeneratedNameError(Services, expect, &api.Pod{}); err != expect {
Is it correct to be using the services strategy when validating pod name generation? In controllers/types.go they're technically the same strategy, just struck me as strange when reading it.

—
Reply to this email directly or view it on GitHub.

pweil- · 2015-02-25T15:26:55Z

My eyes...they bleed! LGTM Some things that jumped out at me that may affect us:

Changes to how ListWatchers are created
Storage config in master has moved from fields to a map
Change to endpoints from a string to a defined type

I can no longer read any secrets code without picturing @pmorie standing behind me whispering "shhh!"

smarterclayton · 2015-02-26T03:45:58Z

The router is fixed in my last commit. Also, you don't have to review the merge commits :). Just my fixes.

On Feb 25, 2015, at 7:27 AM, Paul notifications@github.com wrote:

My eyes...they bleed! LGTM Some things that jumped out at me that may affect us:

Changes to how ListWatchers are created
Storage config in master has moved from fields to a map
Change to endpoints from a string to a defined type
I can no longer read any secrets code without picturing @pmorie standing behind me whispering "shhh!"

—
Reply to this email directly or view it on GitHub.

smarterclayton · 2015-02-27T04:33:01Z

[test]

deads2k · 2015-02-27T13:18:15Z

I'll look at adding a proper error for #1135 today or Monday. I previously misunderstood and thought that self-linking was already tolerant upstream.

The number of required changes is much smaller than I anticipated. Is there going to be a card for someone to take advantage of the generic registry stuff?

…a0236f680ffa8e11037

… pflag

…ner logs

smarterclayton · 2015-03-02T04:04:01Z

@deads2k calling ResourceAccessReviews("") is disallowed now upstream (you can't invoke create on a namespace scoped resource with ""), which causes:

--- FAIL: TestResourceAccessReview (3.91s)
    authorization_test.go:188: expected forbidden error, but didn't get one
    authorization_test.go:194: unexpected error: request construction error: 'an empty namespace may not be set during creation'
    authorization_test.go:199: expected [system:cluster-admins], got []
    authorization_test.go:202: expected [system:kube-client system:openshift-client system:openshift-deployer], got []

My preference is to define GlobalResourceAccessReview() method that takes no namespace and does not set Namespace() on the Create method. Do you have an alternative suggestion?

Empty namespace is now disallowed as an action (you must not invoke request.Namespace() if you don't want a namespace set).

Fix router integration test

deads2k · 2015-03-03T12:48:11Z

My preference is to define GlobalResourceAccessReview() method that takes no namespace and does not set Namespace() on the Create method. Do you have an alternative suggestion

I'd have both. A GlobalResourceAccessReview for cluster-admins who have rights to access resources across all namespaces and a ResourceAccessReview for project admins who only have rights to a single namespace. I had heard rumors that global objects were going to be eliminated...

@derekwaynecarr Will we still have globally objects in a month?

liggitt · 2015-03-03T14:19:00Z

@smarterclayton fyi, I need the secrets stuff in this for #1208

smarterclayton · 2015-03-03T14:43:31Z

Brenton is going to cut beta2 before this lands. Talk to him and make sure the other stuff for beta2 is ready to land.

smarterclayton · 2015-03-04T16:13:43Z

[merge]

openshift-bot · 2015-03-04T16:16:39Z

continuous-integration/openshift-jenkins/merge SUCCESS (https://ci.openshift.redhat.com/jenkins/job/merge_pull_requests_openshift3/1102/) (Image: devenv-fedora_966)

openshift-bot · 2015-03-04T16:16:39Z

Evaluated for origin up to 1c06be8

Merged by openshift-bot

…-3-8 UPSTREAM: 58433: lstat with abs path of parent

pweil- reviewed Feb 25, 2015
View reviewed changes

smarterclayton force-pushed the rebase_kube branch 2 times, most recently from 4f08824 to 6e59b78 Compare February 27, 2015 03:07

smarterclayton and others added 11 commits March 1, 2015 22:36

bump(github.com/GoogleCloudPlatform/kubernetes):6241a211c8f35a6147aa3…

4619ce1

…a0236f680ffa8e11037

UPSTREAM: Add 'release' field to raven-go

1fa6172

UPSTREAM: Disable UIs for Kubernetes and etcd

f5a12f0

UPSTREAM: Disable auto-pull when tag is "latest"

a61fac8

UPSTREAM: need to make sure --help flags is registered before calling…

da8a973

… pflag

UPSTREAM: special command "help" must be aware of context

0b6ceba

UPSTREAM: Make setSelfLink not bail out

2e9532d

Remove references to old volume source, handle endpoints change

4610229

UPSTREAM: Relax constraints on container status while fetching contai…

d8ca27b

…ner logs

UPSTREAM: Validate TCPSocket handler correctly

67a98a0

Better output for integration tests

d1659cb

smarterclayton force-pushed the rebase_kube branch from 6e59b78 to d1659cb Compare March 2, 2015 04:01

Add RootResourceAccessReview to replace use of empty namespace

db19185

Empty namespace is now disallowed as an action (you must not invoke request.Namespace() if you don't want a namespace set).

smarterclayton added this to the 0.5.0 milestone Mar 2, 2015

Fix router integration test

1941235

Merge pull request #6 from liggitt/rebase

1c06be8

Fix router integration test

0xmichalis mentioned this pull request Mar 4, 2015

Implement logs streaming option when starting a build #1197

Merged

openshift-bot pushed a commit that referenced this pull request Mar 4, 2015

Merge pull request #1136 from smarterclayton/rebase_kube

c657948

Merged by openshift-bot

openshift-bot merged commit c657948 into openshift:master Mar 4, 2015

sjenning pushed a commit to sjenning/origin that referenced this pull request Jul 30, 2018

Merge pull request openshift#1136 from gnufied/fix-subpath-file-mount…

9d5ba3c

…-3-8 UPSTREAM: 58433: lstat with abs path of parent

Conversation

smarterclayton commented Feb 25, 2015

Uh oh!

smarterclayton commented Feb 25, 2015

Uh oh!

smarterclayton commented Feb 25, 2015

Uh oh!

smarterclayton commented Feb 25, 2015

Uh oh!

smarterclayton commented Feb 25, 2015

Uh oh!

openshift-bot commented Feb 25, 2015

Uh oh!

smarterclayton commented Feb 25, 2015

Uh oh!

pweil- Feb 25, 2015

Choose a reason for hiding this comment

Uh oh!

pweil- Feb 25, 2015

Choose a reason for hiding this comment

Uh oh!

smarterclayton Feb 26, 2015

Choose a reason for hiding this comment

Uh oh!

pweil- commented Feb 25, 2015

Uh oh!

smarterclayton commented Feb 26, 2015

Uh oh!

smarterclayton commented Feb 27, 2015

Uh oh!

deads2k commented Feb 27, 2015

Uh oh!

smarterclayton commented Mar 2, 2015

Uh oh!

deads2k commented Mar 3, 2015

Uh oh!

liggitt commented Mar 3, 2015

Uh oh!

smarterclayton commented Mar 3, 2015

Uh oh!

smarterclayton commented Mar 4, 2015

Uh oh!

openshift-bot commented Mar 4, 2015

Uh oh!

openshift-bot commented Mar 4, 2015

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

7 participants