Conversation
pweil-
force-pushed
the
rebase
branch
4 times, most recently
from
42b25d2 to
688c92a
Compare
|
Can we pull over NamespaceLifecycle admission control plugin as part of this change? |
|
I pulled a mulligan this morning so I should have everything as of a couple hours ago in my next commit. Will that cover it? |
|
If you can add an anonymous import to the new plugin here: https://github.com/openshift/origin/blob/master/pkg/cmd/server/start/start_allinone.go It should cause it to be pulled over. |
|
ok, I'll do that when I get the unit tests stable. |
|
Thanks! |
pweil-
force-pushed
the
rebase
branch
2 times, most recently
from
f4f3576 to
12c8b9f
Compare
|
@smarterclayton @pmorie @abhgupta @fabianofranz - I think I'm getting close (I hope). I had a successful run of test-go and test-cmd with what is currently in this pr and will be checking out test-end-to-end next @derekwaynecarr I noticed a weird issue with deleting a project in test-cmd. It went into the terminating state as expected but hung there for at least a minute. Issuing a second delete killed it immediately so my last commit has 2 deletes in test-cmd.sh which is wrong (but working). Any advice there? |
|
@pwell - it will hang for a minute until the namespace controller detects that its terminating and cleans up the content in the project. I can't really improve this to make it faster until we get this rebased because origin needs to also take part in the clean-up life-cycle. |
pweil-
force-pushed
the
rebase
branch
2 times, most recently
from
a000a81 to
a887668
Compare
…5fbe1dc1b1971266240
Need to work with go-systemd to see if we can abstract the syscall.CloseOnExec which is not supported on windows.
Conflicts: Godeps/_workspace/src/github.com/GoogleCloudPlatform/kubernetes/pkg/apiserver/resthandler.go
Conflicts: Godeps/_workspace/src/github.com/GoogleCloudPlatform/kubernetes/pkg/kubelet/kubelet.go
…explicitly requested Conflicts: Godeps/_workspace/src/github.com/spf13/cobra/command.go
|
@smarterclayton @pmorie @abhgupta @fabianofranz PTAL - light is green, think this is ready for a review. 2nd to last commit was a merge from master to pick up today's PRs. If that was incorrect please let me know the proper procedure (hopefully not a start over but if so that's not terrible now that I've started over about 10 times) |
|
That's all the comments I had |
|
[test] |
|
continuous-integration/openshift-jenkins/test SUCCESS (https://ci.openshift.redhat.com/jenkins/job/test_pull_requests_openshift3/1533/) |
|
LGTM after squashing. |
|
kicked jenkins |
Something seriously wrong there. |
|
Evaluated for origin up to a24b130 |
|
Light is green, issue was fixed upstream so I've committed it with an UPSTREAM tag. There are still quite a few of the |
|
I thought host was immutable on pod. Will look to see if that is case in Kube |
|
So somewhere we are sending a pod with a host that is set and then we are attempting to update some other part of status ( maybe a label? ) and we are not passing the set host. |
|
It should be (for now, there is a TODO above the validation). I do not get these errors on test-cmd.sh or test-e2e locally for some reason and Jenkins isn't seeing it either. |
|
I'm going to cut a release and then prepare to merge. Do any cleanup you need to this PR ----- Original Message -----
|
|
We think the status_manager error is due to lack of this commit: kubernetes/kubernetes@372c5c9#diff-bf28da68f62a8df6e99e447c4351122d It only affects reporting status while a pod is in pending state and docker container info cant be queried. |
|
@smarterclayton @derekwaynecarr @pweil- @ironcladlou I just discussed squashing this branch with @pweil- at some length on the phone. I do not think there is a good way to squash these commits without doing a soft reset and breaking the changes after the godep bump and upstream patches into a single commits or commits based on sets of related packages. The downside is that we will lose visibility into the commits pulled in and there will then be a rebase conflict with master (if I understand correctly). If we want to keep this as a PR which is merged normally, it should probably go in as-is (with maybe an additional commit to clean up any things in the state of the branch which need it). The alternative is squashing, which is going to have a ton of merge issues that need to be sorted out one by one. |
|
testing out a cleaned up version here: #1519. WTF did you do paul:
|
|
closed in favor of #1519 |
|
Changes Unknown when pulling a24b130 on pweil-:rebase into * on openshift:master*. |
9 participants
WIP for the rebase. Tagging for reviewers
@smarterclayton @pmorie @abhgupta @fabianofranz
The WTF did you do Paul? tracker: https://gist.github.com/pweil-/0753309e70f0604dd758 (rebase actions only, this gist doesn't contain any refactoring info)