Use upgraded container-vm by default on worker nodes for GCE k8s clusters #31023
Conversation
Signed-off-by: Vishnu kannan <vishnuk@google.com>
vishh
added
the
do-not-merge
|
cc @dchen1107 |
|
cc @pwittrock |
k8s-github-robot
added
size/XS
|
I want to make sure all of us on the same page:
|
|
@dchen1107 kubernetes/test-infra#420 is the PR that should enable most of those critical tests. Over the next week, I hope to get to complete test parity. |
vishh
added
release-note-action-required
vishh
changed the title
|
All critical tests have been added at this stage. Let's wait for at-least a day to get enough test data. With that, I'm hoping that we can make an informed decision on the impact of this change. This might push this PR beyond the freeze date. Apologies! |
|
We are holding the PR for testing results. |
|
All critical tests excepting GKE ones are either passing or are not more
flaky than container-vm ones.
#31330 is out to fix the
GKE-GCI tests.
|
|
Update: The GKE tests are running now. Waiting for a couple of runs to gather some concrete data. |
|
Update: #31456 is the only major blockers. All other tests are either passing or are being fixed. |
|
One more issue to track - #31465 |
|
@vishh and I went over all critical builds including GCE & GKE, and most of tests are passed. Several issues were filed separately, and we are going to address them one-by-one. I am going to unblock this pr for now. cc/ @pwittrock, the release czar for 1.4. |
|
/approved-for-1.4 |
dchen1107
removed
the
do-not-merge
|
LGTM |
dchen1107
added
the
lgtm
|
GCE e2e build/test passed for commit ff70760. |
|
Thanks @dchen1107 and @pwittrock On Thu, Aug 25, 2016 at 3:29 PM, Kubernetes Bot notifications@github.com
|
|
Does this not break GlusterFS, CephFS and possibly other Volumes from ever working? |
vishh
changed the title
|
@owenhaynes This PR only changes the default for now. It will still be possible to use the old version of container-vm and one can install gluster or ceph packages. I hope that containerized volume mounters will be ready for production before the old container-vm will be deprecated. |
|
Automatic merge from submit-queue |
|
Please see #32151 (comment) This PR broke CNI on GCE deployments. We either need to get the above PR merged and added to the v1.4 milestone, or revert this change. |
|
cc @thockin |
thockin
added
the
priority/important-soon
|
ACK. @dchen1107 @vishh Please file an issue for the 1.4 milestone and come up with a resolution. We need the resolution merged into 1.4 by Wednesday and then QA'ed manually on the beta build. |
|
PR to set the flag is LGTM'ed, but calico is still broken |
Automatic merge from submit-queue Add flag to set CNI bin dir, and use it on gci nodes **What this PR does / why we need it**: When using `kube-up` on GCE, following #31023 which moved the workers from debian to gci, CNI just isn't working. The root cause is basically as discussed in #28563: one flag (`--network-plugin-dir`) means two different things, and the `configure-helper` script uses it for the wrong purpose. This PR adds a new flag `--cni-bin-dir`, then uses it to configure CNI as desired. As discussed at #28563, I have also added a flag `--cni-conf-dir` so users can be explicit **Which issue this PR fixes** : fixes #28563 **Special notes for your reviewer**: I left the old flag largely alone for backwards-compatibility, with the exception that I stop setting the default when CNI is in use. The value of `"/usr/libexec/kubernetes/kubelet-plugins/net/exec/"` is unlikely to be what is wanted there. **Release note**: ```release-note Added new kubelet flags `--cni-bin-dir` and `--cni-conf-dir` to specify where CNI files are located. Fixed CNI configuration on GCI platform when using CNI. ```
|
@thockin is there an issue for Calico stuff? On Tue, Sep 13, 2016 at 4:41 PM, Tim Hockin notifications@github.com
|
|
@vishh - Will raise one now and report back |
Automatic merge from submit-queue Add flag to set CNI bin dir, and use it on gci nodes **What this PR does / why we need it**: When using `kube-up` on GCE, following kubernetes#31023 which moved the workers from debian to gci, CNI just isn't working. The root cause is basically as discussed in kubernetes#28563: one flag (`--network-plugin-dir`) means two different things, and the `configure-helper` script uses it for the wrong purpose. This PR adds a new flag `--cni-bin-dir`, then uses it to configure CNI as desired. As discussed at kubernetes#28563, I have also added a flag `--cni-conf-dir` so users can be explicit **Which issue this PR fixes** : fixes kubernetes#28563 **Special notes for your reviewer**: I left the old flag largely alone for backwards-compatibility, with the exception that I stop setting the default when CNI is in use. The value of `"/usr/libexec/kubernetes/kubelet-plugins/net/exec/"` is unlikely to be what is wanted there. **Release note**: ```release-note Added new kubelet flags `--cni-bin-dir` and `--cni-conf-dir` to specify where CNI files are located. Fixed CNI configuration on GCI platform when using CNI. ``` (cherry picked from commit c4893df)
This reverts PR kubernetes#31023, which had made GCI the default node image for open source. This revert makes container-vm the default for open source again.
…again Automatic merge from submit-queue Revert "Merge pull request kubernetes#31023 from vishh/gci-default" This reverts PR kubernetes#31023, which had made GCI the default node image for open source. This revert makes container-vm the default for open source again. (cherry picked from commit 3718fd1)
|
@vishh @dchen1107 this PR has a release-note-action-required label on it, but I can't tell what the action is, CHANGELOG.md only has the PR title in it can you please clarify what the action required is? |
vishh
added
release-note-none
|
@spiffxp This PR got reverted. Updated release-note label. |
Automatic merge from submit-queue Add flag to set CNI bin dir, and use it on gci nodes **What this PR does / why we need it**: When using `kube-up` on GCE, following kubernetes#31023 which moved the workers from debian to gci, CNI just isn't working. The root cause is basically as discussed in kubernetes#28563: one flag (`--network-plugin-dir`) means two different things, and the `configure-helper` script uses it for the wrong purpose. This PR adds a new flag `--cni-bin-dir`, then uses it to configure CNI as desired. As discussed at kubernetes#28563, I have also added a flag `--cni-conf-dir` so users can be explicit **Which issue this PR fixes** : fixes kubernetes#28563 **Special notes for your reviewer**: I left the old flag largely alone for backwards-compatibility, with the exception that I stop setting the default when CNI is in use. The value of `"/usr/libexec/kubernetes/kubelet-plugins/net/exec/"` is unlikely to be what is wanted there. **Release note**: ```release-note Added new kubelet flags `--cni-bin-dir` and `--cni-conf-dir` to specify where CNI files are located. Fixed CNI configuration on GCI platform when using CNI. ``` (cherry picked from commit c4893df)
…again Automatic merge from submit-queue Revert "Merge pull request kubernetes#31023 from vishh/gci-default" This reverts PR kubernetes#31023, which had made GCI the default node image for open source. This revert makes container-vm the default for open source again. (cherry picked from commit 3718fd1)
For #25276
Depends on kubernetes/test-infra#417
This change is