Fix working_set calculation in kubelet #29153

vishh · 2016-07-18T23:58:53Z

Bump cadvisor dependencies to latest head.

vishh · 2016-07-19T00:00:31Z

timstclair · 2016-07-19T19:47:10Z

Godeps/Godeps.json

@@ -2160,6 +2160,481 @@
 			"ImportPath": "k8s.io/heapster/metrics/apis/metrics/v1alpha1",
 			"Comment": "v1.1.0-beta2",
 			"Rev": "9cb18ac0ceb193eb530a1fe339355c94ea454d85"
+		},
+		{
+			"ImportPath": "github.com/google/cadvisor/vendor/github.com/golang/glog",


Why are the vendor dependencies getting pulled in?

I'm not sure why godep is doing that. I'm trying to figure out why.

timstclair · 2016-07-19T23:41:52Z

@Random-Liu

Random-Liu · 2016-07-20T01:20:40Z

@vishh If you run godep save twice, the dlopen will be in.
Seems similar with kubernetes/node-problem-detector#7

vishh · 2016-07-20T03:15:32Z

@Random-Liu that seems plausible. Maybe godep is not detecting recursive dependencies in the first run.

vishh · 2016-07-20T03:16:33Z

However, the godep-verify.sh script is going to run godep save just once. I wonder if the script will be happy with running godep save twice.

vishh · 2016-07-20T03:20:51Z

I ran godep-save.sh again and noticed no changes on top of my PR.

Random-Liu · 2016-07-20T03:41:13Z

@vishh It's wired. It works for me, maybe because I changed runc to the same version align with docker v1.11.2? https://github.com/docker/docker/tree/v1.11.2/vendor/src/github.com/opencontainers/runc

Random-Liu · 2016-07-20T03:42:33Z

@vishh However, I just run godep-verify.sh and it also failed and told me to remove dlopen from Godeps...

Your Godeps.json is different:
--- /usr/local/google/home/lantaol/workspace/src/k8s.io/kubernetes/Godeps/Godeps.json   2016-07-19 18:11:32.692929121 -0700
+++ /tmp/gopath.ppLOVU/src/k8s.io/kubernetes/Godeps/Godeps.json 2016-07-19 20:40:34.219648761 -0700
@@ -497,11 +497,6 @@
            "Rev": "8b320e7c550067b1dfb37bd1682e8067023e0751"
        },
        {
-           "ImportPath": "github.com/coreos/go-etcd/etcd",
-           "Comment": "v2.0.0-38-g003851b",
-           "Rev": "003851be7bb0694fe3cc457a49529a19388ee7cf"
-       },
-       {
            "ImportPath": "github.com/coreos/go-oidc/http",
            "Rev": "5cf2aa52da8c574d3aa4458f471ad6ae2240fe6b"
        },
@@ -560,11 +555,6 @@
            "Comment": "v2",
            "Rev": "7f080b6c11ac2d2347c3cd7521e810207ea1a041"
        },
-       {
-           "ImportPath": "github.com/coreos/pkg/dlopen",
-           "Comment": "v2",
-           "Rev": "7f080b6c11ac2d2347c3cd7521e810207ea1a041"
-       },
        {
            "ImportPath": "github.com/coreos/pkg/health",
            "Comment": "v2",

Random-Liu · 2016-07-20T03:43:38Z

@vishh BTW, I don't know why my local godep keeps trying to add go-etcd/etcd...

I've to go home now. Will have another try after I get home and finish dinner. :)

vishh · 2016-07-20T05:37:56Z

@Random-Liu I have updated the presubmit script to run godep twice. I hope this works for now.

vishh · 2016-07-20T14:59:44Z

@k8s-bot test this github issue #IGNORE

Random-Liu · 2016-07-20T17:02:07Z

@freeformz We encountered the same issue again. We have to run godep save twice to include the dependency github.com/coreos/pkg/dlopen.

Any suggestions will be appreciate. :'(

Signed-off-by: Vishnu kannan <vishnuk@google.com>

vishh · 2016-07-20T17:08:29Z

For some reason the ginkgo dependencies were also updated as part of this PR and resulted in an older version of ginkgo as a dep. I tried to fix it now.

vishh · 2016-07-20T17:50:17Z

@timstclair @Random-Liu This PR is finally ready to be merged. PTAL

Random-Liu · 2016-07-20T18:00:16Z

hack/godep-save.sh

@@ -31,4 +31,5 @@ REQUIRED_BINS=(

 pushd "${KUBE_ROOT}" > /dev/null
  GO15VENDOREXPERIMENT=1 ${GODEP} save "${REQUIRED_BINS[@]}"
+  GO15VENDOREXPERIMENT=1 ${GODEP} save "${REQUIRED_BINS[@]}"


Is this for temporary walkaround? If so, maybe we need some comments here to explain or add a TODO.

Random-Liu · 2016-07-20T18:00:42Z

Fixes #28997.

Signed-off-by: Vishnu kannan <vishnuk@google.com>

Random-Liu · 2016-07-20T18:14:30Z

LGTM. :)

k8s-bot · 2016-07-20T18:48:26Z

GCE e2e build/test passed for commit feb7321.

k8s-github-robot · 2016-07-20T20:37:53Z

@k8s-bot test this [submit-queue is verifying that this PR is safe to merge]

k8s-bot · 2016-07-20T21:12:15Z

GCE e2e build/test passed for commit feb7321.

k8s-github-robot · 2016-07-20T21:17:07Z

Automatic merge from submit-queue

apelisse · 2016-07-20T23:56:57Z

This seem to have broken the build.

vishh · 2016-07-21T00:29:15Z

I wonder why this PR was even merged in the first place? Is cross
compilation not required to succeed before merging?
cc @lavalamp

On Wed, Jul 20, 2016 at 4:57 PM, Antoine Pelisse notifications@github.com
wrote:

This seem to have broken the build.

—
You are receiving this because you were mentioned.
Reply to this email directly, view it on GitHub
#29153 (comment),
or mute the thread
https://github.com/notifications/unsubscribe-auth/AGvIKC7WbkzGXQ0GallJQm22ZiD3jCueks5qXrXsgaJpZM4JPQQa
.

lavalamp · 2016-07-21T01:00:11Z

it was too slow-- we need a separate presubmit to test that.

vishh · 2016-07-21T01:03:52Z

Just the cross-compilation part or make release? Can we try
cross-compiling just cmd/kubectl as part of presubmit?

On Wed, Jul 20, 2016 at 6:00 PM, Daniel Smith notifications@github.com
wrote:

it was too slow-- we need a separate presubmit to test that.

—
You are receiving this because you were mentioned.
Reply to this email directly, view it on GitHub
#29153 (comment),
or mute the thread
https://github.com/notifications/unsubscribe-auth/AGvIKFgSyz7KwJrJOTuN-dDWvo_ufol_ks5qXsS9gaJpZM4JPQQa
.

lavalamp · 2016-07-21T01:14:35Z

cross-compile specifically. Can't make existing presubmits slower than they already are. Single-platform make quick-release is already run.

@vishh

Automatic merge from submit-queue Bump cadvisor dependencies to latest head. Fixes #28619 Fixes #28997 This is another try of #29153. To update cadvisor godeps, we did: * Bump up docker version to v1.11.2 for both cadvisor [google/cadvisor#1388] and k8s. * Bump up cadvisor `go-systemd` version to be the same with k8s [google/cadvisor#1390]. Or else, a package `github.com/coreos/pkg/dlopen` will be removed by Godep, because it is used by new `go-systemd` in k8s, but not used by old `go-systemd` in cadvisor. * Bump up runc version to be the same with docker v1.11.2 just in case. * Add `github.com/Azure/go-ansiterm` dependency which is needed by docker v1.11.2. * Change `pkg/util/term/`, because `SetWinsSize` is removed from windows platform in docker v1.11.2. [The first commit] @vishh /cc @ncdc for the `pkg/util/term` change.

googlebot added the cla: yes label Jul 18, 2016

vishh added the release-note Denotes a PR that will be considered when it comes time to generate release notes. label Jul 18, 2016

vishh changed the title ~~Bump cadvisor dependencies to v0.23.7 tag.~~ Fix working_set calculation in cAdvisor Jul 18, 2016

vishh changed the title ~~Fix working_set calculation in cAdvisor~~ Fix working_set calculation in kubelet Jul 19, 2016

vishh added the cherrypick-candidate label Jul 19, 2016

vishh added this to the v1.3 milestone Jul 19, 2016

vishh added the priority/important-soon Must be staffed and worked on either currently, or very soon, ideally in time for the next release. label Jul 19, 2016

vishh assigned timstclair Jul 19, 2016

k8s-github-robot added the size/XS Denotes a PR that changes 0-9 lines, ignoring generated files. label Jul 19, 2016

vishh force-pushed the cadvisor-bump branch from c3fc5da to de05072 Compare July 19, 2016 00:17

fabioy mentioned this pull request Jul 19, 2016

Fix working_set calculation in kubelet #29154

Merged

vishh force-pushed the cadvisor-bump branch 2 times, most recently from d5cb7b2 to ce722fe Compare July 19, 2016 19:42

timstclair reviewed Jul 19, 2016
View reviewed changes

vishh force-pushed the cadvisor-bump branch 2 times, most recently from 12ea7a7 to c0b482a Compare July 20, 2016 00:33

vishh force-pushed the cadvisor-bump branch from 4469961 to 310bee3 Compare July 20, 2016 05:38

vishh assigned Random-Liu Jul 20, 2016

updating cadvisor deps

cdeaef2

Signed-off-by: Vishnu kannan <vishnuk@google.com>

vishh force-pushed the cadvisor-bump branch from 310bee3 to 5a60c64 Compare July 20, 2016 17:06

Random-Liu reviewed Jul 20, 2016
View reviewed changes

run godeps twice in validation to include recursive dependencies

feb7321

Signed-off-by: Vishnu kannan <vishnuk@google.com>

vishh force-pushed the cadvisor-bump branch from 5a60c64 to feb7321 Compare July 20, 2016 18:12

Random-Liu added the lgtm "Looks good to me", indicates that a PR is ready to be merged. label Jul 20, 2016

k8s-github-robot merged commit 115ae62 into kubernetes:master Jul 20, 2016

apelisse mentioned this pull request Jul 21, 2016

Revert "Fix working_set calculation in kubelet" #29349

Merged

This was referenced Jul 21, 2016

Bump cadvisor dependencies to latest head. #29389

Closed

Bump cadvisor dependencies to latest head. #29492

Merged

fabioy removed the cherrypick-candidate label Jul 28, 2016

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Fix working_set calculation in kubelet #29153

Fix working_set calculation in kubelet #29153

vishh commented Jul 18, 2016 •

edited

Loading

vishh commented Jul 19, 2016

timstclair Jul 19, 2016

vishh Jul 19, 2016

timstclair commented Jul 19, 2016

Random-Liu commented Jul 20, 2016

vishh commented Jul 20, 2016

vishh commented Jul 20, 2016

vishh commented Jul 20, 2016

Random-Liu commented Jul 20, 2016

Random-Liu commented Jul 20, 2016 •

edited

Loading

Random-Liu commented Jul 20, 2016

vishh commented Jul 20, 2016

vishh commented Jul 20, 2016

Random-Liu commented Jul 20, 2016 •

edited

Loading

vishh commented Jul 20, 2016

vishh commented Jul 20, 2016

Random-Liu Jul 20, 2016

vishh Jul 20, 2016

Random-Liu commented Jul 20, 2016

Random-Liu commented Jul 20, 2016

k8s-bot commented Jul 20, 2016

k8s-github-robot commented Jul 20, 2016

k8s-bot commented Jul 20, 2016

k8s-github-robot commented Jul 20, 2016

apelisse commented Jul 20, 2016

vishh commented Jul 21, 2016

lavalamp commented Jul 21, 2016

vishh commented Jul 21, 2016

lavalamp commented Jul 21, 2016

Fix working_set calculation in kubelet #29153

Fix working_set calculation in kubelet #29153

Conversation

vishh commented Jul 18, 2016 • edited Loading

vishh commented Jul 19, 2016

timstclair Jul 19, 2016

Choose a reason for hiding this comment

vishh Jul 19, 2016

Choose a reason for hiding this comment

timstclair commented Jul 19, 2016

Random-Liu commented Jul 20, 2016

vishh commented Jul 20, 2016

vishh commented Jul 20, 2016

vishh commented Jul 20, 2016

Random-Liu commented Jul 20, 2016

Random-Liu commented Jul 20, 2016 • edited Loading

Random-Liu commented Jul 20, 2016

vishh commented Jul 20, 2016

vishh commented Jul 20, 2016

Random-Liu commented Jul 20, 2016 • edited Loading

vishh commented Jul 20, 2016

vishh commented Jul 20, 2016

Random-Liu Jul 20, 2016

Choose a reason for hiding this comment

vishh Jul 20, 2016

Choose a reason for hiding this comment

Random-Liu commented Jul 20, 2016

Random-Liu commented Jul 20, 2016

k8s-bot commented Jul 20, 2016

k8s-github-robot commented Jul 20, 2016

k8s-bot commented Jul 20, 2016

k8s-github-robot commented Jul 20, 2016

apelisse commented Jul 20, 2016

vishh commented Jul 21, 2016

lavalamp commented Jul 21, 2016

vishh commented Jul 21, 2016

lavalamp commented Jul 21, 2016

vishh commented Jul 18, 2016 •

edited

Loading

Random-Liu commented Jul 20, 2016 •

edited

Loading

Random-Liu commented Jul 20, 2016 •

edited

Loading