The debian-iptables v12.0.0 build didn't work because of another
previously-undiscovered iptables 1.8.x bug. Work around it for now and
bump the version to v12.0.1; we can revert back to the original
version of the script once iptables 1.8.4 is available in
buster-backports.
default behavior does not change, it uses k8s.gcr.io by default
added two vars: KUBE_DOCKER_REGISTRY, KUBE_BASE_IMAGE_REGISTRY.
KUBE_BASE_IMAGE_REGISTRY is for base image registry of server binaries
KUBE_DOCKER_REGISTRY is for released images registry
user can interact with them by:
`KUBE_DOCKER_REGISTRY=### KUBE_BASE_IMAGE_REGISTRY=### make quick-release`
Signed-off-by: Hui Luo <luoh@vmware.com>
Automatic merge from submit-queue. If you want to cherry-pick this change to another branch, please follow the instructions here: https://github.com/kubernetes/community/blob/master/contributors/devel/cherry-picks.md.
Add ability to build with runtime coverage instrumentation
**What this PR does / why we need it**:
This PR adds the ability to instrument a subset of kubernetes binaries to report code coverage information. The specific use-case is to help determine coverage of our end-to-end Conformance tests, as well as provide data that can be used to help determine where to focus. This PR focuses on making it possible to build with instrumentation; collecting and using the generated coverage data will be done in later PRs. For more details as to the intent, see the [design doc](https://docs.google.com/document/d/1FKMBFxz7vtA-6ZgUkA47F8m6yR00fwqLcXMVJqsHt0g/edit?usp=sharing) (google doc; requires kubernetes-dev membership).
Specifically, this PR adds a new `KUBE_BUILD_WITH_COVERAGE` make variable, which when set will cause `kube-apiserver`, `kube-controller-manager`, `kube-scheduler`, `kube-proxy` and `kubelet` to be built with coverage instrumentation. These coverage-instrumented binaries will flush coverage information to disk every five seconds, defaulting to a temporary directory unless the `KUBE_COVERAGE_FILE` environment variable is set at launch, in which case it will write to that file instead.
The mechanism used to achieve coverage instrumentation is to build the targeted binaries as "unit tests" with coverage enabled, and then rigging the unit tests to just execute the binary's usual entry point. This is implemented only for the bash build system.
/sig testing
```release-note
NONE
```
Automatic merge from submit-queue (batch tested with PRs 67396, 67097, 67395, 67365, 67099). If you want to cherry-pick this change to another branch, please follow the instructions <a href="https://github.com/kubernetes/community/blob/master/contributors/devel/cherry-picks.md">here</a>.
Update to debian-iptables v10.1 and hyperkube-base 0.10.1
**What this PR does / why we need it**: these images are based on the `debian-base` 0.3.2 images, which include CVE fixes (#67026) and permission fixes of the qemu-ARCH-static helper binary (#67026, #67283).
**Release note**:
```release-note
NONE
```
First version of this was only standalone, now the code is being
vendored by anyone who wants to use it. So the standalone binary and
container are no longer useful.
Change-Id: Ib9369de66b4ecb3451f73ba2a252526d6615b96f
Fix kubernetes build scripts to work in case the source directory is in
a directory path which contains a space.
You can prepare such a directory like this:
$ mkdir '/tmp/test dir/'
$ cd '/tmp/test dir/'
$ git clone https://github.com/kubernetes/kubernetes.git
$ cd kubernetes
Then, without the fix:
$ KUBE_FASTBUILD=true KUBE_RELEASE_RUN_TESTS=n build/release.sh
cat: /tmp/test: No such file or directory
cat: dir/kubernetes/build/build-image/cross/VERSION: No such file or directory
cat: /tmp/test: No such file or directory
cat: dír/kubernetes/build/build-image/VERSION: No such file or directory
+++ [0307 18:10:33] Verifying Prerequisites....
cp: target '/tmp/test dir/kubernetes/_output/images/kube-build:build-7c7cd10a18--/Dockerfile' is not a directory
!!! [0307 18:10:33] Call tree:
!!! [0307 18:10:33] 1: build/release.sh:35
kube::build::build_image(...)
!!! Error in build/../build/common.sh:454
Error in build/../build/common.sh:454. '((i<3-1))' exited with status 1
Call stack:
1: build/../build/common.sh:454 kube::build::build_image(...)
2: build/release.sh:35 main(...)
Exiting with status 1
With the fix the compilation succeeds. The fix is done adding double
quotes to required places (and also just in case to other places where
shellcheck recommended adding them).
Note that this fix doesn't help with the official make-based build:
it's tricky to make makefiles work with targets with spaces in their
names.
Automatic merge from submit-queue (batch tested with PRs 59705, 59207, 59677). If you want to cherry-pick this change to another branch, please follow the instructions <a href="https://github.com/kubernetes/community/blob/master/contributors/devel/cherry-picks.md">here</a>.
build: fix a logic error in shell script.
**What this PR does / why we need it**: It looks like that there's a logic error in `build/common.sh`. The return value of a `docker inspect` command is not checked properly, since the value being assigned is actually the previous command's return value (a `0`, because `local` always returns `0` when used like this).
**Which issue(s) this PR fixes** *(optional, in `fixes #<issue number>(, fixes #<issue_number>, ...)` format, will close the issue(s) when PR gets merged)*:
Fixes #
**Special notes for your reviewer**:
**Release note**:
```release-note
NONE
```
This is the 2nd attempt. The previous was reverted while we figured out
the regional mirrors (oops).
New plan: k8s.gcr.io is a read-only facade that auto-detects your source
region (us, eu, or asia for now) and pulls from the closest. To publish
an image, push k8s-staging.gcr.io and it will be synced to the regionals
automatically (similar to today). For now the staging is an alias to
gcr.io/google_containers (the legacy URL).
When we move off of google-owned projects (working on it), then we just
do a one-time sync, and change the google-internal config, and nobody
outside should notice.
We can, in parallel, change the auto-sync into a manual sync - send a PR
to "promote" something from staging, and a bot activates it. Nice and
visible, easy to keep track of.
The right place to assign the "docker inspect" return value is outside
of the subshell. The last return value was coming from something else
than the expected command.
Automatic merge from submit-queue (batch tested with PRs 54071, 58393). If you want to cherry-pick this change to another branch, please follow the instructions <a href="https://github.com/kubernetes/community/blob/master/contributors/devel/cherry-picks.md">here</a>.
Optionally launch build containers inside a specific parent cgroup
**What this PR does / why we need it**: Adds support to the build process for launching containers inside a specific parent cgroup via the `--cgroup-parent` docker arg. We needed this to enforce resource usage when building Kubernetes in our CI environment.
**Special notes for your reviewer**: None
**Release note**:
```release-note
NONE
```
Automatic merge from submit-queue (batch tested with PRs 52959, 53790). If you want to cherry-pick this change to another branch, please follow the instructions <a href="https://github.com/kubernetes/community/blob/master/contributors/devel/cherry-picks.md">here</a>.
Dynamically determine default docker machine memory
**What this PR does / why we need it**:
Currently, if using `docker-machine` for k8s docker operations on a Mac,
we'll create a VM with 4096MB of memory. The machine's RAM will be the
same regardless of the memory available on the local machine. For
example, if the user has 16GB on their local machine, the VM will still
only have 4GB of RAM.
Update the method for defining the `kube-dev` VM with `docker-machine`,
so we give it access to 50% of the total RAM.
**Release note**:
```release-note
NONE
```
Currently, if using `docker-machine` for k8s docker operations on a Mac,
we'll create a VM with 4096MB of memory. The machine's RAM will be the
same regardless of the memory available on the local machine. For
example, if the user has 16GB on their local machine, the VM will still
only have 4GB of RAM.
Update the method for defining the `kube-dev` VM with `docker-machine`,
so we give it access to 50% of the total RAM.
Automatic merge from submit-queue (batch tested with PRs 50602, 51561, 51703, 51748, 49142)
Use arm32v7|arm64v8 images instead of the deprecated armhf|aarch64 image organizations
**What this PR does / why we need it**:
**Which issue this PR fixes** *(optional, in `fixes #<issue number>(, fixes #<issue_number>, ...)` format, will close that issue when PR gets merged)*: fixes#50601
**Special notes for your reviewer**:
/assign @ixdy @jbeda @zmerlynn
**Release note**:
```release-note
Use arm32v7|arm64v8 images instead of the deprecated armhf|aarch64 image organizations
```
Automatic merge from submit-queue (batch tested with PRs 51666, 49829, 51058, 51004, 50938)
add the caller ip into rsync hosts allow list
fixes https://github.com/kubernetes/kubernetes/issues/51003.
This patch allows running build in pods.
```release-note
NONE
```
We don't really need the .git directory to be transferred. This
problem was introduced in the PR:
"Run the update-staging-godeps script inside a docker container"
Fixes#50272
KUBE_ROOT_HASH depends just on the host name and
directory path. So when working with branches, it could lead
to some confusion since the hash is the same even when
you switch from branch to branch. Let us use the git branch
information when we compute the short hash
Fixes#1801
This was broken when we moved to the build container, but no one
noticed. We also likely have another bug, which is that protobuf should
hard fail when we have fields that aren't assigned a tag.
Automatic merge from submit-queue
Improve the multiarch situation; armel => armhf; reenable pcc64le; remove the patched golang
**What this PR does / why we need it**:
- Improves the multiarch situation as described in #38067
- Tries to bump to go1.8 for arm (and later enable ppc64le)
- GOARM 6 => GOARM 7
- Remove the golang 1.7 patch
- armel => armhf
- Bump QEMU version to v2.7.0
**Release note**:
```release-note
Improve the ARM builds and make hyperkube on ARM working again by upgrading the Go version for ARM to go1.8beta2
```
@kubernetes/sig-testing-misc @jessfraz @ixdy @jbeda @david-mcmahon @pwittrock
from etcd.sh split the start process into validate fucntion + start function so that the validate piece can be reused elsewhere. the up-cluster script has been changed to remove duplicate docker logic to the one used in buid-tools/common.sh and the validate etcd function is now used here.
moved docker daemon check function to util.sh and made function name changes and upstream changes.
If you delete a source file, we want to reflect that in the build container. We
only use --delete going that one way as we don't want to accidentally delete
files in the user's source tree.
Automatic merge from submit-queue
Check for rsync and give friendlier message
Fixes#34300.
Not sure if #34309 is the same issue. Hopefully it is the same issue.
Automatic merge from submit-queue
Make sure rsync.sh is executable inside the build image
I kept having the build fail:
```console
$ make quick-release
+++ [1006 18:13:44] Verifying Prerequisites....
+++ [1006 18:13:44] Building Docker image kube-build:build-d3c60cf83f-3-v1.6.3-9
+++ [1006 18:13:54] Creating data container kube-build-data-d3c60cf83f-3-v1.6.3-9
+++ [1006 18:13:55] Syncing sources to container
!!! [1006 18:16:01] Could not connect to rsync container. See build/README.md for setting up remote Docker engine.
make: *** [quick-release] Error 1
```
`docker ps` revealed the issue:
```console
$ docker ps
CONTAINER ID IMAGE COMMAND CREATED STATUS PORTS NAMES
$ docker ps -a
CONTAINER ID IMAGE COMMAND CREATED STATUS PORTS NAMES
75c2a3c40cb3 kube-build:build-d3c60cf83f-3-v1.6.3-9 "/rsyncd.sh" 6 seconds ago Exited (126) 5 seconds ago kube-rsync-d3c60cf83f-3-v1.6.3-9
3eb215e41f36 kube-build:build-d3c60cf83f-3-v1.6.3-9 "chown -R 85078.5000 " 8 seconds ago Exited (0) 6 seconds ago kube-build-data-d3c60cf83f-3-v1.6.3-9
5a2707af2ccd 882577c54f67 "/bin/sh -c 'cd ${K8S" 7 days ago Exited (2) 7 days ago stupefied_goldberg
$ docker logs 75c2a3c40cb3
/bin/bash: /rsyncd.sh: Permission denied
```
I'm not sure why this works on Jenkins but not on my machine.
We were using netcat to try and figure out when the rsync container is ready. Now we instead use rsync itself. I suspect that there was a race condition with some versions of Docker where it would accept connections and then close them during container start.
This fixes#34214 (I think)
make-generated-{protobuf,runtime}.sh was doing some really nasty stuff with how
the build container was managed in order to copy results out. Since we have
more flexibility to grab results out of the build container, we can now avoid
all of this. Ideally we wouldn't have `hack` calling `build` at all, but we
aren't there yet.
We also add "version" to all docker images and containers
This version is to be incremented manually when we change the shape of the build
image (like changing the golang version or the set of volumes in the data
container). This will delete all older versions of images and containers when
the version is different.
Automatic merge from submit-queue
Separate federation build.sh into development and deployment scripts.
The idea behind this separation is that it provides a clear distinction
between the dev environment and the prod environment. The
deploy/deploy.sh script will be shipped to the users, but
develop/develop.sh will be purely for development purposes and won't
be part of a release distribution.
Purely for developer convenience, all the deployment functionality is
made available through the develop/develop.sh script.
This change also copies deploy/* files into the release distribution.
cc @kubernetes/sig-cluster-federation @colhom
```release-note
Federation can now be deployed using the `federation/deploy/deploy.sh` script. This script does not depend on any of the development environment shell library/scripts. This is an alternative to the current `federation-up.sh`/`federation-down.sh` scripts. Both the scripts are going to co-exist in this release, but the `federation-up.sh`/`federation-down.sh` scripts might be removed in a future release in favor of `federation/deploy/deploy.sh` script.
```
Also build the hyperkube docker image on-the-fly.
This is only a temporary fix until the proposal in issue
https://github.com/kubernetes/kubernetes/issues/28630 is implemented.
Also, the new build/deployment method completely obviates this step.
We use debian image instead of busybox and do not build hyperkube as a
static binary yet. Wait until PR
https://github.com/kubernetes/kubernetes/pull/26028 is merged to build
static hyperkube binaries.
The idea behind this separation is that it provides a clear distinction
between the dev environment and the prod environment. The
deploy/deploy.sh script will be shipped to the users, but
develop/develop.sh will be purely for development purposes and won't
be part of a release distribution.
Purely for developer convenience, all the deployment functionality is
made available through the develop/develop.sh script.
This change also copies deploy/* files into the release distribution.
Automatic merge from submit-queue
Disable linux/ppc64le compilation by default
Work-around for #30384.
I'm still testing this locally to see if it actually works. The build is slow. (PR Jenkins won't tell us whether this fixes ppc.)
cc @Random-Liu @spxtr @david-mcmahon @luxas
Repro case:
$ make clean generated_files
$ hack/update-generated-protobuf.sh
This would complain about not finding `fmt`, and it was indicating the wrong
GOROOT. The problem was that the first step built binaries for generating
code, which *embeds* the value of GOROOT into the binary. The whole tree was
bind-mounted into the build container and then JUST the dockerized dir was
mounted over it. The in-container build tried to use the existing binaries,
but GOROOT is wrong.
This change whites-out the whole _output dir.
I first made just an anonymous volume for _output, but docker makes that as
root, which means I can't write to it from our non-root build. So I just put
it in the data container. This seems to work. The biggest change this makes
is that the $GOPATH/bin/ and $GOPATH/pkg/ dirs will persist across dockerized
builds.
This commit removes a part of common.sh script which copied
contrib/ sources for enabled contribs, which resulted in the
duplicated files inside tarball.
Fixes#30150
Automatic merge from submit-queue
Install go-bindata in cross-build image
Another follow-up to #25584.
We need `go-bindata` to create `test/e2e/generated`, and downloading it with `go get` at build time is painful for a variety of reasons. We can just include it in the cross-build image and not worry about it, especially as it updates very infrequently.
This fixes `hack/update-generated-protobuf.sh` as well.
cc @jayunit100 @soltysh
Automatic merge from submit-queue
build: fixed ${KUBE_ROOT} prefix for build scripts
Running `./make-build-image.sh` command inside the `build` directory doesn't work:
```sh
$ cd build
$ ./make-build-image.sh
./../build/common.sh: line 32: hack/lib/init.sh: No such file or directory
```
This PR adds `${KUBE_ROOT}` prefix for the `source` bash function. Also I added braces to unify the code style.
[]()
If the docker-machine certificates get in a bad state, the current behavior
causes an infinite loop waiting for `docker-machine env` to return. Now it will
echo the certificate error and prompt the user to regenerate.
This logs a false "error" message, so it's time to go. It was needed to ensure
nobody has stale build images laying around, but that was quite a while ago, so
it's probably safe now.
Automatic merge from submit-queue
Substitute federation_domain_map parameter with its value in node bootstrap scripts.
This PR also removes the substitution code we added to the build scripts.
**Release Note**
```release-note
If you use one of the kube-dns replication controller manifest in `cluster/saltbase/salt/kube-dns`, i.e. `cluster/saltbase/salt/kube-dns/{skydns-rc.yaml.base,skydns-rc.yaml.in}`, either substitute one of `__PILLAR__FEDERATIONS__DOMAIN__MAP__` or `{{ pillar['federations_domain_map'] }}` with the corresponding federation name to domain name value or remove them if you do not support cluster federation at this time. If you plan to substitute the parameter with its value, here is an example for `{{ pillar['federations_domain_map'] }`
pillar['federations_domain_map'] = "- --federations=myfederation=federation.test"
where `myfederation` is the name of the federation and `federation.test` is the domain name registered for the federation.
```
cc @erictune @kubernetes/sig-cluster-federation @MikeSpreitzer @luxas
[]()
Automatic merge from submit-queue
Add upgrade Docker VM
Add an Error Message to upgarde your Docker VM if needed, example output:
```bash
+++ [0622 13:19:48] No docker host is set. Checking options for setting one...
+++ [0622 13:19:49] docker-machine was found.
+++ [0622 13:19:49] A Docker host using docker-machine named 'kube-dev' is ready to go!
Can't connect to 'docker' daemon. please fix and retry.
Possible causes:
- On Mac OS X, DOCKER_HOST hasn't been set. You may need to:
- Create and start your VM using docker-machine or boot2docker:
- docker-machine create -d virtualbox --virtualbox-memory 4096 --virtualbox-cpu-count -1 kube-dev
- boot2docker init && boot2docker start
- Set your environment variables using:
- eval $(docker-machine env kube-dev)
- $(boot2docker shellinit)
- On Linux, user isn't in 'docker' group. Add and relogin.
- Something like 'sudo usermod -a -G docker jscheuermann'
- RHEL7 bug and workaround: https://bugzilla.redhat.com/show_bug.cgi?id=1119282#c8
- On Linux, Docker daemon hasn't been started or has crashed.
!!! Error in hack/../hack/update-generated-protobuf.sh:53
'return 1' exited with status 1
Call stack:
1: hack/../hack/update-generated-protobuf.sh:53 main(...)
Exiting with status 1
Updating generated-protobuf FAILED
$docker info
Error response from daemon: client is newer than server (client API version: 1.24, server API version: 1.23)
```
After running `docker-machine upgrade kube-dev` everything is fine again. So we should add a hint in the error message that this can also happen.
Automatic merge from submit-queue
Enable all ppc64le builds, except for hyperkube
Partially fixes: #25886
Talked to @Pensu, and all other binaries seem to work fine
@david-mcmahon @ixdy @Pensu @smarterclayton
Auto generated docs are **NO LONGER CHECKED IN**, only placeholders.
To generate them, e.g. before exporting docs, run hack/generate-docs.sh.
hack/verify-generated-docs.sh ensures that generated docs are merely the
placeholder text.
hack/update-generated-docs.sh puts the placeholder text in the proper
places.
The old munge behavior is moved into hack/{update|verify}-munge-docs.sh.
Automatic merge from submit-queue
Switch DNS addons from skydns to kubedns
Change GCI and trusty cluster-helper scripts to use kubedns instead of skydns.
Unified skydns templates using a simple underscore based template and
added transform sed scripts to transform into salt and sed yaml
templates
Moved all content out of cluster/addons/dns into build/kube-dns and
saltbase/salt/kube-dns