Automatic merge from submit-queue (batch tested with PRs 65348, 65599, 65635, 65688, 65691). 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>.
test/e2e_node/system/types_unix: support ZFS
**What this PR does / why we need it**:
Docker validation tests in the case of ZFS used as the graph driver
fail due to "zfs" not being present in the default Docker specification.
Add "zfs" in the GraphDriver slice.
kubeadm relies on the `DockerValidator` and pre-flight checks would fail if the user is using ZFS.
**Which issue(s) this PR fixes** *(optional, in `fixes #<issue number>(, fixes #<issue_number>, ...)` format, will close the issue(s) when PR gets merged)*:
Updates kubernetes/kubeadm#930
**Special notes for your reviewer**:
NONE
/cc @kubernetes/sig-node-pr-reviews
/cc @kubernetes/sig-cluster-lifecycle-pr-reviews
/cc @kvaps (reported by)
/area node-e2e
/area kubeadm
**Release note**:
```release-note
Unix: support ZFS as a valid graph driver for Docker
```
Docker validation tests in the case of ZFS used as the graph driver
fail due to "zfs" not being present in the default Docker specification.
Add "zfs" in the GraphDriver slice.
Kubelet restart process seems to get a bit slower recently. From running
the gpu_device_plugin e2e_node test on GCE, I saw it took ~37 seconds
for kubelet to start CM DeviceManager after it restarts, and then took
~12 seconds for the gpu device plugin to re-register. As the result,
this e2e_node test fails because the current 10 sec waiting time is too
small. Restarting a container also seems to get slower that it sometimes
exceeds the current 2 min waiting time in ensurePodContainerRestart().
This change increase both waiting time to 5 min to leave enough space
on slower machines.
Automatic merge from submit-queue (batch tested with PRs 65449, 65373, 49410). 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>.
add kernel config locations for fedora and atomic
**What this PR does / why we need it**:
* Fedora stores its kernel configs in /usr/lib/modules/$(uname -r)/config
* Fedora/CentOS/RHEL atomic hosts use /usr/lib/ostree-boot/$(uname -r), though this location is deprecated
* The lack of these locations in the validator is causing kubeadm to hang on "failed to parse kernel config" in its preflight checking on fedora and atomic host
**Special notes for your reviewer**:
**Release note**:
```release-note
```
Automatic merge from submit-queue. 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>.
Add metadata to kubelet eviction event annotations
**What this PR does / why we need it**:
Add annotations to kubelet eviction events. Annotations include
"offending_containers" : comma-seperated list of containers.
"offending_containers_usage": comma-seperated list of usage.
"starved_resource": v1.ResourceName of the starved resource
**Special notes for your reviewer**:
Adding annotations to events required changing the `EventRecorder` interface to add a `AnnotatedEventf` function, which can add annotations to an event.
**Release note**:
```release-note
NONE
```
/assign @dchen1107
cc @mwielgus @schylek @kgrygiel
Automatic merge from submit-queue (batch tested with PRs 64456, 64457, 64472). 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>.
e2e node: mark pod cgroup test as [NodeConformance]
**What this PR does / why we need it**:
**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
```
Automatic merge from submit-queue (batch tested with PRs 64456, 64457, 64472). 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>.
Fix dynamic kubelet config tests
**What this PR does / why we need it**:
The default for dynamic kubelet config is now true, so if it is unset, assume it is enabled for testing.
**Release note**:
```release-note
NONE
```
/sig node
/kind bug
/priority critical-urgent
/assign @mtaufen
Automatic merge from submit-queue (batch tested with PRs 64175, 63893). 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>.
add colon separators to improve readability of test names
```release-note
NONE
```
Automatic merge from submit-queue (batch tested with PRs 61963, 64279, 64130, 64125, 64049). 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>.
Add node-exclusive tags to tests in test/e2e_node
Original issue: #59001
Depends on: #64128
Add node-exclusive tags based on the [proposal](https://docs.google.com/document/d/1BdNVUGtYO6NDx10x_fueRh_DLT-SVdlPC_SsXjYCHOE/edit#)
Follow-up PRs will:
- Tag the tests in `test/e2e/common`
- Change the test job configurations to use the new tests
- Remove the unused, non-node-exclusive tags in `test/e2e_node`
**Release note**:
```release-note
NONE
```
Automatic merge from submit-queue (batch tested with PRs 64013, 63896, 64139, 57527, 62102). 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>.
add dynamic config metrics
This PR exports config-releated metrics from the Kubelet.
The Guages for active, assigned, and last-known-good config can be used
to identify config versions and produce aggregate counts across several
nodes. The error-reporting Gauge can be used to determine whether a node
is experiencing a config-related error, and to prodouce an aggregate
count of nodes in an error state.
https://github.com/kubernetes/features/issues/281
```release-note
The Kubelet now exports metrics that report the assigned (node_config_assigned), last-known-good (node_config_last_known_good), and active (node_config_active) config sources, and a metric indicating whether the node is experiencing a config-related error (node_config_error). The config source metrics always report the value 1, and carry the node_config_name, node_config_uid, node_config_resource_version, and node_config_kubelet_key labels, which identify the config version. The error metric reports 1 if there is an error, 0 otherwise.
```
Automatic merge from submit-queue (batch tested with PRs 64013, 63896, 64139, 57527, 62102). 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>.
Refactor test utils that deal with Kubelet metrics for clarity
I found these functions hard to understand, because the names did not
accurately reflect their behavior. For example, GetKubeletMetrics
assumed that all of the metrics passed in were measuring latency.
The caller of GetKubeletMetrics was implicitly making this assumption,
but it was not obvious at the call site.
```release-note
NONE
```
This PR exports config-releated metrics from the Kubelet.
The Guages for active, assigned, and last-known-good config can be used
to identify config versions and produce aggregate counts across several
nodes. The error-reporting Gauge can be used to determine whether a node
is experiencing a config-related error, and to prodouce an aggregate
count of nodes in an error state.
Automatic merge from submit-queue (batch tested with PRs 63881, 64046, 63409, 63402, 63221). 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>.
Kubelet responds to ConfigMap mutations for dynamic Kubelet config
This PR makes dynamic Kubelet config easier to reason about by leaving less room for silent skew scenarios. The new behavior is as follows:
- ConfigMap does not exist: Kubelet reports error status due to missing source
- ConfigMap is created: Kubelet starts using it
- ConfigMap is updated: Kubelet respects the update (but we discourage this pattern, in favor of incrementally migrating to a new ConfigMap)
- ConfigMap is deleted: Kubelet keeps using the config (non-disruptive), but reports error status due to missing source
- ConfigMap is recreated: Kubelet respects any updates (but, again, we discourage this pattern)
This PR also makes a small change to the config checkpoint file tree structure, because ResourceVersion is now taken into account when saving checkpoints. The new structure is as follows:
```
- dir named by --dynamic-config-dir (root for managing dynamic config)
| - meta
| - assigned (encoded kubeletconfig/v1beta1.SerializedNodeConfigSource object, indicating the assigned config)
| - last-known-good (encoded kubeletconfig/v1beta1.SerializedNodeConfigSource object, indicating the last-known-good config)
| - checkpoints
| - uid1 (dir for versions of object identified by uid1)
| - resourceVersion1 (dir for unpacked files from resourceVersion1)
| - ...
| - ...
```
fixes: #61643
```release-note
The dynamic Kubelet config feature will now update config in the event of a ConfigMap mutation, which reduces the chance for silent config skew. Only name, namespace, and kubeletConfigKey may now be set in Node.Spec.ConfigSource.ConfigMap. The least disruptive pattern for config management is still to create a new ConfigMap and incrementally roll out a new Node.Spec.ConfigSource.
```
I found these functions hard to understand, because the names did not
accurately reflect their behavior. For example, GetKubeletMetrics
assumed that all of the metrics passed in were measuring latency.
The caller of GetKubeletMetrics was implicitly making this assumption,
but it was not obvious at the call site.
Automatic merge from submit-queue (batch tested with PRs 63865, 57849, 63932, 63930, 63936). 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>.
Eviction Node e2e test checks for eviction reason
**What this PR does / why we need it**:
Currently, the eviction test simply ensures that pods are marked `Failed`. However, this could occur because of an OOM, rather than an eviction.
To ensure that pods are actually being evicted, check for the Reason in the pod status to ensure it is evicted.
**Release note**:
```release-note
NONE
```
cc @kubernetes/sig-node-pr-reviews
Updates dynamic Kubelet config to use a structured status, rather than a
node condition. This makes the status machine-readable, and thus more
useful for config orchestration.
Fixes: #56896
Automatic merge from submit-queue (batch tested with PRs 63624, 59847). 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>.
explicit kubelet config key in Node.Spec.ConfigSource.ConfigMap
This makes the Kubelet config key in the ConfigMap an explicit part of
the API, so we can stop using magic key names.
As part of this change, we are retiring ConfigMapRef for ConfigMap.
```release-note
You must now specify Node.Spec.ConfigSource.ConfigMap.KubeletConfigKey when using dynamic Kubelet config to tell the Kubelet which key of the ConfigMap identifies its config file.
```
The names were made invalid for the CgroupName refactor in #62541, so
update them here.
Furthermore, as the new names are now compatible with what
EnforceNodeAllocatable wants, reuse the constants there as well.
Tested:
$ make test-e2e-node REMOTE=true HOSTS=test-cos-beta-67-10575-27-0 FOCUS='Validate Node Allocatable' SKIP='' TEST_ARGS='--feature-gates=DynamicKubeletConfig=true'
• [SLOW TEST:39.488 seconds]
[k8s.io] Node Container Manager [Serial]
Validate Node Allocatable
set's up the node and runs the test
Ran 1 of 261 Specs in 57.348 seconds
SUCCESS! -- 1 Passed | 0 Failed | 0 Pending | 260 Skipped
This makes the Kubelet config key in the ConfigMap an explicit part of
the API, so we can stop using magic key names.
As part of this change, we are retiring ConfigMapRef for ConfigMap.
Automatic merge from submit-queue (batch tested with PRs 61455, 63346, 63130, 63404). 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>.
[Device-Plugin]: Extend e2e test to cover node allocatables
**What this PR does / why we need it**:
Extends device plugin e2e to cover node allocatable
**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
```
/sig node
/area hw-accelerators
/cc @jiayingz @vishh @RenaudWasTaken
Automatic merge from submit-queue. 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>.
Use a []string for CgroupName, which is a more accurate internal representation
**What this PR does / why we need it**:
This is purely a refactoring and should bring no essential change in behavior.
It does clarify the cgroup handling code quite a bit.
It is preparation for further changes we might want to do in the cgroup hierarchy. (But it's useful on its own, so even if we don't do any, it should still be considered.)
**Special notes for your reviewer**:
The slice of strings more precisely captures the hierarchic nature of the cgroup paths we use to represent pods and their groupings.
It also ensures we're reducing the chances of passing an incorrect path format to a cgroup driver that requires a different path naming, since now explicit conversions are always needed.
The new constructor `NewCgroupName` starts from an existing `CgroupName`, which enforces a hierarchy where a root is always needed. It also performs checking on the component names to ensure invalid characters ("/" and "_") are not in use.
A `RootCgroupName` for the top of the cgroup hierarchy tree is introduced.
This refactor results in a net reduction of around 30 lines of code,
mainly with the demise of ConvertCgroupNameToSystemd which had fairly
complicated logic in it and was doing just too many things.
There's a small TODO in a helper `updateSystemdCgroupInfo` that was introduced to make this commit possible. That logic really belongs in libcontainer, I'm planning to send a PR there to include it there. (The API already takes a field with that information, only that field is only processed in cgroupfs and not systemd driver, we should fix that.)
Tested: By running the e2e-node tests on both Ubuntu 16.04 (with cgroupfs driver) and CentOS 7 (with systemd driver.)
**NOTE**: I only tested this with dockershim, we should double-check that this works with the CRI endpoints too, both in cgroupfs and systemd modes.
/assign @derekwaynecarr
/assign @dashpole
/assign @Random-Liu
**Release note**:
```release-note
NONE
```
Automatic merge from submit-queue. 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 all script shebangs to use /usr/bin/env interpreter instead of /bin/interpreter
This is required to support systems where bash doesn't reside in /bin (such as NixOS, or the *BSD family) and allow users to specify a different interpreter version through $PATH manipulation.
https://www.cyberciti.biz/tips/finding-bash-perl-python-portably-using-env.html
```release-note
Use /usr/bin/env in all script shebangs to increase portability.
```
The slice of strings more precisely captures the hierarchic nature of
the cgroup paths we use to represent pods and their groupings.
It also ensures we're reducing the chances of passing an incorrect path
format to a cgroup driver that requires a different path naming, since
now explicit conversions are always needed.
The new constructor NewCgroupName starts from an existing CgroupName,
which enforces a hierarchy where a root is always needed. It also
performs checking on the component names to ensure invalid characters
("/" and "_") are not in use.
A RootCgroupName for the top of the cgroup hierarchy tree is introduced.
This refactor results in a net reduction of around 30 lines of code,
mainly with the demise of ConvertCgroupNameToSystemd which had fairly
complicated logic in it and was doing just too many things.
There's a small TODO in a helper updateSystemdCgroupInfo that was
introduced to make this commit possible. That logic really belongs in
libcontainer, I'm planning to send a PR there to include it there.
(The API already takes a field with that information, only that field is
only processed in cgroupfs and not systemd driver, we should fix that.)
Tested by running the e2e-node tests on both Ubuntu 16.04 (with cgroupfs
driver) and CentOS 7 (with systemd driver.)
Automatic merge from submit-queue (batch tested with PRs 58474, 60034, 62101, 63198). 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>.
Fix wrong usage of kubelet option
**What this PR does / why we need it**:
"--allow-privileged true" is incorrect usage of boolean option.
It means setting '--allow-priviledged' to its default value plus
non-existing subcommand 'true'.
"--allow-privileged false" is even more confusing as it sets
allow-priviledged flag to its default value 'true'
This is true for any boolean command line option.
Fixed this by using correct syntax --allow-priviledged=true
**Special notes for your reviewer**:
This is a show-stopper for PR #61833
**Release note**:
```release-note
NONE
```
Automatic merge from submit-queue. 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>.
E2e path utils
**What this PR does / why we need it**:
A bunch of useful methods for getting k8s paths and stuff are secreted away in `e2e_node`. This PR pulls them out so they can be used in other E2E method.
**Which issue(s) this PR fixes** *(optional, in `fixes #<issue number>(, fixes #<issue_number>, ...)` format, will close the issue(s) when PR gets merged)*:
**Special notes for your reviewer**:
This is motivated by the upcoming kubeadm-specific E2E tests. Those tests will be added in a follow-up to this PR.
**Release note**:
```release-note
NONE
```
https://github.com/kubernetes/kubernetes/pull/62913 switched from using a client pool, where each groupVersionResource got its own rest client, to a single client.
This increases the QPS to account for increased requests using a single rest client rate limiter.
Automatic merge from submit-queue. 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>.
unpack dynamic kubelet config payloads to files
This PR unpacks the downloaded ConfigMap to a set of files on the node.
This enables other config files to ride alongside the
KubeletConfiguration, and the KubeletConfiguration to refer to these
cohabitants with relative paths.
This PR also stops storing dynamic config metadata (e.g. current,
last-known-good config records) in the same directory as config
checkpoints. Instead, it splits the storage into `meta` and
`checkpoints` dirs.
The current store dir structure is as follows:
```
- dir named by --dynamic-config-dir (root for managing dynamic config)
| - meta (dir for metadata, e.g. which config source is currently assigned, last-known-good)
| - current (a serialized v1 NodeConfigSource object, indicating the assigned config)
| - last-known-good (a serialized v1 NodeConfigSource object, indicating the last-known-good config)
| - checkpoints (dir for config checkpoints)
| - uid1 (dir for unpacked config, identified by uid1)
| - file1
| - file2
| - ...
| - uid2
| - ...
```
There are some likely changes to the above structure before dynamic config goes beta, such as renaming "current" to "assigned" for clarity, and extending the checkpoint identifier to include a resource version, as part of resolving #61643.
```release-note
NONE
```
/cc @luxas @smarterclayton
This PR unpacks the downloaded ConfigMap to a set of files on the node.
This enables other config files to ride alongside the
KubeletConfiguration, and the KubeletConfiguration to refer to these
cohabitants with relative paths.
This PR also stops storing dynamic config metadata (e.g. current,
last-known-good config records) in the same directory as config
checkpoints. Instead, it splits the storage into `meta` and
`checkpoints` dirs.
"--allow-privileged true" is incorrect usage of boolean option.
It means setting '--allow-priviledged' to its default value plus
non-existing subcommand 'true'.
"--allow-privileged false" is even more confusing as it sets
allow-priviledged flag to its default value 'true'
This is true for any boolean command line option.
Fixed this by using correct syntax --allow-priviledged=true
Fixed generating of kubelet command line in addKubeletConfigFlags
function.
Automatic merge from submit-queue (batch tested with PRs 62192, 61866, 62206, 62360). 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>.
Remove rkt references in the codebase
```release-note
None
```
Automatic merge from submit-queue. 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>.
Fixes restartKubelet in test/e2e_node failure.
Looks like there is some recent change on how we start kubelet service
in test_e2e_node. Fixes restartKubelet() to get right kubelet service
name to cope with the change.
**What this PR does / why we need it**:
**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 kubelet-serial-gce-e2e test failure:
https://k8s-testgrid.appspot.com/wg-resource-management#kubelet-serial-gce-e2e
Thanks a lot to @mindprince for noticing this!
**Special notes for your reviewer**:
**Release note**:
```release-note
```
Looks like there is some recent change on how we start kubelet service
in test_e2e_node. Fixes restartKubelet() to get right kubelet service
name to cope with the change.
Automatic merge from submit-queue (batch tested with PRs 61498, 62030). 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>.
Delete in-tree support for NVIDIA GPUs.
This removes the alpha Accelerators feature gate which was deprecated in 1.10 (#57384).
The alternative feature DevicePlugins went beta in 1.10 (#60170).
Fixes#54012
```release-note
Support for "alpha.kubernetes.io/nvidia-gpu" resource which was deprecated in 1.10 is removed. Please use the resource exposed by DevicePlugins instead ("nvidia.com/gpu").
```
Automatic merge from submit-queue (batch tested with PRs 61482, 61740). 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>.
Make systemd service name for kubelet use a timestamp in e2e-node tests.
**What this PR does / why we need it**:
This makes it easier to figure out which execution was last when looking at the output of `systemd list-units kubelet-*.service`.
We try to find the name of the /tmp/node-e2e-* directory and use the same timestamp if we can. Otherwise, we just call Now() again, which isn't as nice (as the unit name and directory name will not match) but will still produce unit names that will be ordered when launching multiple subsequent executions on the same host.
**Which issue(s) this PR fixes** *(optional, in `fixes #<issue number>(, fixes #<issue_number>, ...)` format, will close the issue(s) when PR gets merged)*:
N/A
**Special notes for your reviewer**:
Tested using `make test-e2e-node REMOTE=true` and then checking `systemctl list-units kubelet-*.service` on the target host.
```
$ systemctl list-units kubelet-*.service
kubelet-20180326T142016.service loaded active exited /tmp/node-e2e-20180326T142016/kubelet --kubeconfig /tmp/node-e2e-20180326T142016/kubeconfig --root-dir /var/lib/kubelet ...
kubelet-20180326T143550.service loaded active exited /tmp/node-e2e-20180326T143550/kubelet --kubeconfig /tmp/node-e2e-20180326T143550/kubeconfig --root-dir /var/lib/kubelet ...
```
The units are sorted in the order they were launched.
**Release note**:
```release-note
NONE
```
The numbers will only be available when docker.service has its own
memory and cpu cgroups, which doesn't necessarily happen unless the unit
has Delegate=yes configured.
Let's work around that by checking the status of Delegate, in the case
where we are:
* running Docker
* running Systemd
* able to check the status through systemctl
* the status is explicitly Delegate=no (the default)
If all of those are true, let's make CPU and Memory expectations
optional.
Tested: make test-e2e-node REMOTE=true HOSTS=centos-e2e-node FOCUS="Summary API"
This is necessary to show any RootFs usage on systems where the backing
filesystem of overlay2 is xfs.
The current test only created directories (for mount points) in the
upper layer of the overlay. Outside of the mount namespace, only the
directories are visible. When running `du` on those, usually filesystems
will show some usage, but not xfs, which shows a disk usage of 0 for
directories.
Fix this by creating a file in the root directory, outside the volumes,
in order to trigger some disk usage that can be measured by `du`.
Tested: make test-e2e-node REMOTE=true HOSTS=centos-e2e-node FOCUS="Summary API"
Automatic merge from submit-queue (batch tested with PRs 61829, 61908, 61307, 61872, 60100). 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>.
node authorizer sets up access rules for dynamic config
This PR makes the node authorizer automatically set up access rules for
dynamic Kubelet config.
I also added some validation to the node strategy, which I discovered we
were missing while writing this.
This PR is based on another WIP from @liggitt.
```release-note
The node authorizer now automatically sets up rules for Node.Spec.ConfigSource when the DynamicKubeletConfig feature gate is enabled.
```
This makes it easier to figure out which execution was last when looking
at the output of `systemd list-units kubelet-*.service`.
We try to find the name of the /tmp/node-e2e-* directory and use the
same timestamp if we can. Otherwise, we just call Now() again, which
isn't as nice (as the unit name and directory name will not match) but
will still produce unit names that will be ordered when launching
multiple subsequent executions on the same host.
Curl is more ubiquitous than wget. For instance, the GCE centos-7 and
rhel-7 image families ship curl by default, but not wget.
Looking at the shell scripts under cluster/, they tend to use curl more
than wget. (The ones that use wget, such as get-kube.sh, try curl first
and only fallback to wget if it's not available.)
Tested: by running node-e2e-test on Ubuntu, COS and CentOS.
This PR makes the node authorizer automatically set up access rules for
dynamic Kubelet config.
I also added some validation to the node strategy, which I discovered we
were missing while writing this.
Automatic merge from submit-queue. 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>.
Replace package "golang.org/x/net/context" with "context"
**What this PR does / why we need it**:
Replace package "golang.org/x/net/context" with "context"
**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#60560
**Special notes for your reviewer**:
As of Go 1.7 this package(golang.org/x/net/context) is available in the standard library under the name context. see (https://godoc.org/golang.org/x/net/context)
It is almost machinery replace.
**Release note**:
```release-note
NONE
```
Automatic merge from submit-queue (batch tested with PRs 61378, 60915, 61499, 61507, 61478). 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>.
Capture pod startup phases as metrics
Learning from https://github.com/kubernetes/kubernetes/issues/60589, we should also start collecting and graphing sub-parts of pod-startup latency.
/sig scalability
/kind feature
/priority important-soon
/cc @wojtek-t
```release-note
NONE
```
Automatic merge from submit-queue (batch tested with PRs 59740, 59728, 60080, 60086, 58714). 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>.
more concise to merge the slice
**What this PR does / why we need it**:
more concise to merge the slice
**Special notes for your reviewer**:
Automatic merge from submit-queue. 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>.
Fixes the races around devicemanager Allocate() and endpoint deletion.
There is a race in predicateAdmitHandler Admit() that getNodeAnyWayFunc()
could get Node with non-zero deviceplugin resource allocatable for a
non-existing endpoint. That race can happen when a device plugin fails,
but is more likely when kubelet restarts as with the current registration
model, there is a time gap between kubelet restart and device plugin
re-registration. During this time window, even though devicemanager could
have removed the resource initially during GetCapacity() call, Kubelet
may overwrite the device plugin resource capacity/allocatable with the
old value when node update from the API server comes in later. This
could cause a pod to be started without proper device runtime config set.
To solve this problem, introduce endpointStopGracePeriod. When a device
plugin fails, don't immediately remove the endpoint but set stopTime in
its endpoint. During kubelet restart, create endpoints with stopTime set
for any checkpointed registered resource. The endpoint is considered to be
in stopGracePeriod if its stoptime is set. This allows us to track what
resources should be handled by devicemanager during the time gap.
When an endpoint's stopGracePeriod expires, we remove the endpoint and
its resource. This allows the resource to be exported through other channels
(e.g., by directly updating node status through API server) if there is such
use case. Currently endpointStopGracePeriod is set as 5 minutes.
Given that an endpoint is no longer immediately removed upon disconnection,
mark all its devices unhealthy so that we can signal the resource allocatable
change to the scheduler to avoid scheduling more pods to the node.
When a device plugin endpoint is in stopGracePeriod, pods requesting the
corresponding resource will fail admission handler.
Tested:
Ran GPUDevicePlugin e2e_node test 100 times and all passed now.
**What this PR does / why we need it**:
**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 https://github.com/kubernetes/kubernetes/issues/60176
**Special notes for your reviewer**:
**Release note**:
```release-note
Fixes the races around devicemanager Allocate() and endpoint deletion.
```
There is a race in predicateAdmitHandler Admit() that getNodeAnyWayFunc()
could get Node with non-zero deviceplugin resource allocatable for a
non-existing endpoint. That race can happen when a device plugin fails,
but is more likely when kubelet restarts as with the current registration
model, there is a time gap between kubelet restart and device plugin
re-registration. During this time window, even though devicemanager could
have removed the resource initially during GetCapacity() call, Kubelet
may overwrite the device plugin resource capacity/allocatable with the
old value when node update from the API server comes in later. This
could cause a pod to be started without proper device runtime config set.
To solve this problem, introduce endpointStopGracePeriod. When a device
plugin fails, don't immediately remove the endpoint but set stopTime in
its endpoint. During kubelet restart, create endpoints with stopTime set
for any checkpointed registered resource. The endpoint is considered to be
in stopGracePeriod if its stoptime is set. This allows us to track what
resources should be handled by devicemanager during the time gap.
When an endpoint's stopGracePeriod expires, we remove the endpoint and
its resource. This allows the resource to be exported through other channels
(e.g., by directly updating node status through API server) if there is such
use case. Currently endpointStopGracePeriod is set as 5 minutes.
Given that an endpoint is no longer immediately removed upon disconnection,
mark all its devices unhealthy so that we can signal the resource allocatable
change to the scheduler to avoid scheduling more pods to the node.
When a device plugin endpoint is in stopGracePeriod, pods requesting the
corresponding resource will fail admission handler.
Automatic merge from submit-queue. 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>.
Add node-e2e test for ShareProcessNamespace
**What this PR does / why we need it**: Adds a node-e2e test for kubernetes/features#495
**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#59554
**Special notes for your reviewer**: This requires a feature gate to be enabled in both the kubelet and API server. I'm not sure which jenkins configs need to be updated (or if these are even still used) so I just updated a pile of them.
opened kubernetes/test-infra#7030 for https://github.com/kubernetes/test-infra/blob/master/jobs/config.json
**Release note**:
```release-note
NONE
```
Automatic merge from submit-queue. 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>.
fix todo:Move function readinessCheck to util
**What this PR does / why we need it**:
fix todo:Move function readinessCheck to util in test/e2e_node/services
**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
```
Automatic merge from submit-queue (batch tested with PRs 60376, 55584, 60358, 54631, 60291). 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>.
remove gcloud docker -- since it's deprecated
docker handles this now and it raises an error.
try 3
```release-note
NONE
```
Automatic merge from submit-queue (batch tested with PRs 60236, 60332, 57375, 60451, 57408). 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 device plugin e2e_node test to not changing Kubelet config
as DevicePlugins feature is enabled by default now.
**What this PR does / why we need it**:
**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
```
Automatic merge from submit-queue (batch tested with PRs 59159, 60318, 60079, 59371, 57415). 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>.
Made a couple API changes to deviceplugin/v1beta1 to avoid future
incompatible API changes:
- Add GetDevicePluginOptions rpc call. This is needed when we switch
from Registration service to probe-based plugin watcher.
- Change AllocateRequest and AllocateResponse to allow device requests
from multiple containers in a pod. Currently only made mechanical
change on the devicemanager and test code to cope with the API but
still issues an Allocate call per container. We can modify the
devicemanager in 1.11 to issue a single Allocate call per pod.
The change will also facilitate incremental API change to communicate
pod level information through Allocate rpc if there is such future
need.
**What this PR does / why we need it**:
Made a couple API changes to deviceplugin/v1beta1 to avoid future incompatible API changes.
**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 https://github.com/kubernetes/kubernetes/issues/59370
**Special notes for your reviewer**:
**Release note**:
```release-note
```
Automatic merge from submit-queue (batch tested with PRs 60324, 60269, 59771, 60314, 59941). 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>.
expunge the word 'manifest' from Kubelet's config API
The word 'manifest' technically refers to a container-group specification
that predated the Pod abstraction. We should avoid using this legacy
terminology where possible. Fortunately, the Kubelet's config API will
be beta in 1.10 for the first time, so we still had the chance to make
this change.
I left the flags alone, since they're deprecated anyway.
I changed a few var names in files I touched too, but this PR is the
just the first shot, not the whole campaign
(`git grep -i manifest | wc -l -> 1248`).
```release-note
Some field names in the Kubelet's now v1beta1 config API differ from the v1alpha1 API: PodManifestPath is renamed to PodPath, ManifestURL is renamed to PodURL, ManifestURLHeader is renamed to PodURLHeader.
```
Automatic merge from submit-queue. 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 test framework featuregates type
**What this PR does / why we need it**:
A cleanup following #53025 and #57962.
**Which issue(s) this PR fixes** *(optional, in `fixes #<issue number>(, fixes #<issue_number>, ...)` format, will close the issue(s) when PR gets merged)*:
ref: #53025
and #57962.
**Special notes for your reviewer**:
but yeah, not sure if it's worthy to do this :)
**Release note**:
```release-note
NONE
```
incompatible changes:
- Add GetDevicePluginOptions rpc call. This is needed when we switch
from Registration service to probe-based plugin watcher.
- Change AllocateRequest and AllocateResponse to allow device requests
from multiple containers in a pod. Currently only made mechanical
change on the devicemanager and test code to cope with the API but
still issues an Allocate call per container. We can modify the
devicemanager in 1.11 to issue a single Allocate call per pod.
The change will also facilitate incremental API change to communicate
pod level information through Allocate rpc if there is such future
need.
The word 'manifest' technically refers to a container-group specification
that predated the Pod abstraction. We should avoid using this legacy
terminology where possible. Fortunately, the Kubelet's config API will
be beta in 1.10 for the first time, so we still had the chance to make
this change.
I left the flags alone, since they're deprecated anyway.
I changed a few var names in files I touched too, but this PR is the
just the first shot, not the whole campaign
(`git grep -i manifest | wc -l -> 1248`).
Automatic merge from submit-queue. 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>.
set default enabled admission plugins by official document
**What this PR does / why we need it**:
https://kubernetes.io/docs/admin/admission-controllers/#is-there-a-recommended-set-of-admission-controllers-to-use
recommend running the following set of admission controllers
```
If you previously had not set the `--admission-control` flag, your cluster behavior may change (to be more standard). See [https://kubernetes.io/docs/admin/admission-controllers/] for explanation of admission control.
```
**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
Set default enabled admission plugins `NamespaceLifecycle,LimitRanger,ServiceAccount,PersistentVolumeLabel,DefaultStorageClass,DefaultTolerationSeconds,MutatingAdmissionWebhook,ValidatingAdmissionWebhook,ResourceQuota`
```
Automatic merge from submit-queue (batch tested with PRs 60158, 60156, 58111, 57583, 60055). 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>.
Bugfix/logf newline
**What this PR does / why we need it**:
Removes all redundant new lines being passed into the `Logf()` function. This involved going through code in both `test/e2e` and `test/e2e_node`, finding the newline redundancies in calls to `Logf()` and removing them.
**Which issue(s) this PR fixes**:
Fixes [#57102](https://github.com/kubernetes/kubernetes/issues/57102)
**Release note**:
```release-note
NONE
```
Automatic merge from submit-queue (batch tested with PRs 58716, 59977, 59316, 59884, 60117). 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>.
remove deprecated /proxy paths
These were deprecated in v1.2.
ref https://github.com/kubernetes/kubernetes/issues/59885
```release-note
kube-apiserver: the root /proxy paths have been removed (deprecated since v1.2). Use the /proxy subresources on objects that support HTTP proxying.
```
@kubernetes/sig-api-machinery-api-reviews
Automatic merge from submit-queue. 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 bazelbuild/rules_go, kubernetes/repo-infra, and gazelle dependencies
**What this PR does / why we need it**: updates our bazelbuild/rules_go dependency in order to bump everything to go1.9.4. I'm separating this effort into two separate PRs, since updating rules_go requires a large cleanup, removing an attribute from most build rules.
**Release note**:
```release-note
NONE
```
Automatic merge from submit-queue (batch tested with PRs 59927, 59989, 59950). 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>.
Fix e2e node setKubeletConfiguration helper
The helper should have been using `apiequality.Semantic.DeepEqual`,
instead of `reflect.DeepEqual`. Previously, nil vs empty containers
were treated as not equal, but they should be considered equal for
objects managed by Kubernetes API machinery, like KubeletConfiguration.
This should fix the failing eviction tests.
```release-note
NONE
```
Automatic merge from submit-queue (batch tested with PRs 59683, 59964, 59841, 59936, 59686). 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>.
Reevaluate eviction thresholds after reclaim functions
**What this PR does / why we need it**:
When the node comes under `DiskPressure` due to inodes or disk space, the eviction manager runs garbage collection functions to clean up dead containers and unused images.
Currently, we use the strategy of trying to measure the disk space and inodes freed by garbage collection. However, as #46789 and #56573 point out, there are gaps in the implementation that can cause extra evictions even when they are not required. Furthermore, for nodes which frequently cycle through images, it results in a large number of evictions, as running out of inodes always causes an eviction.
This PR changes this strategy to call the garbage collection functions and ignore the results. Then, it triggers another collection of node-level metrics, and sees if the node is still under DiskPressure.
This way, we can simply observe the decrease in disk or inode usage, rather than trying to measure how much is freed.
**Which issue(s) this PR fixes**:
Fixes#46789Fixes#56573
Related PR #56575
**Special notes for your reviewer**:
This will look cleaner after #57802 removes arguments from [makeSignalObservations](https://github.com/kubernetes/kubernetes/pull/57802/files#diff-9e5246d8c78d50ce4ba440f98663f3e9R719).
**Release note**:
```release-note
NONE
```
/sig node
/kind bug
/priority important-soon
cc @kubernetes/sig-node-pr-reviews
The helper should have been using `apiequality.Semantic.DeepEqual`,
instead of `reflect.DeepEqual`. Previously, nil vs empty containers
were treated as not equal, but they should be considered equal for
objects managed by Kubernetes API machinery, like KubeletConfiguration.
This should fix the failing eviction tests.
Automatic merge from submit-queue (batch tested with PRs 57136, 59920). 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>.
Updated PID pressure node condition.
**Which issue(s) this PR fixes** *(optional, in `fixes #<issue number>(, fixes #<issue_number>, ...)` format, will close the issue(s) when PR gets merged)*:
part of #54313
**Release note**:
```release-note
Updated PID pressure node condition
```
Automatic merge from submit-queue (batch tested with PRs 59353, 59905, 53833). 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>.
Graduate kubeletconfig API group to beta
Regarding https://github.com/kubernetes/features/issues/281, this PR moves the kubeletconfig API group to beta.
After #53088, the KubeletConfiguration type should not contain any deprecated or experimental fields, and we should not have to remove any more fields from the type before graduating it to beta.
We need the community to double check for two things, however:
1. Are there any fields currently in the KubeletConfiguration type that you were going to mark deprecated this quarter, but haven't yet?
2. Are there any fields currently in the KubeletConfiguration type that are experimental or alpha, but were not explicitly denoted as such?
Please comment on this PR if you can answer "yes" to either of those two questions. Please cc anyone with a stake in the kubeletconfig API, so we get as much coverage as possible.
/cc @thockin @dchen1107 @Random-Liu @yujuhong @dashpole @tallclair @vishh @abw @freehan @dnardo @bowei @MrHohn @luxas @liggitt @ncdc @derekwaynecarr @mikedanese
@kubernetes/sig-network-pr-reviews, @kubernetes/sig-node-pr-reviews
```release-note
action required: The `kubeletconfig` API group has graduated from alpha to beta, and the name has changed to `kubelet.config.k8s.io`. Please use `kubelet.config.k8s.io/v1beta1`, as `kubeletconfig/v1alpha1` is no longer available.
```
**TODO:**
- [x] Move experimental/non-gated-alpha/soon-to-be-deprecated fields to `KubeletFlags`
- [x] #53088
- [x] #54154
- [x] #54160
- [x] #55562
- [x] #55983
- [x] #57851
- [x] Lift embedded structure out of strings
- [x] #53025
- [x] #54643
- [x] #54823
- [x] #55254
- [x] Resolve relative paths against the location config files are loaded from
- [x] #55648
- [x] Rename to `kubelet.config.k8s.io`
- [x] Comments
- [x] Make sure existing comments at least read sensibly.
- [x] Note default values in comments on the versioned struct.
- [x] Remove any reference to default values in comments on the internal struct.
- [x] Most fields should be `+optional` and `omitempty`. Add where necessary. ~Where omitted, explicitly comment.~ Edit: We should not distinguish between nil and empty, see below items.
- [x] Ensure defaults are specified via `pkg/kubelet/apis/kubelet.config.k8s.io/v1beta1/defaults.go`, not `cmd/kubelet/app/options/options.go`.
- [x] #57770
- [x] Ensure kubeadm does not persist v1alpha1 KubeletConfiguration objects (or feature-gates this functionality)
- [x] Don't make a distinction between empty and nil, because of #43203.
- [x] #59515
- [x] #59681
- [x] Take the opportunity to fix insecure Kubelet defaults @tallclair
- [x] #59666
- [x] Remove CAdvisorPort from KubeletConfiguration wrt #56523.
- [x] #59580
- [x] Hide `ConfigTrialDuration` until we're more sure what to do with it.
- [x] #59628
- [x] Fix `// default: x` comments after rebasing on recent changes.
This is a more accurate name for the condition, as it describes the
status of the Kubelet's configuration.
Also cleans up capitalization of internal names.
Automatic merge from submit-queue (batch tested with PRs 59653, 58812, 59582, 59665, 59511). 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>.
Add node e2e tests to verify HugePages feature
**What this PR does / why we need it**:
Add node e2e tests to verify HugePages feature.
**Special notes for your reviewer**:
Test follows same pattern as pod container manager tests.
**Release note**:
```release-note
NONE
```
Automatic merge from submit-queue. 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>.
Ignore 0% and 100% eviction thresholds
Primarily, this gives a way to explicitly disable eviction, which is
necessary to use omitempty on EvictionHard.
See: https://github.com/kubernetes/kubernetes/pull/53833#discussion_r166672137
As justification for this approach, neither 0% nor 100% make sense as
eviction thresholds; in the "less-than" case, you can't have less than
0% of a resource and 100% perpetually evicts; in the
"greater-than" case (assuming we ever add a resource with this
semantic), the reasoning is the reverse (not more than 100%, 0%
perpetually evicts).
```release-note
Eviction thresholds set to 0% or 100% are now ignored.
```
Primarily, this gives a way to explicitly disable eviction, which is
necessary to use omitempty on EvictionHard.
See: https://github.com/kubernetes/kubernetes/pull/53833#discussion_r166672137
As justification for this approach, neither 0% nor 100% make sense as
eviction thresholds; in the "less-than" case, you can't have less than
0% of a resource and 100% perpetually evicts; in the
"greater-than" case (assuming we ever add a resource with this
semantic), the reasoning is the reverse (not more than 100%, 0%
perpetually evicts).
Automatic merge from submit-queue. 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>.
fix all the typos across the project
**What this PR does / why we need it**:
There are lots of typos across the project. We should avoid small PRs on fixing those annoying typos, which is time-consuming and low efficient.
This PR does fix all the typos across the project currently. And with #59463, typos could be avoided when a new PR gets merged.
**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**:
/sig testing
/area test-infra
/sig release
/cc @ixdy
/assign @fejta
**Release note**:
```release-note
None
```
Automatic merge from submit-queue (batch tested with PRs 57824, 58806, 59410, 59280). 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>.
2nd try at using a vanity GCR name
The 2nd commit here is the changes relative to the reverted PR. Please focus review attention on that.
This is the 2nd attempt. The previous try (#57573) 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.
xref https://github.com/kubernetes/release/issues/281
TL;DR:
* The new `staging-k8s.gcr.io` is where we push images. It is literally an alias to `gcr.io/google_containers` (the existing repo) and is hosted in the US.
* The contents of `staging-k8s.gcr.io` are automatically synced to `{asia,eu,us)-k8s.gcr.io`.
* The new `k8s.gcr.io` will be a read-only alias to whichever regional repo is closest to you.
* In the future, images will be promoted from `staging` to regional "prod" more explicitly and auditably.
```release-note
Use "k8s.gcr.io" for pulling container images rather than "gcr.io/google_containers". Images are already synced, so this should not impact anyone materially.
Documentation and tools should all convert to the new name. Users should take note of this in case they see this new name in the system.
```
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.
Automatic merge from submit-queue (batch tested with PRs 59276, 51042, 58973, 59377, 59472). 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 Container Runtime Interface to use enumerated namespace modes
**What this PR does / why we need it**: This updates the CRI as described in the [Shared PID Namespace](https://github.com/kubernetes/community/blob/master/contributors/design-proposals/node/pod-pid-namespace.md#container-runtime-interface-changes) proposal. This change to the alpha API is not backwards compatible: implementations of the CRI will need to update to the new API version.
**Which issue(s) this PR fixes** *(optional, in `fixes #<issue number>(, fixes #<issue_number>, ...)` format, will close the issue(s) when PR gets merged)*:
WIP #1615
**Special notes for your reviewer**:
/assign @yujuhong
**Release note**:
```release-note
[action-required] The Container Runtime Interface (CRI) version has increased from v1alpha1 to v1alpha2. Runtimes implementing the CRI will need to update to the new version, which configures container namespaces using an enumeration rather than booleans.
```
This also incorporates the version string into the package name so
that incompatibile versions will fail to connect.
Arbitrary choices:
- The proto3 package name is runtime.v1alpha2. The proto compiler
normally translates this to a go package of "runtime_v1alpha2", but
I renamed it to "v1alpha2" for consistency with existing packages.
- kubelet/apis/cri is used as "internalapi". I left it alone and put the
public "runtimeapi" in kubelet/apis/cri/runtime.
Automatic merge from submit-queue. 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>.
reopen#58913 Fix TODO move GetPauseImageNameForHostArch func
**What this PR does / why we need it**:
reopen#58913 Fix TODO move GetPauseImageNameForHostArch func,because of I squash to a single commit wrong,so recommit one,and close the #58913
**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**:
/assign @liggitt
**Release note**:
```release-note
NONE
```
Automatic merge from submit-queue (batch tested with PRs 57683, 59116, 58728, 59140, 58976). 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>.
Use node-e2e framework for testing cadvisor
**What this PR does / why we need it**:
With cadvisor checked out in your gopath, we can now run cadvisor integration tests: `make test-e2e-node TEST_SUITE=cadvisor`.
This has a number of advantages:
* we can use the same images to test both, configured the same way.
* we will now get cadvisor logs from the integration test.
* we can now use the familiar node-e2e arguments to specify images to test with cadvisor
* no more managing snowflake VMs for cadvisor.
**Special notes for your reviewer**:
cadvisor doesnt currently produce junit* files, so I removed that as a requirement.
This wont actually work until https://github.com/google/cadvisor/pull/1868 is merged as well.
Related issue:
https://github.com/kubernetes/test-infra/issues/190
**Release note**:
```release-note
NONE
```
/assign @Random-Liu
/sig node
/priority important-soon
/kind cleanup
Automatic merge from submit-queue. 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>.
Fix PodPidsLimit and ConfigTrialDuration on internal KubeletConfig type
They should both follow the convention of not being a pointer on the internal type.
This required adding a conversion function between `int64` and `*int64`. A side effect is this removes a warning in the generated code for the apps API group.
@dims
```release-note
NONE
```
Automatic merge from submit-queue. 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>.
Fixes for HostIPC tests to work when Docker has SELinux support enabled.
**What this PR does / why we need it**:
Fixes for HostIPC tests to work when Docker has SELinux support enabled.
**Which issue(s) this PR fixes** *(optional, in `fixes #<issue number>(, fixes #<issue_number>, ...)` format, will close the issue(s) when PR gets merged)*:
N/A
**Special notes for your reviewer**:
The core of the matter is to use `ipcs` from util-linux rather than the one from busybox. The typical SELinux policy has enough to allow Docker containers (running under svirt_lxc_net_t SELinux type) to access IPC information by reading the contents of the files under /proc/sysvipc/, but not by using the shmctl etc. syscalls.
The `ipcs` implementation in busybox will use `shmctl(0, SHM_INFO, ...)` to detect whether it can read IPC info (see source code [here](https://git.busybox.net/busybox/tree/util-linux/ipcs.c?h=1_28_0#n138)), while the one in util-linux will prefer to read from the /proc files directly if they are available (see source code [here](https://github.com/karelzak/util-linux/blob/v2.27.1/sys-utils/ipcutils.c#L108)).
It turns out the SELinux policy doesn't allow the shmctl syscalls in an unprivileged container, while access to it through the /proc interface is fine. (One could argue this is a bug in the SELinux policy, but getting it fixed on stable OSs is hard, and it's not that hard for us to test it with an util-linux `ipcs`, so I propose we do so.)
This PR also contains a refactor of the code setting IpcMode, since setting it in the "common options" function is misleading, as on containers other than the sandbox, it ends up always getting overwritten, so let's only set it to "host" in the Sandbox.
It also has a minor fix for the `ipcmk` call, since support for size suffix was only introduced in recent versions of it.
**Release note**:
```release-note
NONE
```
They should both follow the convention of not being a pointer on the
internal type. This required adding a conversion function between
`int64` and `*int64`.
A side effect is this removes a warning in the generated code for the
apps API group.
Automatic merge from submit-queue (batch tested with PRs 57467, 58996). 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>.
Set generate-kubelet-config-file to true by default.
This should fix the flaky suite.
https://k8s-testgrid.appspot.com/sig-node-kubelet#kubelet-flaky-gce-e2e
@mtaufen /cc @kubernetes/sig-node-bugs
Signed-off-by: Lantao Liu <lantaol@google.com>
**What this PR does / why we need it**:
**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
```
Automatic merge from submit-queue (batch tested with PRs 57467, 58996). 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>.
Remove flaky label from Eviction tests
**What this PR does / why we need it**:
All eviction tests in the flaky suite are no longer flaky. Remove the flaky label to move them from the flaky suite to the serial suite.
I removed the QoS-based memory eviction test since it does not reflect the current eviction strategy.
**Release note**:
```release-note
NONE
```
/assign @mtaufen @Random-Liu
/sig node
/priority important-soon
/kind cleanup
Automatic merge from submit-queue (batch tested with PRs 58777, 58978, 58977, 58775). 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>.
Skip NoNewPrivileges test when SELinux is enabled
**What this PR does / why we need it**:
A bug in the SELinux policy prevented NoNewPrivileges from working on Docker with SELinux support enabled.
The problem has been fixed upstream (see projectatomic/container-selinux#45)
But hasn't been backported yet (a fix might come in RHEL 7.5)
For now, let's skip the NoNewPrivileges test when SELinux support is enabled in Docker.
Tested:
- Before this commit, the test fails:
```
$ make test-e2e-node REMOTE=true FOCUS="allow privilege escalation"
(on a host with SELinux enabled)
• [SLOW TEST:22.798 seconds] (passed)
[k8s.io] Security Context
when creating containers with AllowPrivilegeEscalation
should allow privilege escalation when true
• Failure [16.539 seconds]
[k8s.io] Security Context
when creating containers with AllowPrivilegeEscalation
should not allow privilege escalation when false [It]
wait for pod "alpine-nnp-false-aef03e47-0090-11e8-886f-42010af00009" to success
Expected success, but got an error:
<*errors.errorString | 0xc4204e26d0>: {
s: "pod \"alpine-nnp-false-aef03e47-0090-11e8-886f-42010af00009\" failed with reason: \"\", message: \"\"",
}
pod "alpine-nnp-false-aef03e47-0090-11e8-886f-42010af00009" failed with reason: "", message: ""
• [SLOW TEST:26.572 seconds] (passed)
[k8s.io] Security Context
when creating containers with AllowPrivilegeEscalation
should allow privilege escalation when not explicitly set and uid != 0
Ran 3 of 257 Specs in 45.364 seconds
FAIL! -- 2 Passed | 1 Failed | 0 Pending | 254 Skipped
Ginkgo ran 1 suite in 49.389123442s
Test Suite Failed
```
- After this commit, the test is skipped:
```
$ make test-e2e-node REMOTE=true FOCUS="allow privilege escalation"
(on a host with SELinux enabled)
S [SKIPPING] in Spec Setup (BeforeEach) [12.452 seconds]
S [SKIPPING] in Spec Setup (BeforeEach) [16.298 seconds]
S [SKIPPING] in Spec Setup (BeforeEach) [18.183 seconds]
Ran 0 of 257 Specs in 39.174 seconds
SUCCESS! -- 0 Passed | 0 Failed | 0 Pending | 257 Skipped
Ginkgo ran 1 suite in 43.570630357s
Test Suite Passed
```
- No changes when SELinux is disabled:
```
$ make test-e2e-node REMOTE=true FOCUS="allow privilege escalation"
(on a host with SELinux disabled)
• [SLOW TEST:15.013 seconds]
[k8s.io] Security Context
when creating containers with AllowPrivilegeEscalation
should not allow privilege escalation when false
• [SLOW TEST:19.155 seconds]
[k8s.io] Security Context
when creating containers with AllowPrivilegeEscalation
should allow privilege escalation when true
• [SLOW TEST:21.087 seconds]
[k8s.io] Security Context
when creating containers with AllowPrivilegeEscalation
should allow privilege escalation when not explicitly set and uid != 0
Ran 3 of 259 Specs in 38.560 seconds
SUCCESS! -- 3 Passed | 0 Failed | 0 Pending | 256 Skipped
Ginkgo ran 1 suite in 41.937918928s
Test Suite Passed
```
**Which issue(s) this PR fixes** *(optional, in `fixes #<issue number>(, fixes #<issue_number>, ...)` format, will close the issue(s) when PR gets merged)*:
N/A
**Special notes for your reviewer**:
N/A
**Release note**:
```release-note
NONE
```
Automatic merge from submit-queue. 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>.
Removal of KubeletConfigFile feature gate: Step 1
This feature gate was redundant with the `--config` flag, which already
enables/disables loading Kubelet config from a file.
Since the gate guarded an alpha feature, removing it is not a violation
of our API guidelines.
Some stuff in `kubernetes/test-infra` currently sets the gate,
so removing will be a 3 step process:
1. This PR, which makes the gate a no-op.
2. Stop setting the gate in `kubernetes/test-infra`.
3. Completely remove the gate (this PR will get the release note).
```release-note
NONE
```
This ensures the `ipcs` command from util-linux will be used, which
succeeds when Docker is running with SELinux enabled (while the one from
busybox fails.)
Tested: On a host with Docker running with SELinux enabled:
$ make test-e2e-node REMOTE=true FOCUS="host IPC"
• [SLOW TEST:17.272 seconds] (passed)
[k8s.io] Security Context
when creating a pod in the host IPC namespace
should show the shared memory ID in the host IPC containers
• [SLOW TEST:20.419 seconds] (passed)
[k8s.io] Security Context
when creating a pod in the host IPC namespace
should not show the shared memory ID in the non-hostIPC containers
Ran 2 of 257 Specs in 43.934 seconds
SUCCESS! -- 2 Passed | 0 Failed | 0 Pending | 255 Skipped
Expand the use of "1M" to the corresponding number of bytes, since
support for size suffix was only added to `ipcmk` in util-linux 2.27
which is not yet available in some Linux distributions.
Tested by running `make test-e2e-node` against distributions with ipcmk
that supports and doesn't support the suffix syntax, all of them passed.
A bug in the SELinux policy prevented NoNewPrivileges from working on
Docker with SELinux support enabled.
The problem has been fixed upstream:
https://github.com/projectatomic/container-selinux/issues/45
But hasn't been backported yet (a fix might come in RHEL 7.5)
For now, let's skip the NoNewPrivileges test when SELinux support is
enabled in Docker.
Tested:
- Before this commit, the test fails:
$ make test-e2e-node REMOTE=true FOCUS="allow privilege escalation"
(on a host with SELinux enabled)
• [SLOW TEST:22.798 seconds] (passed)
[k8s.io] Security Context
when creating containers with AllowPrivilegeEscalation
should allow privilege escalation when true
• Failure [16.539 seconds]
[k8s.io] Security Context
when creating containers with AllowPrivilegeEscalation
should not allow privilege escalation when false [It]
wait for pod "alpine-nnp-false-aef03e47-0090-11e8-886f-42010af00009" to success
Expected success, but got an error:
<*errors.errorString | 0xc4204e26d0>: {
s: "pod \"alpine-nnp-false-aef03e47-0090-11e8-886f-42010af00009\" failed with reason: \"\", message: \"\"",
}
pod "alpine-nnp-false-aef03e47-0090-11e8-886f-42010af00009" failed with reason: "", message: ""
• [SLOW TEST:26.572 seconds] (passed)
[k8s.io] Security Context
when creating containers with AllowPrivilegeEscalation
should allow privilege escalation when not explicitly set and uid != 0
Ran 3 of 257 Specs in 45.364 seconds
FAIL! -- 2 Passed | 1 Failed | 0 Pending | 254 Skipped
Ginkgo ran 1 suite in 49.389123442s
Test Suite Failed
- After this commit, the test is skipped:
$ make test-e2e-node REMOTE=true FOCUS="allow privilege escalation"
(on a host with SELinux enabled)
S [SKIPPING] in Spec Setup (BeforeEach) [12.452 seconds]
S [SKIPPING] in Spec Setup (BeforeEach) [16.298 seconds]
S [SKIPPING] in Spec Setup (BeforeEach) [18.183 seconds]
Ran 0 of 257 Specs in 39.174 seconds
SUCCESS! -- 0 Passed | 0 Failed | 0 Pending | 257 Skipped
Ginkgo ran 1 suite in 43.570630357s
Test Suite Passed
- No changes when SELinux is disabled:
$ make test-e2e-node REMOTE=true FOCUS="allow privilege escalation"
(on a host with SELinux disabled)
• [SLOW TEST:15.013 seconds]
[k8s.io] Security Context
when creating containers with AllowPrivilegeEscalation
should not allow privilege escalation when false
• [SLOW TEST:19.155 seconds]
[k8s.io] Security Context
when creating containers with AllowPrivilegeEscalation
should allow privilege escalation when true
• [SLOW TEST:21.087 seconds]
[k8s.io] Security Context
when creating containers with AllowPrivilegeEscalation
should allow privilege escalation when not explicitly set and uid != 0
Ran 3 of 259 Specs in 38.560 seconds
SUCCESS! -- 3 Passed | 0 Failed | 0 Pending | 256 Skipped
Ginkgo ran 1 suite in 41.937918928s
Test Suite Passed
This feature gate was redundant with the `--config` flag, which already
enables/disables loading Kubelet config from a file.
Since the gate guarded an alpha feature, removing it is not a violation
of our API guidelines.
Some stuff in `kubernetes/test-infra` currently sets the gate,
so removing will be a 3 step process:
1. This PR, which makes the gate a no-op.
2. Stop setting the gate in `kubernetes/test-infra`.
3. Completely remove the gate.
The log path test is not expected to pass unless the Docker is using the
JSON logging driver, since that's what the log path is trying to find.
When Docker is using the journald logging driver, there will be no JSON
files in the logging directories for it to find.
Furthermore, when SELinux support is enabled in the Docker daemon,
SELinux will prevent processes running inside Docker containers from
accessing the log files owned by Docker (which is what this test is
trying to accomplish), so let's also skip this test in case SELinux
support is enabled.
Tested:
- With Docker daemon started using --log-driver=journald:
S [SKIPPING] in Spec Setup (BeforeEach) [8.193 seconds]
[k8s.io] ContainerLogPath
Pod with a container
printed log to stdout
should print log to correct log path [BeforeEach]
Jan 3 18:33:44.869: Skipping because Docker daemon is using a logging driver other than "json-file": journald
- With Docker daemon started using --selinux-enabled:
S [SKIPPING] in Spec Setup (BeforeEach) [8.488 seconds]
[k8s.io] ContainerLogPath
Pod with a container
printed log to stdout
should print log to correct log path [BeforeEach]
Jan 3 18:35:58.909: Skipping because Docker daemon is running with SELinux support enabled
- With Docker started using JSON logging driver and with SELinux disabled:
• [SLOW TEST:16.352 seconds] (passed)
[k8s.io] ContainerLogPath
Pod with a container
printed log to stdout
should print log to correct log path
Ran 1 of 256 Specs in 36.428 seconds
SUCCESS! -- 1 Passed | 0 Failed | 0 Pending | 255 Skipped
Automatic merge from submit-queue. 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>.
Benchmark test non docker specific
**What this PR does / why we need it**:
This will make benchmark test generic to all container runtimes
**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#58321
**Special notes for your reviewer**:
**Release note**:
```release-note
none
```
/cc @Random-Liu
Automatic merge from submit-queue. 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>.
Fix policy conflict in the CPU manager node e2e test.
**What this PR does / why we need it**:
After graduation of the CPU manager feature to Beta, the CPU manager `none` policy is ON by default. But when the CPU manager is set to use `static` policy in the node e2e test, there will always be a conflict with the policy checkpointed in the disk. This PR fixes that by deleting the state file where required.
Manually tested in an `n1-standard-4` instance with `Ubuntu 16.04` image on GCP, which is the same machine and image type as one of the configs used in the node e2e tests.
Use the following command to run the test locally:
`make test-e2e-node TEST_ARGS='--feature-gates=DynamicKubeletConfig=true' FOCUS="CPU Manager" SKIP="" PARALLELISM=1`
CC @ConnorDoyle @derekwaynecarr
Automatic merge from submit-queue. 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>.
Add balajismaniam and ConnorDoyle to node-e2e approvers.
**What this PR does / why we need it**:
- Add balajismaniam and ConnorDoyle to node-e2e approvers.
**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**:
_Rationale:_ We are maintaining node e2e tests for the CPU manager component, and would also like to help with the rest of review load in this package. Both Balaji and I are approvers for the cpumanager and cpuset packages in the Kubelet container manager.
**Release note**:
```release-note
NONE
```
Automatic merge from submit-queue (batch tested with PRs 58216, 58193, 53033, 58219, 55921). 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>.
Use GinkgoRecover to avoid panic.
See this in the test:
```
I0111 14:28:31.010] panic:
I0111 14:28:31.010] Your test failed.
I0111 14:28:31.010] Ginkgo panics to prevent subsequent assertions from running.
I0111 14:28:31.011] Normally Ginkgo rescues this panic so you shouldn't see it.
I0111 14:28:31.011]
I0111 14:28:31.011] But, if you make an assertion in a goroutine, Ginkgo can't capture the panic.
I0111 14:28:31.011] To circumvent this, you should call
I0111 14:28:31.011]
I0111 14:28:31.012] defer GinkgoRecover()
I0111 14:28:31.012]
I0111 14:28:31.012] at the top of the goroutine that caused this panic.
I0111 14:28:31.012]
I0111 14:28:31.012]
I0111 14:28:31.012] goroutine 1028 [running]:
I0111 14:28:31.013] k8s.io/kubernetes/vendor/github.com/onsi/ginkgo.Fail(0xc421098000, 0xb0, 0xc420da24c8, 0x1, 0x1)
I0111 14:28:31.013] /go/src/k8s.io/kubernetes/_output/local/go/src/k8s.io/kubernetes/vendor/github.com/onsi/ginkgo/ginkgo_dsl.go:255 +0xda
I0111 14:28:31.014] k8s.io/kubernetes/vendor/github.com/onsi/gomega/internal/assertion.(*Assertion).match(0xc4220bd700, 0x9e897e0, 0xa123640, 0x0, 0x0, 0x0, 0x0, 0xa123640)
I0111 14:28:31.014] /go/src/k8s.io/kubernetes/_output/local/go/src/k8s.io/kubernetes/vendor/github.com/onsi/gomega/internal/assertion/assertion.go:69 +0x1ef
I0111 14:28:31.014] k8s.io/kubernetes/vendor/github.com/onsi/gomega/internal/assertion.(*Assertion).NotTo(0xc4220bd700, 0x9e897e0, 0xa123640, 0x0, 0x0, 0x0, 0xc4220bd700)
I0111 14:28:31.015] /go/src/k8s.io/kubernetes/_output/local/go/src/k8s.io/kubernetes/vendor/github.com/onsi/gomega/internal/assertion/assertion.go:43 +0xae
I0111 14:28:31.015] k8s.io/kubernetes/test/e2e_node.deletePodsSync.func1(0xc421485220, 0xc421321680, 0xc421517180)
I0111 14:28:31.015] /go/src/k8s.io/kubernetes/_output/local/go/src/k8s.io/kubernetes/test/e2e_node/resource_collector.go:382 +0x320
I0111 14:28:31.015] created by k8s.io/kubernetes/test/e2e_node.deletePodsSync
I0111 14:28:31.016] /go/src/k8s.io/kubernetes/_output/local/go/src/k8s.io/kubernetes/test/e2e_node/resource_collector.go:375 +0x9e
```
e.g.: https://storage.googleapis.com/kubernetes-jenkins/logs/ci-cri-containerd-node-e2e-serial/17/build-log.txt
**Release note**:
```release-note
none
```
Automatic merge from submit-queue. 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>.
Add stub device plugin for conformance e2e test
**What this PR does / why we need it**:
Add stub device plugin for conformance e2e test
- extend [device_plugin_stub](https://github.com/kubernetes/kubernetes/blob/master/pkg/kubelet/cm/deviceplugin/device_plugin_stub.go) to support e2e test
- add test suite with this device-plugin-stub
- simulate more use cases by deploying some pods to request these resources
**Which issue this PR fixes**:
fixes#52861
**Special notes for your reviewer**:
@vishh @jiayingz PTAL.
**Release note**:
```release-note
None
```
Automatic merge from submit-queue. 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>.
[Alpha: DynamicKubeletConfig] Double check before setKubeletConfiguration
**What this PR does / why we need it**:
Double check the `newCfg` is not equal to the `oldCfg` before we call `setKubeletConfiguration(newCfg)` in `tempSetCurrentKubeletConfig()`.
**Which issue(s) this PR fixes**:
Fixes https://github.com/kubernetes/kubernetes/issues/57701
**Special notes for your reviewer**:
/area kubelet
/sig node
/assign @mtaufen
/cc @vishh @jiayingz @derekwaynecarr @dchen1107 @liggitt
PTAL, Thanks!
**Release note**:
```release-note
NONE
```
Automatic merge from submit-queue. 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>.
Node e2e non docker specific
Fixes https://github.com/kubernetes/kubernetes/issues/57977.
Make node e2e test generic to container runtimes.
With this change, other than tests with `[Feature:Docker]`, all tests can run against all CRI container runtimes.
Note that this PR also marks cpu manager test as `Serial`, because it restarts kubelet during the test. It doesn't cause problem in regular node e2e suite today, because it is skipped if node has less than 2 CPUs, which is the case for our test environment. /cc @balajismaniam
@yujuhong @mrunalp @feiskyer
/cc @dashpole @balajismaniam @bprashanth Because I addressed your comments.
/cc @kubernetes/sig-node-pr-reviews
**Release note**:
```release-note
none
```
Automatic merge from submit-queue. 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>.
e2e node framework can generate a base kubelet config file
Fixes#57980
This allows the e2e node test framework to generate a Kubelet config file containing the defaults it would typically pass to a test via flags, rather than passing these defaults as flags.
```release-note
NONE
```
This allows the e2e node test framework to generate a kubelet config
file containing the defaults it would typically pass to a test via
flags, rather than passing these defaults as flags.
Automatic merge from submit-queue (batch tested with PRs 57532, 57392). 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>.
Gke spec tweaks
**What this PR does / why we need it**:
This PR removes two unnecessary requirements for the GKE node image validation spec:
- The "vim" package doesn't need to be installed, as a vim environment is already available and the full "vim" pacakge installation takes some precious disk space.
- The linux headers are not needed for the kubernetes node and cluster tests to succeed, and again, take unnecessary disk space.
**Special notes for your reviewer**:
None.
**Release note**:
```release-note
NONE
```
Automatic merge from submit-queue. 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>.
dynamic config test: use a hyphen between the config name and the unique suffix
These are painful to read right now due to the lack of hyphen.
```release-note
NONE
```
Automatic merge from submit-queue (batch tested with PRs 57434, 57221, 57417, 57474, 57481). 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>.
Send an event just before the Kubelet restarts to use a new config
**What this PR does / why we need it**:
This PR makes the Kubelet send events for configuration changes. This makes it much easier to see a recent history of configuration changes for the Kubelet.
**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#56895
**Special notes for your reviewer**:
**Release note**:
```release-note
NONE
```
/cc @dchen1107 @liggitt @dashpole
Automatic merge from submit-queue (batch tested with PRs 54278, 56259, 56762). 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>.
make sure of deleting archive
**What this PR does / why we need it**:
Exit() causes the current program to exit with the given status code, but deferred function does not run.
Automatic merge from submit-queue. 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>.
Version bump to etcd v3.2.11, grpc v1.7.5
Fix https://github.com/kubernetes/kubernetes/issues/56114: Update to etcd client 3.2.11
Version bumps:
- etcd from 3.1.10 to 3.2.11
- grpc from 1.3.0 to 1.7.5
- grpc-gateway from v1.1.0-25-g84398b9 to v1.3.0
TODO:
- [x] Apply etcd [3.2 client upgrade guide](https://github.com/coreos/etcd/blob/master/Documentation/upgrades/upgrade_3_2.md)
- [x] Apply grpc API changes in 1.6.0 and 1.7.0 [release notes](https://github.com/grpc/grpc-go/releases)
- [x] bbolt was pulled in transitively, why? We have tests that embed etcd, so we must vendor the etcd server and all it's dependencies.
- [x] Upgrade to containerd v1.0.0? Currently kubernetes depends on containerd v1.0.0-beta.2-159-g27d450a0 which depends on grpc v1.3.0, but containerd v1.0.0 depends on grpc 1.7.2. Not needed. The containerd grpc upgrade required [no code changes](ce3e32680d).
- [x] Fix all failing tests
- [x] Ensure we can safely upgrade grpc to 1.7.5 given that docker and cAdvisor still depend on grpc 1.3.0 (both in the versions we vend and on master for both projects). Should we hold off on this change until we have a docker release that uses gprc 1.7.x?
- [x] Wait for grpc 1.7.5 to be released (it will include https://github.com/grpc/grpc-go/pull/1747). Once released, bump grpc version in this PR and remove workarounds in `hack/godep-save.sh`.
Transitive dependencies on grpc:
- docker depends on grpc, but according to the package dependency graph (`go list -f '{{ .Deps }}'`) there are no dependencies from kubernetes to grpc via docker packages.
- containerd v1.0.0 depends on grpc 1.7.2, we should upgrade to containerd v1.0.0 soon, this can be done in a separate PR
- cadvisor depends on grpc 1.3.0 on master, it should upgrade it to grpc 1.7.5, this can be done in a separate PR
**Release note**:
```release-note
Upgrade to etcd client 3.2.11 and grpc 1.7.5 to improve HA etcd cluster stability.
```
The linux headers take significant disk space and are not necessary to
run kubernetes on a GKE node. User logging on to a node can trivially
install the kernel headers should they need to by running "apt-get install
linux-headers-$(uname -r)".
Signed-off-by: Chris Glass <chris.glass@canonical.com>
The minimal Ubuntu image used on GKE nodes provides the vim editor as
part of system packages, as "vim.tiny". People logging on the nodes have a vim
environment available despite the "vim" package not being installed.
Signed-off-by: Chris Glass <chris.glass@canonical.com>
Automatic merge from submit-queue (batch tested with PRs 55475, 57155, 57260, 57222). 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>.
Fix e2e local test
**What this PR does / why we need it**:
Fix some issue on local e2e_node test: `Can't start e2e service "kubelet"`
**Which issue(s) this PR fixes**:
Fixes https://github.com/kubernetes/kubernetes/issues/54622
**Special notes for your reviewer**:
**Release note**:
```release-note
```
/sig node
Automatic merge from submit-queue. 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>.
e2e_node: small fixes to setup_host.sh for Ubuntu Trusty
**What this PR does / why we need it**:
Two small fixes for how nsenter is built from source and installed on Ubuntu Trusty:
1. Use mktemp for the creating the build directory instead of a hard coded name
2. Use current (2.31) util-linux instead of 3+ year old version
**Which issue(s) this PR fixes**:
Fixes#56106
**Special notes for your reviewer**:
See https://github.com/kubernetes/kubernetes/issues/56106 for some thoughts on other ways to address this. My patch for util-linux 2.31 may just be a band-aid?
**Release note**:
```release-note
NONE
```
Kubernetes Submit Queue