This refactor removes the legacy KubeletConfig object and adds a new
KubeletDeps object, which contains injected runtime objects and
separates them from static config. It also reduces NewMainKubelet to two
arguments: a KubeletConfiguration and a KubeletDeps.
Some mesos and kubemark code was affected by this change, and has been
modified accordingly.
And a few final notes:
KubeletDeps:
KubeletDeps will be a temporary bin for things we might consider
"injected dependencies", until we have a better dependency injection
story for the Kubelet. We will have to discuss this eventually.
RunOnce:
We will likely not pull new KubeletConfiguration from the API server
when in runonce mode, so it doesn't make sense to make this something
that can be configured centrally. We will leave it as a flag-only option
for now. Additionally, it is increasingly looking like nobody actually uses the
Kubelet's runonce mode anymore, so it may be a candidate for deprecation
and removal.
Automatic merge from submit-queue
rkt: Improve support for privileged pod (pod whose all containers are privileged)
Fix https://github.com/kubernetes/kubernetes/issues/31100
This takes advantage of https://github.com/coreos/rkt/pull/2983 . By appending the new `--all-run` insecure-options to `rkt run-prepared` command when all the containers are privileged. The pod now gets more privileged power.
Automatic merge from submit-queue
Add sysctl support
Implementation of proposal https://github.com/kubernetes/kubernetes/pull/26057, feature https://github.com/kubernetes/features/issues/34
TODO:
- [x] change types.go
- [x] implement docker and rkt support
- [x] add e2e tests
- [x] decide whether we want apiserver validation
- ~~[ ] add documentation~~: api docs exist. Existing PodSecurityContext docs is very light and links back to the api docs anyway: 6684555ed9/docs/user-guide/security-context.md
- [x] change PodSecurityPolicy in types.go
- [x] write admission controller support for PodSecurityPolicy
- [x] write e2e test for PodSecurityPolicy
- [x] make sure we are compatible in the sense of https://github.com/kubernetes/kubernetes/blob/master/docs/devel/api_changes.md
- [x] test e2e with rkt: it only works with kubenet, not with no-op network plugin. The later has no sysctl support.
- ~~[ ] add RunC implementation~~ (~~if that is already in kube,~~ it isn't)
- [x] update whitelist
- [x] switch PSC fields to annotations
- [x] switch PSP fields to annotations
- [x] decide about `--experimental-whitelist-sysctl` flag to be additive or absolute
- [x] decide whether to add a sysctl node whitelist annotation
### Release notes:
```release-note
The pod annotation `security.alpha.kubernetes.io/sysctls` now allows customization of namespaced and well isolated kernel parameters (sysctls), starting with `kernel.shm_rmid_forced`, `net.ipv4.ip_local_port_range`, `net.ipv4.tcp_max_syn_backlog` and `net.ipv4.tcp_syncookies` for Kubernetes 1.4.
The pod annotation `security.alpha.kubernetes.io/unsafeSysctls` allows customization of namespaced sysctls where isolation is unclear. Unsafe sysctls must be enabled at-your-own-risk on the kubelet with the `--experimental-allowed-unsafe-sysctls` flag. Future versions will improve on resource isolation and more sysctls will be considered safe.
```
Automatic merge from submit-queue
Increase request timeout based on termination grace period
When terminationGracePeriodSeconds is set to > 2 minutes (which is
the default request timeout), ContainerStop() times out at 2 minutes.
We should check the timeout being passed in and bump up the
request timeout if needed.
Fixes#31219
Automatic merge from submit-queue
Kubelet code move: volume / util
Addresses some odds and ends that I apparently missed earlier. Preparation for kubelet code-move ENDGAME.
cc @kubernetes/sig-node
Automatic merge from submit-queue
Kubelet: implement GetNetNS for new runtime api
Kubelet: implement GetNetNS for new runtime api.
CC @yujuhong @thockin @kubernetes/sig-node @kubernetes/sig-rktnetes
Automatic merge from submit-queue
Add kubelet --network-plugin-mtu flag for MTU selection
* Add network-plugin-mtu option which lets us pass down a MTU to a network provider (currently processed by kubenet)
* Add a test, and thus make sysctl testable
When terminationGracePeriodSeconds is set to > 2 minutes (which is
the default request timeout), ContainerStop() times out at 2 minutes.
We should check the timeout being passed in and bump up the
request timeout if needed.
Fixes#31219
Automatic merge from submit-queue
[kubelet test] Improve node status test debug info
I find the output format `%v` of glog couldn't output useful information of an `api.Node` object. The output of this line https://github.com/kubernetes/kubernetes/blob/master/pkg/kubelet/kubelet_node_status_test.go#L492
is
```
kubelet_node_status_test.go:491: expected
&TypeMeta{Kind:,APIVersion:,}
, got
&TypeMeta{Kind:,APIVersion:,}
```
- It's difficult for me to tell the difference between expected and got.
- I prefer to use `diff.ObjectDiff(expectedNode, updatedNode)` to output the debug information as it will point out the starting character of the different objects.
I think this line https://github.com/kubernetes/kubernetes/blob/master/pkg/kubelet/kubelet_node_status_test.go#L647 can use `diff.ObjectDiff()` as well.
The related issus is #30952
MTU selection is difficult, and if there is a transport such as IPSEC in
use may be impossible. So we allow specification of the MTU with the
network-plugin-mtu flag, and we pass this down into the network
provider.
Currently implemented by kubenet.
Automatic merge from submit-queue
Kubelet: pass pod name/namespace/uid in new runtime API
First part of #30463.
Pass pod name/namespace/uid in new runtime API and change dockershim to build unique sandbox/container name based on them.
CC @yujuhong @euank @yifan-gu @kubernetes/sig-node
Automatic merge from submit-queue
remove duplicate code in updatePodCIDR
As kl.runtimeState.podCIDR() is a sync method, need fetch lock and release lock, so we only invoke once here
Kubernetes uses Docker labels as storage for some internal labels. The
majority of these labels are not meaningful metric labels and a few of
them are even harmful as they're not static and cause wrong aggregation
results.
This change provides a custom labels func to only attach meaningful
labels to cAdvisor exported metrics.
Automatic merge from submit-queue
Kubelet: implement GetPods for new runtime API
Implement GetPods for kuberuntime. Part of #28789 .
CC @yujuhong @Random-Liu
Automatic merge from submit-queue
Kubelet: add --container-runtime-endpoint and --image-service-endpoint
Flag `--container-runtime-endpoint` (overrides `--container-runtime`) is introduced to identify the unix socket file of the remote runtime service. And flag `--image-service-endpoint` is introduced to identify the unix socket file of the image service.
This PR is part of #28789 Milestone 0.
CC @yujuhong @Random-Liu
Automatic merge from submit-queue
rkt: Support subPath volume mounts feature
So that at most one volume object will be created for every unique
host path. Also the volume's name is random generated UUID to avoid
collision since the mount point's name passed by kubelet is not
guaranteed to be unique when 'subpath' is specified.
Should partially fix https://github.com/kubernetes/kubernetes/issues/26986
The non-existing host path creation issue is not touched here.
cc @kubernetes/sig-rktnetes
also cc @kubernetes/sig-node for the Mount name comments I added.
Automatic merge from submit-queue
kubelet status manager: Fix nil in error message due to var shadowing
Variable shadowing can cause this log message to print a nil:
```go
glog.Warningf("Failed to update status for pod %q: %v", format.Pod(pod), err)
```
@kubernetes/rh-cluster-infra
The serviceAccountName is occasionally useful for clients running on
Kube that need to know who they are when talking to other components.
The nodeName is useful for PetSet or DaemonSet pods that need to make
calls back to the API to fetch info about their node.
Both fields are immutable, and cannot easily be retrieved in another
way.
Automatic merge from submit-queue
Fix image inspection and matching
An image string could contain a hostname (e.g., "docker.io") or not. The same
applies to the RepoTags returned from an image inspection. To determine whether
the image docker pulled matches what the user ask for, we check if the either
string is the suffix of the other.
/cc @dims @dchen1107 @Random-Liu
This fixes#30710
Automatic merge from submit-queue
Always return command output for exec probes and kubelet RunInContainer
Always return command output for exec probes and kubelet RunInContainer, even if the command invocation returns nonzero.
When #24921 replaced RunInContainer with ExecInContainer, it introduced a change where an exec probe that failed no longer included the stdout/stderr from the probe in the event. For example, when running at log level 4, you see:
```
I0816 15:01:36.259826 29713 exec.go:38] Exec probe response: "Failed to access the status endpoint : HTTP Error 404: Not Found.\nHawkular metrics has only been running for 7\n seconds not aborting yet.\n"
```
But the event looks like this:
```
54s 22s 5 hawkular-metrics-hjme4 Pod spec.containers{hawkular-metrics} Warning Unhealthy {kubelet corbeau} Readiness probe failed:
```
Note the absence of the exec probe response after "Readiness probe failed". This PR restores the previous behavior.
cc @kubernetes/rh-cluster-infra @mwringe
xref https://github.com/openshift/origin/issues/10424
Automatic merge from submit-queue
Unblock iterative development on pod-level cgroups
In order to allow forward progress on this feature, it takes the commits from #28017#29049 and then it globally disables the flag that allows these features to be exercised in the kubelet. The flag can be re-added to the kubelet when its actually ready.
/cc @vishh @dubstack @kubernetes/rh-cluster-infra
Automatic merge from submit-queue
dockertools: Don't use network plugin if net=host
I'm pretty sure this was just an oversight the first time around.
Before: `E0815 18:06:17.627468 976 docker_manager.go:350] NetworkPlugin kubenet failed on the status hook for pod 'sleep' - Unexpected command output Device "eth0" does not exist.`
After: No such logline is printed
The pod IP reported in `describe` is the same either way
cc @kubernetes/sig-node
Automatic merge from submit-queue
rkt: Do not error out when there are unrecognized lines in os-release
Also fix the error handling which will cause panic. Also fix the error handling which will cause panic.
cc @kubernetes/sig-rktnetes
So that at most one volume object will be created for every unique
host path. Also the volume's name is random generated UUID to avoid
collision since the mount point's name passed by kubelet is not
guaranteed to be unique when 'subpath' is specified.
Automatic merge from submit-queue
kubelet/api: split RuntimeService interface
Splits `RuntimeService` interface into smaller interfaces
to make testing easier and delineate the responsibilities.
Its a non-breaking change to the previous users of `api.RuntimeService`
Automatic merge from submit-queue
Add Events for operation_executor to show status of mounts, failed/successful to show in describe events
Fixes#27590
@saad-ali @pmorie @erinboyd
After talking with @pmorie last week about the above issue, I decided to poke around and see if I could remedy. The refactoring broke my previous UXP merged PR's that correctly showed failed mount errors in the describe events. However, Not sure I implemented correctly, but it tested out and seems to be working, let me know what I missed or if this is not the correct approach.
```
Events:
FirstSeen LastSeen Count From SubobjectPath Type Reason Message
--------- -------- ----- ---- ------------- -------- ------ -------
2m 2m 1 {default-scheduler } Normal Scheduled Successfully assigned nfs-bb-pod1 to 127.0.0.1
44s 44s 1 {kubelet 127.0.0.1} Warning FailedMount Unable to mount volumes for pod "nfs-bb-pod1_default(a94f64f1-37c9-11e6-9aa5-52540073d346)": timeout expired waiting for volumes to attach/mount for pod "nfs-bb-pod1"/"default". list of unattached/unmounted volumes=[nfsvol]
44s 44s 1 {kubelet 127.0.0.1} Warning FailedSync Error syncing pod, skipping: timeout expired waiting for volumes to attach/mount for pod "nfs-bb-pod1"/"default". list of unattached/unmounted volumes=[nfsvol]
38s 38s 1 {kubelet } Warning FailedMount Unable to mount volumes for pod "a94f64f1-37c9-11e6-9aa5-52540073d346": Mount failed: exit status 32
Mounting arguments: nfs1.rhs:/opt/data99 /var/lib/kubelet/pods/a94f64f1-37c9-11e6-9aa5-52540073d346/volumes/kubernetes.io~nfs/nfsvol nfs []
Output: mount.nfs: Connection timed out
Resolution hint: Check and make sure the NFS Server exists (ensure that correct IPAddress/Hostname was given) and is available/reachable.
Also make sure firewall ports are open on both client and NFS Server (2049 v4 and 2049, 20048 and 111 for v3).
Use commands telnet <nfs server> <port> and showmount <nfs server> to help test connectivity.
```
New flag --container-runtime-endpoint (overrides --container-runtime)
is introduced to kubelet which identifies the unix socket file of
the remote runtime service. And new flag --image-service-endpoint is
introduced to kubelet which identifies the unix socket file of the
image service.
An image string could contain a hostname (e.g., "docker.io") or not. The same
applies to the RepoTags returned from an image inspection. To determine whether
the image docker pulled matches what the user ask for, we check if the either
string is the suffix of the other.
Automatic merge from submit-queue
Fix default resource limits (node allocatable) for downward api volumes and env vars
@kubernetes/rh-cluster-infra @pmorie @derekwaynecarr
Automatic merge from submit-queue
Add note: kubelet manages only k8s containers.
Kubelet wrote log when accesing container which was not created in k8s, what could confuse users. That's why we added note about it in documentation and lowered log level of the message to 5.
Here is example of the message:
```
> Apr 19 11:50:32 openshift-114.lab.sjc.redhat.com atomic-openshift-node[9551]:
I0419 11:50:32.194020 9600 docker.go:363]
Docker Container: /tiny_babbage is not managed by kubelet.
```
bug 1328441
Bugzilla link https://bugzilla.redhat.com/show_bug.cgi?id=1328441
Automatic merge from submit-queue
syncNetworkUtil in kubelet and fix loadbalancerSourceRange on GCE
fixes: #29997#29039
@yujuhong Can you take a look at the kubelet part?
@girishkalele KUBE-MARK-DROP is the chain for dropping connections. Marked connection will be drop in INPUT/OUTPUT chain of filter table. Let me know if this is good enough for your use case.
Automatic merge from submit-queue
Implement AppArmor Kubelet support
Includes PR https://github.com/kubernetes/kubernetes/pull/29812
Implements the Kubelet logic for AppArmor based on the alpha API proposed [here](https://github.com/kubernetes/kubernetes/blob/master/docs/proposals/apparmor.md). Also adds an E2E test, and I ran manual tests.
Remaining work: PodSecurityPolicy support, profile loader daemon, documentation, (maybe) beta API.
/cc @jfrazelle @Amey-D @kubernetes/sig-node
*Note on release-note-none: I am implementing AppArmor over multiple PRs. I will submit a single release note once the implementation is done to cover all of them.*
This was already handled in most places. I think this is the only
remaining instance of it in the docker package.
This could lead to confusing results. E.g. if `networkPlugin` was cni,
it could lead to error logs about not getting network status for host
pods if eth0 didn't exist on the host.
Automatic merge from submit-queue
Fix image verification when hostname is present in image
Deal better with the situation where a image name contains
a hostname as well.
Fixes#30580
Automatic merge from submit-queue
Add volume reconstruct/cleanup logic in kubelet volume manager
Currently kubelet volume management works on the concept of desired
and actual world of states. The volume manager periodically compares the
two worlds and perform volume mount/unmount and/or attach/detach
operations. When kubelet restarts, the cache of those two worlds are
gone. Although desired world can be recovered through apiserver, actual
world can not be recovered which may cause some volumes cannot be cleaned
up if their information is deleted by apiserver. This change adds the
reconstruction of the actual world by reading the pod directories from
disk. The reconstructed volume information is added to both desired
world and actual world if it cannot be found in either world. The rest
logic would be as same as before, desired world populator may clean up
the volume entry if it is no longer in apiserver, and then volume
manager should invoke unmount to clean it up.
Fixes https://github.com/kubernetes/kubernetes/issues/27653
Automatic merge from submit-queue
CRI: remove pod sandbox resources
The pod-level resources need further discussion. Remove it from CRI for now.
See the original discussion in #29871
Currently kubelet volume management works on the concept of desired
and actual world of states. The volume manager periodically compares the
two worlds and perform volume mount/unmount and/or attach/detach
operations. When kubelet restarts, the cache of those two worlds are
gone. Although desired world can be recovered through apiserver, actual
world can not be recovered which may cause some volumes cannot be cleaned
up if their information is deleted by apiserver. This change adds the
reconstruction of the actual world by reading the pod directories from
disk. The reconstructed volume information is added to both desired
world and actual world if it cannot be found in either world. The rest
logic would be as same as before, desired world populator may clean up
the volume entry if it is no longer in apiserver, and then volume
manager should invoke unmount to clean it up.
Automatic merge from submit-queue
move syncNetworkConfig to Init for cni network plugin
start syncNetworkConfig routine in `Init` instead of probing. This fixes a bug where the syncNetworkConfig runs periodically even `cni` network plugin is not in use.
Automatic merge from submit-queue
Remove resource specifications from CRI until further notice
See #29871 for the discussion issue.
cc @dchen1107 @vishh @yujuhong @euank @yifan-gu @feiskyer
Automatic merge from submit-queue
Validate SHA/Tag when checking docker images
Docker API does not validate the tag/sha, for example, all the following
calls work say for a alpine image with short SHA "4e38e38c8ce0"
echo -e "GET /images/alpine:4e38e38c8ce0/json HTTP/1.0\r\n" | nc -U /var/run/docker.sock
echo -e "GET /images/alpine:4e38e38c/json HTTP/1.0\r\n" | nc -U /var/run/docker.sock
echo -e "GET /images/alpine:4/json HTTP/1.0\r\n" | nc -U /var/run/docker.sock
So we should check the response from the Docker API and look for the tags or SHA explicitly.
Fixes#30355
Automatic merge from submit-queue
Set pod state as "unknown" when CNI plugin fails
Before this change, CNI plugin failure didn't change anything in the pod status, so pods having containers without requested network were "running".
Fixes#29148
Automatic merge from submit-queue
Remove kubelet pkill dependency
Issue #26093 identified pkill as one of the dependencies of kublet
which could be worked around. Build on the code introduced for pidof
and regexp for the process(es) we need to send a signal to.
Related to #26093
Automatic merge from submit-queue
[kubelet] Introduce --protect-kernel-defaults flag to make the tunable behaviour configurable
Let's make the default behaviour of kernel tuning configurable. The default behaviour is kept modify as has been so far.
Docker API does not validate the tag/sha, for example, all the following
calls work say for a alpine image with short SHA "4e38e38c8ce0"
echo -e "GET /images/alpine:4e38e38c8ce0/json HTTP/1.0\r\n" | nc -U /var/run/docker.sock
echo -e "GET /images/alpine:4e38e38c/json HTTP/1.0\r\n" | nc -U /var/run/docker.sock
echo -e "GET /images/alpine:4/json HTTP/1.0\r\n" | nc -U /var/run/docker.sock
So we should check the response from the Docker API and look for the
tags or SHA explicitly.
Fixes#30355
Issue #26093 identified pkill as one of the dependencies of kublet
which could be worked around. Build on the code introduced for pidof
and regexp for the process(es) we need to send a signal to.
Related to #26093
Automatic merge from submit-queue
Remove kubelet dependency on pidof
Issue #26093 identified pidof as one of the dependencies of kublet
which could be worked around. In this PR, we just look at /proc
to construct the list of pids we need for a specified process
instead of running "pidof" executable
Related to #26093
<!-- Reviewable:start -->
---
This change is [<img src="https://reviewable.kubernetes.io/review_button.svg" height="34" align="absmiddle" alt="Reviewable"/>](https://reviewable.kubernetes.io/reviews/kubernetes/kubernetes/30002)
<!-- Reviewable:end -->
Issue #26093 identified pidof as one of the dependencies of kublet
which could be worked around. In this PR, we just look at /proc
to construct the list of pids we need for a specified process
instead of running "pidof" executable
Related to #26093
Automatic merge from submit-queue
Kubelet: Add container ports label.
Addresses https://github.com/kubernetes/kubernetes/pull/30049#discussion_r73983952.
Add container ports label, although we don't use it now, it will make it easier to switch to new runtime interface in the future.
@yujuhong @feiskyer
Automatic merge from submit-queue
Add total inodes to kubelet summary api
Needed to support inode based eviction thresholds as a percentage.
/cc @ronnielai @vishh @kubernetes/rh-cluster-infra
Before this change, CNI plugin failure didn't change anything in
the pod status, so pods having containers without requested
network were "running".
Fixes#29148
Automatic merge from submit-queue
[kubelet] Auto-discover node IP if neither cloud provider exists and IP is not explicitly specified
One example where the earlier implementation failed is when running kubelet on CoreOS (bare-metal), where the nameserver is set to `8.8.8.8`. kubelet tries to lookup the node name agains Google DNS, which obviously fails. The kubelet won't recover after that.
The workaround hsa been to set `--hostname-override` to an IP address, but it's quite annoying to try to make a multi-distro way of getting the IP in bash for example. This way is much cleaner.
Refactored the function a little bit at the same time
@vishh @yujuhong @resouer @Random-Liu
Automatic merge from submit-queue
Verify volume.GetPath() never returns ""
Add a new helper method volume.GetPath(Mounter) instead of calling
the GetPath() of the Mounter directly. Check if GetPath() is returning
a "" and convert that into an error.
Fixes#23163
Add a new helper method volume.GetPath(Mounter) instead of calling
the GetPath() of the Mounter directly. Check if GetPath() is returning
a "" and convert that into an error. At this point, we only have
information about the type of the Mounter, so let's log that if
there is a problem
Fixes#23163
Automatic merge from submit-queue
Node disk pressure should induce image gc
If the node reports disk pressure, prior to evicting pods, the node should clean up unused images.
Automatic merge from submit-queue
Consolidating image pulling logic
Moving image puller logic into image manager by consolidating 2 pullers into one implementation.
Automatic merge from submit-queue
Kubelet: add kubeGenericRuntimeManager for new runtime API
Part of #28789. Add `kubeGenericRuntimeManager` for kubelet new runtime API #17048.
Note that:
- To facilitate code reviewing, #28396 is splited into a few small PRs. This is the first part.
- This PR also fixes some syntax errors in `api.proto`.
- This PR is depending on #29811 (already merged).
CC @yujuhong @Random-Liu @kubernetes/sig-node
Automatic merge from submit-queue
Kubelet: add gRPC implementation of new runtime interface
Add gRPC implementation of new runtime interface.
CC @yujuhong @Random-Liu @kubernetes/sig-node
Automatic merge from submit-queue
Kubelet: add fake kube runtime
Add a new fake kube runtime with kubelet using the new runtime API.
CC @yujuhong @Random-Liu
Automatic merge from submit-queue
pods which can not be admitted should return directly
if the pod can not be admitted, the code runPod(pod, retryDelay) should not be run.
Automatic merge from submit-queue
Use the CNI bridge plugin to set hairpin mode
Following up this part of #23711:
> I'd like to wait until containernetworking/cni#175 lands and then just pass the request through to CNI.
The code here just
* passes the required setting down from kubenet to CNI
* disables `DockerManager` from doing hairpin-veth, if kubenet is in use
Note to test you need a very recent version of the CNI `bridge` plugin; the one brought in by #28799 should be OK.
Also relates to https://github.com/kubernetes/kubernetes/issues/19766#issuecomment-232722864
If the resource in the delete call does not exist, the runtime should
not return an error. This eliminates the need for kubelet to define a
resource "not found" error that every runtime has to return.
Automatic merge from submit-queue
pkg/various: plug leaky time.New{Timer,Ticker}s
According to the documentation for Go package time, `time.Ticker` and
`time.Timer` are uncollectable by garbage collector finalizers. They
leak until otherwise stopped. This commit ensures that all remaining
instances are stopped upon departure from their relative scopes.
Similar efforts were incrementally done in #29439 and #29114.
```release-note
* pkg/various: plugged various time.Ticker and time.Timer leaks.
```
Automatic merge from submit-queue
Initial support for pod eviction based on disk
This PR adds the following:
1. node reports disk pressure condition based on configured thresholds
1. scheduler does not place pods on nodes reporting disk pressure
1. kubelet will not admit any pod when it reports disk pressure
1. kubelet ranks pods for eviction when low on disk
1. kubelet evicts greediest pod
Follow-on PRs will need to handle:
1. integrate with new image gc PR (https://github.com/kubernetes/kubernetes/pull/27199)
1. container gc policy should always run (will not be launched from eviction, tbd who does that)
1. this means kill pod is fine for all eviction code paths since container gc will remove dead container
1. min reclaim support will just poll summary provider (derek will do follow-on)
1. need to know if imagefs is same device as rootfs from summary (derek follow-on)
/cc @vishh @kubernetes/sig-node
Michael Taufen
