This tag contains some fixes for hostprocess containers, mainly around
fixing task stats which regressed from a change in v0.9.3.
https://github.com/microsoft/hcsshim/releases/tag/v0.9.4
Signed-off-by: Daniel Canter <dcanter@microsoft.com>
The ghcr.io/containerd/registry:2.7 image does not support the ppc64le
architecture, causing the TestCRIImagePullTimeout tests to fail when
executed on a ppc64le device.
Replace the ghcr.io/containerd/registry:2.7 image name and index with
the ghcr.io/containerd/volume-ownership:2.1 image and index in the
HoldingContentOpenWriter test.
Change the image pulled in the NoDataTransferred test to the
ghcr.io/containerd/volume-ownership:2.1 image.
Signed-off-by: James Jenkins <James.Jenkins@ibm.com>
This tag brings in a new field to fix an HNS issue in ws2019 as well as
an optimization for collecting Windows stats (memory, cpu, iops).
Signed-off-by: Daniel Canter <dcanter@microsoft.com>
`ioutil` has been deprecated by golang. All the code in `ioutil` just
forwards functionality to code in either the `io` or `os` packages.
See https://github.com/golang/go/pull/51961 for more info.
Signed-off-by: Jeff Widman <jeff@jeffwidman.com>
Introduce cni-bridge-fp as CNI bridge plugin wrapper binary for CRI
testing.
With CNI `io.kubernetes.cri.pod-annotations` capability enabled, the user
can inject the failpoint setting by pod's annotation
`cniFailpointControlStateDir`, which stores each pod's failpoint setting
named by `${K8S_POD_NAMESPACE}-${K8S_POD_NAME}.json`.
When the plugin is invoked, the plugin will check the CNI_ARGS to get
the failpoint for the CNI_COMMAND from disk. For the testing, the user
can prepare setting before RunPodSandbox.
Signed-off-by: Wei Fu <fuweid89@gmail.com>
If there is any unskipable error during setuping shim plugins, we should
fail return error to prevent from leaky shim instance. For example,
there is error during init task plugin, the shim ttrpc server will not
contain any shim API method. The any call to the shim will receive that
failed to create shim task: service containerd.task.v2.Task: not implemented
Then containerd can't use `Shutdown` to let the shim close. The shim
will be leaky. And also fail return if there is no ttrpc service.
Signed-off-by: Wei Fu <fuweid89@gmail.com>
Added new runc shim binary in integration testing.
The shim is named by io.containerd.runc-fp.v1, which allows us to use
additional OCI annotation `io.containerd.runtime.v2.shim.failpoint.*` to
setup shim task API's failpoint. Since the shim can be shared with
multiple container, like what kubernetes pod does, the failpoint will be
initialized during setup the shim server. So, the following the
container's OCI failpoint's annotation will not work.
This commit also updates the ctr tool that we can use `--annotation` to
specify annotations when run container. For example:
```bash
➜ ctr run -d --runtime runc-fp.v1 \
--annotation "io.containerd.runtime.v2.shim.failpoint.Kill=1*error(sorry)" \
docker.io/library/alpine:latest testing sleep 1d
➜ ctr t ls
TASK PID STATUS
testing 147304 RUNNING
➜ ctr t kill -s SIGKILL testing
ctr: sorry: unknown
➜ ctr t kill -s SIGKILL testing
➜ sudo ctr t ls
TASK PID STATUS
testing 147304 STOPPED
```
The runc-fp.v1 shim is based on core runc.v2. We can use it to inject
failpoint during testing complicated or big transcation API, like
kubernetes PodRunPodsandbox.
Signed-off-by: Wei Fu <fuweid89@gmail.com>
Currently, the runc shimv2 commandline manager doesn't support ttrpc
server's customized option, for example, the ttrpc server interceptor.
This commit is to allow the task plugin can return the
`UnaryServerInterceptor` option to the manager so that the task plugin
can do enhancement before handling the incoming request, like API-level
failpoint control.
Signed-off-by: Wei Fu <fuweid89@gmail.com>
Failpoint is used to control the fail during API call when testing, especially
the API is complicated like CRI-RunPodSandbox. It can help us to test
the unexpected behavior without mock. The control design is based on freebsd
fail(9), but simpler.
REF: https://www.freebsd.org/cgi/man.cgi?query=fail&sektion=9&apropos=0&manpath=FreeBSD%2B10.0-RELEASE
Signed-off-by: Wei Fu <fuweid89@gmail.com>
For k8s e2e node tests, we want to be able to test both cgroupv1 and
cgroupv2 variants. Currently, the latest version of COS (M97) is running
as cgroupv2 and we would like the add the ability to switch to cgroupv1
for running tests on cgroupv1.
To do this, add a new flag, "CGROUP_MODE" which can be set as v1 or v2.
If the current cgroup configuration does not match the desired config,
the cgroup config will be set via `cgroup_helper` COS utility, and the
node will be rebooted.
Signed-off-by: David Porter <porterdavid@google.com>
We are constantly getting the message below from Cirrus CI;
> Machine type with name 'c2-standard-2' does not exist in zone 'us-central1-c'."
Seems Cirrus CI is constructing the machine type based on
the "cpu" field from the YAML file. However c2-standard's minimal vCPU
count is 4 based on the official document below.
https://cloud.google.com/compute/docs/compute-optimized-machines
Signed-off-by: Kazuyoshi Kato <katokazu@amazon.com>
