This change is needed for running the latest containerd inside Docker
that is not aware of the recently added caps (BPF, PERFMON, CHECKPOINT_RESTORE).
Without this change, containerd inside Docker fails to run containers with
"apply caps: operation not permitted" error.
See kubernetes-sigs/kind 2058
NOTE: The caller process of this function is now assumed to be as
privileged as possible.
Signed-off-by: Akihiro Suda <akihiro.suda.cz@hco.ntt.co.jp>
Previously we simply ignored any not found error when loading the
containerd config. This created unintuitive behavior:
- If the user specified a path that didn't exist via --config, we would
silently ignore the error.
- If a config specified an import that didn't exist, we would silently
ignore the error.
In either of these cases, it appears we would end up using a potentially
corrupted config, as it would contain any files that were merged into it
before the not found error was hit.
However, we can't just remove the check for !os.IsNotExist(err),
as we shouldn't throw an error when --config is not passed, but the
default config doesn't exist.
This change updates the logic to only attempt to load the config if
we know it exists, or the user passed --config.
Signed-off-by: Kevin Parsons <kevpar@microsoft.com>
full diff: 0b889c03f1...8a68de567b
- utils: export RunningInUserNS()
- Support updates for net_cls (adds netclsController.Update())
- Only append Hugetlb in Subsystems list when available
- dependency updates (for compatibility with recent Go versions):
- go.mod: github.com/cilium/ebpf v0.2.0
- go.mod: github.com/coreos/go-systemd v22.1.0
- go.mod: sirupsen/logrus v1.7.0
Signed-off-by: Sebastiaan van Stijn <github@gone.nl>
https://github.com/actions/setup-go/tree/v2.1.3#v2
The V2 offers:
- Adds GOBIN to the PATH
- Proxy Support
- stable input
- Bug Fixes (including issues around version matching and semver)
Signed-off-by: Sebastiaan van Stijn <github@gone.nl>
The current code simply ignores the full binary path when starting the
shimv2 process, and instead fallbacks to a binary in the path, and this
is problematic (and confusing) for those using CRI-O, which has this
bits vendored.
The reason it's problematic with CRI-O is because the user can simply
set the full binary path and, instead of having that executed, CRI-O
will simply fail to create the container unless that binary is part of
the path, which may not be case in a few different scenarios (testing
being the most common one).
Fixes: #5006
Signed-off-by: Fabiano Fidêncio <fidencio@redhat.com>
Newer golangci-lint needs explicit `//` separator. Otherwise it treats
the entire line (`staticcheck deprecated ... yet`) as a name.
https://golangci-lint.run/usage/false-positives/#nolint
Signed-off-by: Kazuyoshi Kato <katokazu@amazon.com>
setupLoop()'s Autoclear (LO_FLAGS_AUTOCLEAR) will destruct the
loopback device when all associated file descriptors are closed.
However this behavior didn't work before since setupLoop() was
returning a file name. The looppack device was destructed at
the end of the function when LoopParams had Autoclear = true.
Fixes#4969.
Signed-off-by: Kazuyoshi Kato <katokazu@amazon.com>
download.fedoraproject.org gives HTTP 404 at times,
breaking the CI. Let's give it another chance.
Signed-off-by: Sebastiaan van Stijn <github@gone.nl>
Looks like this import was not needed for the test; simplified the test
by just using the device-path (a counter would work, but for debugging,
having the list of paths can be useful).
Signed-off-by: Sebastiaan van Stijn <github@gone.nl>
For some tools having the actual image name in the annotations is helpful for
debugging and auditing the workload.
Signed-off-by: Michael Crosby <michael@thepasture.io>
The variable `buf` is connected to containerd's stdout. So, accessing
the variable before killing containerd can cause a race condition.
Signed-off-by: Kazuyoshi Kato <katokazu@amazon.com>
This test has been flaky in GitHub Actions. This change logs the
values from devmapper to further investigate the issue.
Signed-off-by: Kazuyoshi Kato <katokazu@amazon.com>
Use the existing `.Close()` method instead of implementing the same
logic in this function.
The defer sets `cios` to `nil` if an error occurred to preserve the
existing behavior.
Signed-off-by: Sebastiaan van Stijn <github@gone.nl>
The defer functions were checking the local variable, and would therefore
not be executed, as the function returned if an error occurred.
Perhaps best illustrated when renaming the local variables;
if fifos.Stdin != "" {
l, err1 := winio.ListenPipe(fifos.Stdin, nil)
if err1 != nil {
return nil, errors.Wrapf(err1, "failed to create stdin pipe %s", fifos.Stdin)
}
defer func(l net.Listener) {
if err1 != nil {
l.Close()
}
}(l)
// ...
}
if fifos.Stdout != "" {
l, err2 := winio.ListenPipe(fifos.Stdout, nil)
if err2 != nil {
return nil, errors.Wrapf(err2, "failed to create stdout pipe %s", fifos.Stdout)
}
defer func(l net.Listener) {
if err2 != nil {
l.Close()
}
}(l)
// ....
}
This patch changes the function to use a named return variable, and to use
a single `defer()` that closes all pipes.
Signed-off-by: Sebastiaan van Stijn <github@gone.nl>
This change is mostly defensive; when checking for the returned
error, it's easy to make a mistake, and check for a "local" error,
not the actual returned error.
This patch changes the function to use a named return variable,
which is checked in the defer.
Signed-off-by: Sebastiaan van Stijn <github@gone.nl>
On newer kernels and systems, AppArmor will block sending signals in
many scenarios by default resulting in strange behaviours (container
programs cannot signal each other, or host processes like containerd
cannot signal containers).
The reason this happens only on some distributions (and is not a kernel
regression) is that the kernel doesn't enforce signal mediation unless
the profile contains signal rules. However because our profies #include
the distribution-managed <abstractions/base>, some distributions added
signal rules -- which results in AppArmor enforcing signal mediation and
thus a regression. On these systems, containers cannot send and receive
signals at all -- meaning they cannot signal each other and the
container runtime cannot kill them either.
This issue was fixed in Docker in 2018[1] but this code was copied
before then and thus the patches weren't carried. It also contains a new
fix for a more esoteric case[2]. Ideally this code should live in a
project like "containerd/apparmor" so that Docker, libpod, and
containerd can share it, but that's probably something to do separately.
In addition, the copyright header is updated to reference that the code
is copied from Docker (and thus was not written entirely by the
containerd authors).
[1]: https://github.com/docker/docker/pull/37831
[2]: https://github.com/docker/docker/pull/41337
Signed-off-by: Aleksa Sarai <cyphar@cyphar.com>
