Before this patch, a valid, but empty set of credentials would still
write a config-file to the container and set `DOCKER_CONFIG`:
mkdir -p tmpConfig
export DOCKER_CONFIG=$PWD/tmpConfig
echo '{}' > "${DOCKER_CONFIG}/config.json"
docker run --rm --use-api-socket alpine cat /run/secrets/docker/config.json
{
"auths": {}
}
echo '{"auths": {}}' > "${DOCKER_CONFIG}/config.json"
docker run --rm --use-api-socket alpine cat /run/secrets/docker/config.json
{
"auths": {}
}
echo '{"auths": {"https://index.docker.io/v1/": {"auth": "am9lam9lOmhlbGxv"}}}' > "${DOCKER_CONFIG}/config.json"
docker run --rm --use-api-socket alpine cat /run/secrets/docker/config.json
{
"auths": {
"https://index.docker.io/v1/": {
"auth": "am9lam9lOmhlbGxv"
}
}
}
With this patch, the `DOCKER_CONFIG` env-var and config-file are only created
if we have credentials to set;
mkdir -p tmpConfig
export DOCKER_CONFIG=$PWD/tmpConfig
echo '{}' > "${DOCKER_CONFIG}/config.json"
docker run --rm --use-api-socket alpine cat /run/secrets/docker/config.json
cat: can't open '/run/secrets/docker/config.json': No such file or directory
echo '{"auths": {}}' > "${DOCKER_CONFIG}/config.json"
docker run --rm --use-api-socket alpine cat /run/secrets/docker/config.json
cat: can't open '/run/secrets/docker/config.json': No such file or directory
echo '{"auths": {"https://index.docker.io/v1/": {"auth": "am9lam9lOmhlbGxv"}}}' > "${DOCKER_CONFIG}/config.json"
docker run --rm --use-api-socket alpine cat /run/secrets/docker/config.json
{
"auths": {
"https://index.docker.io/v1/": {
"auth": "am9lam9lOmhlbGxv"
}
}
}
Signed-off-by: Sebastiaan van Stijn <github@gone.nl>
Go maintainers started to unconditionally update the minimum go version
for golang.org/x/ dependencies to go1.23, which means that we'll no longer
be able to support any version below that when updating those dependencies;
> all: upgrade go directive to at least 1.23.0 [generated]
>
> By now Go 1.24.0 has been released, and Go 1.22 is no longer supported
> per the Go Release Policy (https://go.dev/doc/devel/release#policy).
>
> For golang/go#69095.
This updates our minimum version to go1.23, as we won't be able to maintain
compatibility with older versions because of the above.
Signed-off-by: Sebastiaan van Stijn <github@gone.nl>
Adds a flag to the create and run command, `--use-api-socket`, that can
be used to start a container with the correctly configured parameters to
ensure that accessing the docker socket will work with out managing bind
mounts and authentication injection.
The implementation in this PR resolves the tokens for the current
credential set in the client and then copies it into a container at the
well know location of /run/secrets/docker/config.json, setting
DOCKER_CONFIG to ensure it is resolved by existing tooling. We use a
compose-compatible secret location with the hope that the CLI and
compose can work together seamlessly.
The bind mount for the socket is resolved from the current context,
erroring out if the flag is set and the provided socket is not a unix
socket.
There are a few drawbacks to this approach but it resolves a long
standing pain point. We'll continue to develop this as we understand
more use cases but it is marked as experimental for now.
Signed-off-by: Stephen Day <stephen.day@docker.com>
Same functionality, but implemented with atomicwriter. There's a slight
difference in error-messages produced (but can be adjusted if we want).
Before:
docker container export -o ./no/such/foo mycontainer
failed to export container: invalid output path: directory "no/such" does not exist
docker container export -o /no/permissions mycontainer
failed to export container: stat /no/permissions: permission denied
After:
docker container export -o ./no/such/foo mycontainer
failed to export container: invalid file path: stat no/such: no such file or directory
docker container export -o /no/permissions mycontainer
failed to export container: failed to stat output path: lstat /no/permissions: permission denied
Signed-off-by: Sebastiaan van Stijn <github@gone.nl>
This looks like a copy/paste from other tests, because this test
does not test anything related to docker content trust.
Signed-off-by: Sebastiaan van Stijn <github@gone.nl>
Restore part of the code removed by 966b44183f
that closed the stream. It's required now because the Run command won't
finish before the output stream was processed by the caller.
Signed-off-by: Paweł Gronowski <pawel.gronowski@docker.com>
Fix a regression introduced by 30c4637f03
which made the `docker run` command produce potentially truncated
stdout/stderr output.
Previous implementation stopped the content streaming as soon as the
container exited which would potentially truncate a long outputs.
This change fixes the issue by only canceling the IO stream immediately
if neither stdout nor stderr is attached.
Signed-off-by: Paweł Gronowski <pawel.gronowski@docker.com>
Cobra now defines a CompletionFunc for the same, so we can alias
it to that, and stop using our own definition.
Signed-off-by: Sebastiaan van Stijn <github@gone.nl>
This validation is now handled by the API-client since [moby@5d6b566],
so no longer needed to be done in the cli. This function was only used
internally and has no external consumers, so removing it without
deprecating first.
[moby@5d6b566]: 5d6b56699d
Signed-off-by: Sebastiaan van Stijn <github@gone.nl>
Signed-off-by: Paweł Gronowski <pawel.gronowski@docker.com>
This patch fixes the `TestRunAttachTermination` flaky runs.
It seems like we weren't halting on the `waitFunc` so if the
process was fast enough to setup the signal handler and execute
`waitExitOrRemoved`. We now instead wait for the `killCh` channel
to close inside the mocked `waitFunc`.
Signed-off-by: Alano Terblanche <18033717+Benehiko@users.noreply.github.com>
Use a consistent approach for producing warnings, but add a TODO for moving
this warning to the daemon, which can make a better call if it will work
or not (depending on networking mode).
This warning was originally added in [moby@afa92a9], before integration with
libnetwork, and this warning may be incorrect in many scenarios.
While updating, also removing the custom regular expression used to
detect if the IP is a loopback address, and using go's netip package
instead.
[moby@afa92a9]: afa92a9af0
Signed-off-by: Sebastiaan van Stijn <github@gone.nl>
This warning was originally added in [moby@3aa70c1], and moved to be printed
on both `run` and `create` in commit 7c514a31c9.
However, [moby@57f1305] (docker 19.03, API 1.40) moved such warnings to
the daemon side. The patch mentioned this issue:
> This patch will have one side-effect; docker cli's that also perform this check
> client-side will print the warning twice; this can be addressed by disabling
> the cli-side check for newer API versions, but will generate a bit of extra
> noise when using an older CLI.
The CLI does not take this into account currently, and still prints warnings
twice; even in cases where the option is not supported by the daemon, and
discarded:
On a host without OomKillDisable support:
docker create --oom-kill-disable alpine
WARNING: Disabling the OOM killer on containers without setting a '-m/--memory' limit may be dangerous.
WARNING: Your kernel does not support OomKillDisable. OomKillDisable discarded.
On a host that supports it:
docker create --oom-kill-disable alpine
WARNING: Disabling the OOM killer on containers without setting a '-m/--memory' limit may be dangerous.
WARNING: OOM killer is disabled for the container, but no memory limit is set, this can result in the system running out of resources.
This patch removes the client-side warning, leaving it to the daemon to
report if any warnings should produced (and the client to print them).
With this patch applied:
On a host without OomKillDisable support:
docker create --oom-kill-disable alpine
WARNING: Your kernel does not support OomKillDisable. OomKillDisable discarded.
On a host that supports it:
docker create --oom-kill-disable alpine
WARNING: OOM killer is disabled for the container, but no memory limit is set, this can result in the system running out of resources.
[moby@3aa70c1]: 3aa70c1948
[moby@57f1305]: 57f1305e74
Signed-off-by: Sebastiaan van Stijn <github@gone.nl>
This function was shared between "image" and "container" packages,
all of which needed the trust package, so move it there instead.
Signed-off-by: Sebastiaan van Stijn <github@gone.nl>
cli/command/container/cp.go:206:3: naked return in func `resolveLocalPath` with 5 lines of code (nakedret)
return
^
Signed-off-by: Sebastiaan van Stijn <github@gone.nl>
cli/command/container/client_test.go:78:7: unused-receiver: method receiver 'f' is not referenced in method's body, consider removing or renaming it as _ (revive)
func (f *fakeClient) ContainerExecStart(context.Context, string, container.ExecStartOptions) error {
^
cli/command/container/create_test.go:383:7: unused-receiver: method receiver 'f' is not referenced in method's body, consider removing or renaming it as _ (revive)
func (f fakeNotFound) NotFound() {}
^
cli/command/container/create_test.go:384:7: unused-receiver: method receiver 'f' is not referenced in method's body, consider removing or renaming it as _ (revive)
func (f fakeNotFound) Error() string { return "error fake not found" }
^
Signed-off-by: Sebastiaan van Stijn <github@gone.nl>
Previously, multiple suggestions were provided when completing
commands like `run`, `history` and `push`. This change
limits completion to a single suggestion for the above and 2 suggestions for `tag`
Signed-off-by: Mohammed Aminu Futa <mohammedfuta2000@gmail.com>
Without breaking API compatibility, this patch allows us to know whether
a returned `cli/StatusError` was caused by a context cancellation or
not, which we can use to provide a nicer UX and not print the Go
"context canceled" error message if this is the cause.
Signed-off-by: Laura Brehm <laurabrehm@hey.com>
- use apiClient instead of client for the API client to
prevent shadowing imports.
- use dockerCLI with Go's standard camelCase casing.
- suppress some errors to make my IDE and linters happier
Signed-off-by: Sebastiaan van Stijn <github@gone.nl>
- use Println to print newline instead of custom format
- suppress some errors to make my IDE and linters happier
Signed-off-by: Sebastiaan van Stijn <github@gone.nl>
While there may be reasons to keep pkg/errors in production code,
we don't need them for these tests.
Signed-off-by: Sebastiaan van Stijn <github@gone.nl>
The '--link' option should only be migrated to an endpoint
option if the network is user-defined ... there was already
an exception for network "default", but not for "bridge".
Signed-off-by: Rob Murray <rob.murray@docker.com>
Allows for the `jsonmessage.DisplayJSONMessagesStream` function
to correctly return when the context is cancelled with the appropriate
reason (`ctx.Error()`) instead of just a nil error.
Follow-up to 30a73ff19c
Signed-off-by: Alano Terblanche <18033717+Benehiko@users.noreply.github.com>
Co-authored-by: Paweł Gronowski <pawel.gronowski@docker.com>
Currently the cp will tar from the same directory it will untar into
simultaneously. There is a race between reading the file and truncating
the file for write, however, the race will not show up with a large
enough buffer on the tar side if buffered before the copy begins.
Also removes the unnecessary deferred removal, the removal is handled by
cleanup and respects the no cleanup env.
Signed-off-by: Derek McGowan <derek@mcg.dev>
If STDOUT or STDERR are attached and the container exits, the streams
will be closed by the daemon while the container is exiting, causing
the streamer to return an error
61b02e636d/cli/command/container/hijack.go (L53)
that gets sent
61b02e636d/cli/command/container/run.go (L278)
and received
61b02e636d/cli/command/container/run.go (L225)
on `errCh`.
However, if only STDIN is attached, it's not closed (since this is
attached to the user's TTY) when the container exits, so the streamer
doesn't exit and nothing gets sent on `errCh`, meaning the CLI execution
hangs receiving on `errCh` on L231.
Change the logic to receive on both `errCh` and `statusChan` – this way,
if the container exits, we get notified on `statusChan` (even if only
STDIN is attached), and can cancel the streamer and exit.
Signed-off-by: Laura Brehm <laurabrehm@hey.com>
Now, if running in "detached" mode, we early exit at L222.
Similarly, if `attachContainer` errors out, it returns an error that
gets handled on L190.
As such, `errCh` can never be nil on L231. Remove the nil check.
Signed-off-by: Laura Brehm <laurabrehm@hey.com>
Since everything else after the `apiClient.ContainerStart` block is
under an `if attach` conditional, we can move the "detached" early exit
up.
Signed-off-by: Laura Brehm <laurabrehm@hey.com>
