Golang uses a `sync.Once` when determining the proxy to use. This means
that it's not possible to test the proxy configuration in unit tests,
because the proxy configuration will be "fixated" the first time Golang
detects the proxy configuration.
This patch changes TestNewAPIClientFromFlagsWithHttpProxyEnv to an e2e
test so that we can verify the CLI picks up the proxy configuration.
Signed-off-by: Sebastiaan van Stijn <github@gone.nl>
(cherry picked from commit 40c6b117e7)
Signed-off-by: Sebastiaan van Stijn <github@gone.nl>
Some tests were using domain names that were intended to be "fake", but are
actually registered domain names (such as mycorp.com).
Even though we were not actually making connections to these domains, it's
better to use domains that are designated for testing/examples in RFC2606:
https://tools.ietf.org/html/rfc2606
Signed-off-by: Sebastiaan van Stijn <github@gone.nl>
(cherry picked from commit f3886f354a)
Signed-off-by: Sebastiaan van Stijn <github@gone.nl>
> Legacy PEM encryption as specified in RFC 1423 is insecure by design. Since
> it does not authenticate the ciphertext, it is vulnerable to padding oracle
> attacks that can let an attacker recover the plaintext
From https://go-review.googlesource.com/c/go/+/264159
> It's unfortunate that we don't implement PKCS#8 encryption so we can't
> recommend an alternative but PEM encryption is so broken that it's worth
> deprecating outright.
Signed-off-by: Sebastiaan van Stijn <github@gone.nl>
(cherry picked from commit 15535d4594)
Signed-off-by: Sebastiaan van Stijn <github@gone.nl>
This extends #2929 to Darwin as well as Linux.
Running the example in https://github.com/golang/go/issues/37942
I see lots of:
```
dave@m1 sigurg % uname -ms
Darwin arm64
dave@m1 sigurg % go run main.go
received urgent I/O condition: 2021-05-21 16:03:03.482211 +0100 BST m=+0.014553751
received urgent I/O condition: 2021-05-21 16:03:03.507171 +0100 BST m=+0.039514459
```
Signed-off-by: David Scott <dave@recoil.org>
(cherry picked from commit cedaf44ea2)
Signed-off-by: Sebastiaan van Stijn <github@gone.nl>
Commit fff164c22e modified ForwardAllSignals to
take `SIGURG` signals into account, which can be generated by the Go runtime
on Go 1.14 and up as an interrupt to support pre-emptable system calls on Linux.
With the updated code, the signal (`s`) would sometimes be `nil`, causing spurious
(but otherwise harmless) warnings to be printed;
Unsupported signal: <nil>. Discarding.
To debug this issue, I patched v20.10.4 to handle `nil`, and added a debug line
to print the signal in all cases;
```patch
diff --git a/cli/command/container/signals.go b/cli/command/container/signals.go
index 06e4d9eb6..0cb53ef06 100644
--- a/cli/command/container/signals.go
+++ b/cli/command/container/signals.go
@@ -22,8 +22,9 @@ func ForwardAllSignals(ctx context.Context, cli command.Cli, cid string, sigc <-
case <-ctx.Done():
return
}
+ fmt.Fprintf(cli.Err(), "Signal: %v\n", s)
if s == signal.SIGCHLD || s == signal.SIGPIPE {
```
When running a cross-compiled macOS binary with Go 1.13 (`make -f docker.Makefile binary-osx`):
# regular "docker run" (note that the `<nil>` signal only happens "sometimes"):
./build/docker run --rm alpine/git clone https://github.com/docker/getting-started.git
Cloning into 'getting-started'...
Signal: <nil>
# when cancelling with CTRL-C:
./build/docker run --rm alpine/git clone https://github.com/docker/getting-started.git
^CSignal: interrupt
Cloning into 'getting-started'...
error: could not lock config file /git/getting-started/.git/config: No such file or directory
fatal: could not set 'core.repositoryformatversion' to '0'
Signal: <nil>
Signal: <nil>
When running a macOS binary built with Go 1.15 (`DISABLE_WARN_OUTSIDE_CONTAINER=1 make binary`):
# regular "docker run" (note that the `<nil>` signal only happens "sometimes"):
# this is the same as on Go 1.13
./build/docker run --rm alpine/git clone https://github.com/docker/getting-started.git
Cloning into 'getting-started'...
Signal: <nil>
# when cancelling with CTRL-C:
./build/docker run --rm alpine/git clone https://github.com/docker/getting-started.git
Cloning into 'getting-started'...
^CSignal: interrupt
Signal: urgent I/O condition
Signal: urgent I/O condition
fatal: --stdin requires a git repository
fatal: index-pack failed
Signal: <nil>
Signal: <nil>
This patch checks if the channel is closed, and removes the warning (to prevent warnings if new
signals are added that are not in our known list of signals)
We should also consider updating `notfiyAllSignals()`, which currently forwards
_all_ signals (`signal.Notify(sigc)` without passing a list of signals), and
instead pass it "all signals _minus_ the signals we don't want forwarded":
35f023a7c2/cli/command/container/signals.go (L55)
Signed-off-by: Sebastiaan van Stijn <github@gone.nl>
(cherry picked from commit 9342ec6b71)
Signed-off-by: Sebastiaan van Stijn <github@gone.nl>
We refactorted `ForwardAllSignals` so it blocks but did not update the
call in `start` to call it in a goroutine.
Signed-off-by: Brian Goff <cpuguy83@gmail.com>
(cherry picked from commit e1a7517514)
Signed-off-by: Sebastiaan van Stijn <github@gone.nl>
The warning should be ignored on cgroup v2 hosts.
Relevant: 8086443a44
Signed-off-by: Akihiro Suda <akihiro.suda.cz@hco.ntt.co.jp>
(cherry picked from commit 05ec0188fa)
Signed-off-by: Sebastiaan van Stijn <github@gone.nl>
The kernel memory limit is deprecated in Docker 20.10.0,
and its support was removed in runc v1.0.0-rc94.
So, this warning can be safely removed.
Relevant: b8ca7de823
Signed-off-by: Akihiro Suda <akihiro.suda.cz@hco.ntt.co.jp>
(cherry picked from commit 731f52cfe8)
Signed-off-by: Sebastiaan van Stijn <github@gone.nl>
This reverts commit 3c87f01b18.
This commit introduced two regressions;
- spurious "Unsupported signal: <nil>. Discarding."
- docker start --attach hanging if the container does not
have a TTY attached
Reverting for now, while we dug deeper into what's causing
the regression.
Signed-off-by: Sebastiaan van Stijn <github@gone.nl>
Prior to this change, progressbars would sometimes be hidden, and the function
would return early. In addition, the direction of the progressbars would sometimes
be "incrementing" (similar to "docker service update"), and sometimes be "decrementing"
(to indicate a "rollback" is being performed).
This fix makes sure that we always proceed with the "verifying" step, and now
prints a message _after_ the verifying stage was completed;
$ docker service rollback foo
foo
overall progress: rolling back update: 5 out of 5 tasks
1/5: running [> ]
2/5: starting [===========> ]
3/5: starting [===========> ]
4/5: running [> ]
5/5: running [> ]
verify: Service converged
rollback: rollback completed
$ docker service rollback foo
foo
overall progress: rolling back update: 1 out of 1 tasks
1/1: running [> ]
verify: Service converged
rollback: rollback completed
Signed-off-by: Sebastiaan van Stijn <github@gone.nl>
(cherry picked from commit 104469be0b)
Signed-off-by: Sebastiaan van Stijn <github@gone.nl>
Before this change:
--------------------------------------------
$ docker service create --replicas=1 --name foo -p 8080:80 nginx:alpine
t33qvykv8y0zbz266rxynsbo3
overall progress: 1 out of 1 tasks
1/1: running [==================================================>]
verify: Service converged
$ echo $?
0
$ docker service update --replicas=5 foo
foo
overall progress: 5 out of 5 tasks
1/5: running [==================================================>]
2/5: running [==================================================>]
3/5: running [==================================================>]
4/5: running [==================================================>]
5/5: running [==================================================>]
verify: Service converged
$ echo $?
0
$ docker service rollback foo
foo
rollback: manually requested rollback
overall progress: rolling back update: 1 out of 1 tasks
1/1: running [> ]
verify: Service converged
$ echo $?
0
$ docker service rollback foo
foo
service rolled back: rollback completed
$ echo $?
1
After this change:
--------------------------------------------
$ docker service create --replicas=1 --name foo -p 8080:80 nginx:alpine
t33qvykv8y0zbz266rxynsbo3
overall progress: 1 out of 1 tasks
1/1: running [==================================================>]
verify: Service converged
$ echo $?
0
$ docker service update --replicas=5 foo
foo
overall progress: 5 out of 5 tasks
1/5: running [==================================================>]
2/5: running [==================================================>]
3/5: running [==================================================>]
4/5: running [==================================================>]
5/5: running [==================================================>]
verify: Waiting 1 seconds to verify that tasks are stable...
$ echo $?
0
$ docker service rollback foo
foo
rollback: manually requested rollback
overall progress: rolling back update: 1 out of 1 tasks
1/1: running [> ]
verify: Service converged
$ echo $?
0
$ docker service rollback foo
foo
service rolled back: rollback completed
$ echo $?
0
$ docker service ps foo
ID NAME IMAGE NODE DESIRED STATE CURRENT STATE ERROR PORTS
4dt4ms4c5qfb foo.1 nginx:alpine docker-desktop Running Running 2 minutes ago
Remaining issues with reconciliation
--------------------------------------------
Note that both before, and after this change, the command sometimes terminates
early, and does not wait for the service to reconcile; this is most apparent
when rolling back is scaling up (so more tasks are deployed);
$ docker service rollback foo
foo
service rolled back: rollback completed
$ docker service rollback foo
foo
rollback: manually requested rollback
overall progress: rolling back update: 1 out of 5 tasks
1/5: pending [=================================> ]
2/5: running [> ]
3/5: pending [=================================> ]
4/5: pending [=================================> ]
5/5: pending [=================================> ]
service rolled back: rollback completed
Signed-off-by: Sebastiaan van Stijn <github@gone.nl>
(cherry picked from commit ce26a165b0)
Signed-off-by: Sebastiaan van Stijn <github@gone.nl>
Commit f32731f902 fixed a potential panic
when an error was returned while trying to get existing credentials.
However, other code paths currently use the result of `GetDefaultAuthConfig()`
even in an error condition; this resulted in a panic, because a `nil` was
returned.
Signed-off-by: Sebastiaan van Stijn <github@gone.nl>
(cherry picked from commit c2820a7e3b)
Signed-off-by: Sebastiaan van Stijn <github@gone.nl>
On Windows, the os/exec.{Command,CommandContext,LookPath} functions
resolve command names that have neither path separators nor file extension
(e.g., "git") by first looking in the current working directory before
looking in the PATH environment variable.
Go maintainers intended to match cmd.exe's historical behavior.
However, this is pretty much never the intended behavior and as an abundance of precaution
this patch prevents that when executing commands.
Example of commands that docker.exe may execute: `git`, `docker-buildx` (or other cli plugin), `docker-credential-wincred`, `docker`.
Note that this was prompted by the [Go 1.15.7 security fixes](https://blog.golang.org/path-security), but unlike in `go.exe`,
the windows path lookups in docker are not in a code path allowing remote code execution, thus there is no security impact on docker.
Signed-off-by: Tibor Vass <tibor@docker.com>
(cherry picked from commit 8d199d5bba)
Signed-off-by: Tibor Vass <tibor@docker.com>
full diff: 75b288015a...c1f2f97bff
relevant changes:
- pkcs12: document that we use the wrong PEM type
- pkcs12: drop PKCS#12 attributes with unknown OIDs
- ocsp: Improve documentation for ParseResponse and ParseResponseForCert
other changes (not in vendor);
- ssh: improve error message for KeyboardInteractiveChallenge
- ssh: remove slow unnecessary diffie-hellman-group-exchange primality check
- ssh/terminal: replace with a golang.org/x/term wrapper
- Deprecates ssh/terminal in favor of golang.org/x/term
- ssh/terminal: add support for zos
- ssh/terminal: bump x/term dependency to fix js/nacl
- nacl/auth: use Size instead of KeySize for Sum output
- sha3: remove go:nocheckptr annotation
Signed-off-by: Sebastiaan van Stijn <github@gone.nl>
This hack was added in an attempt to continue supporting the experimental
(non-buildkit) `--platform` option, by dynamically updating the API version
required if buildkit isn't enabled.
This hack didn't work, however, because at the moment the override is
added, the command is not yet attached to the "root" (`docker`) command,
and because of that, the command itself is the `root` command;
`cmd.Root()` returned the `build` command.
As a result, validation steps defined as `PersistentPreRunE` on the root
command were not executed, causing invalid flags/options to not producing
an error.
Attempts to use an alternative approach (for example, cobra supports both
a `PersistentPreRun` and `PersistentPreRunE`) did not work either, because
`PersistentPreRunE` takes precedence over `PersistentPreRun`, and only one
will be executed.
Now that `--platform` should be supported for other cases than just for
experimental (LCOW), let's remove the 'experimental' check, and just assume
it's supported for API v1.32 and up.
Signed-off-by: Sebastiaan van Stijn <github@gone.nl>
While performance will be worse, we can safely ignore the --stream
option when used, and print a deprecation warning instead of failing
the build.
With this patch:
echo -e "FROM scratch\nLABEL foo=bar" | docker build --stream -
DEPRECATED: The experimental --stream flag has been removed and the build context
will be sent non-streaming. Enable BuildKit instead with DOCKER_BUILDKIT=1
to stream build context, see https://docs.docker.com/go/buildkit/
Sending build context to Docker daemon 2.048kB
Step 1/2 : FROM scratch
--->
Step 2/2 : LABEL foo=bar
---> Running in 99e4021085b6
Removing intermediate container 99e4021085b6
---> 1a7a41be241f
Successfully built 1a7a41be241f
Signed-off-by: Sebastiaan van Stijn <github@gone.nl>
The CLI disabled experimental features by default, requiring users
to set a configuration option to enable them.
Disabling experimental features was a request from Enterprise users
that did not want experimental features to be accessible.
We are changing this policy, and now enable experimental features
by default. Experimental features may still change and/or removed,
and will be highlighted in the documentation and "usage" output.
For example, the `docker manifest inspect --help` output now shows:
EXPERIMENTAL:
docker manifest inspect is an experimental feature.
Experimental features provide early access to product functionality. These features
may change between releases without warning or can be removed entirely from a future
release. Learn more about experimental features: https://docs.docker.com/go/experimental/
Signed-off-by: Sebastiaan van Stijn <github@gone.nl>
When initializing the API client, the User-Agent was added to any custom
HTTPHeaders that were configured. However, because the map was not properly
dereferenced, the original map was modified, causing the User-Agent to also
be saved to config.json after `docker login` and `docker logout`:
Before this change;
$ cat ~/.docker/config.json
cat: can't open '/root/.docker/config.json': No such file or directory
$ docker login -u myusername
Password:
...
Login Succeeded
$ cat ~/.docker/config.json
{
"auths": {
"https://index.docker.io/v1/": {
"auth": "<base64 auth>"
}
},
"HttpHeaders": {
"User-Agent": "Docker-Client/19.03.12 (linux)"
}
}
$ docker logout
{
"auths": {},
"HttpHeaders": {
"User-Agent": "Docker-Client/19.03.12 (linux)"
}
}
After this change:
$ cat ~/.docker/config.json
cat: can't open '/root/.docker/config.json': No such file or directory
$ docker login -u myusername
Password:
...
Login Succeeded
$ cat ~/.docker/config.json
{
"auths": {
"https://index.docker.io/v1/": {
"auth": "<base64 auth>"
}
}
}
$ docker logout
Removing login credentials for https://index.docker.io/v1/
$ cat ~/.docker/config.json
{
"auths": {}
}
Signed-off-by: Sebastiaan van Stijn <github@gone.nl>
This implements a special "RESET" value that can be used to reset the
list of capabilities to add/drop when updating a service.
Given the following service;
| CapDrop | CapAdd |
| -------------- | ------------- |
| CAP_SOME_CAP | |
When updating the service, and applying `--cap-drop RESET`, the "drop" list
is reset to its default:
| CapDrop | CapAdd |
| -------------- | ------------- |
| | |
When updating the service, and applying `--cap-drop RESET`, combined with
`--cap-add CAP_SOME_CAP` and `--cap-drop CAP_SOME_OTHER_CAP`:
| CapDrop | CapAdd |
| -------------- | ------------- |
| CAP_FOO_CAP | CAP_SOME_CAP |
Signed-off-by: Sebastiaan van Stijn <github@gone.nl>
Adding/removing capabilities when updating a service is considered a tri-state;
- if the capability was previously "dropped", then remove it from "CapabilityDrop",
but do NOT add it to "CapabilityAdd". However, if the capability was not yet in
the service's "CapabilityDrop", then simply add it to the service's "CapabilityAdd"
- likewise, if the capability was previously "added", then remove it from
"CapabilityAdd", but do NOT add it to "CapabilityDrop". If the capability was
not yet in the service's "CapabilityAdd", then simply add it to the service's
"CapabilityDrop".
In other words, given a service with the following:
| CapDrop | CapAdd |
| -------------- | ------------- |
| CAP_SOME_CAP | |
When updating the service, and applying `--cap-add CAP_SOME_CAP`, the previously
dropped capability is removed:
| CapDrop | CapAdd |
| -------------- | ------------- |
| | |
When updating the service a second time, applying `--cap-add CAP_SOME_CAP`,
capability is now added:
| CapDrop | CapAdd |
| -------------- | ------------- |
| | CAP_SOME_CAP |
Signed-off-by: Sebastiaan van Stijn <github@gone.nl>
When creating and updating services, we need to avoid unneeded service churn.
The interaction of separate lists to "add" and "drop" capabilities, a special
("ALL") capability, as well as a "relaxed" format for accepted capabilities
(case-insensitive, `CAP_` prefix optional) make this rather involved.
This patch updates how we handle `--cap-add` / `--cap-drop` when _creating_ as
well as _updating_, with the following rules/assumptions applied:
- both existing (service spec) and new (values passed through flags or in
the compose-file) are normalized and de-duplicated before use.
- the special "ALL" capability is equivalent to "all capabilities" and taken
into account when normalizing capabilities. Combining "ALL" capabilities
and other capabilities is therefore equivalent to just specifying "ALL".
- adding capabilities takes precedence over dropping, which means that if
a capability is both set to be "dropped" and to be "added", it is removed
from the list to "drop".
- the final lists should be sorted and normalized to reduce service churn
- no validation of capabilities is handled by the client. Validation is
delegated to the daemon/server.
When deploying a service using a docker-compose file, the docker-compose file
is *mostly* handled as being "declarative". However, many of the issues outlined
above also apply to compose-files, so similar handling is applied to compose
files as well to prevent service churn.
Signed-off-by: Sebastiaan van Stijn <github@gone.nl>
Sebastiaan van Stijn
Samuel Karp
Sebastiaan van Stijn
Mathieu Champlon
David Scott
Brian Goff
Akihiro Suda
Silvin Lubecki
Jim Lin
Alexey Igrychev
Tibor Vass
Chris Crone
Djordje Lukic
Tibor Vass
bn4t
Jorge Vallecillo
Jennings Zhang
Albin Kerouanton