Retries after v1 fallbacks were added in #20411. The test still appears
to be flaky. There are two potential problems. The initial pull was not
protected against pulling from v1, so it could be giving us a different
hello-world image to compare against. Also, after experiencing a v1
fallback, we need to restore the original image before doing the next
pull, because otherwise the "Image is up to date for hello-world:latest"
message will not show up as expected.
See #17214.
Signed-off-by: Aaron Lehmann <aaron.lehmann@docker.com>
Further differentiate the APIEndpoint used with V2 with the endpoint type which is only used for v1 registry interactions
Rename Endpoint to V1Endpoint and remove version ambiguity
Use distribution token handler for login
Signed-off-by: Derek McGowan <derek@mcgstyle.net>
Signed-off-by: Aaron Lehmann <aaron.lehmann@docker.com>
Concurrent uploads which share layers worked correctly as of #18353,
but unfortunately #18785 caused a regression. This PR removed the logic
that shares digests between different push sessions. This overlooked the
case where one session was waiting for another session to upload a
layer.
This commit adds back the ability to propagate this digest information,
using the distribution.Descriptor type because this is what is received
from stats and uploads, and also what is ultimately needed for building
the manifest.
Surprisingly, there was no test covering this case. This commit adds
one. It fails without the fix.
See recent comments on #9132.
Signed-off-by: Aaron Lehmann <aaron.lehmann@docker.com>
This test is often failing on remote daemons. We tried many approaches
to fix it but none worked. In order to make the CI more reliable, this
will skip the test when running against a remote daemon (e.g. win2lin).
Signed-off-by: Tibor Vass <tibor@docker.com>
I noticied an inconsistency when reviewing docker/pull/20692.
Changing Ip to IP and Nf to NF.
More info: The golang folks recommend that you keep the initials consistent:
https://github.com/golang/go/wiki/CodeReviewComments#initialisms.
Signed-off-by: Christy Perez <christy@linux.vnet.ibm.com>
* Remvoe integration test of TestDaemonStartWithDaemonCommand
* Rewrite as unit test
Related issue #19425
Signed-off-by: Wen Cheng Ma <wenchma@cn.ibm.com>
This removes the email prompt when you use docker login, and also removes the ability to register via the docker cli. Docker login, will strictly be used for logging into a registry server.
Signed-off-by: Ken Cochrane <kencochrane@gmail.com>
This change implements communication with an external credentials store,
ala git-credential-helper. The client falls back the plain text store,
what we're currently using, if there is no remote store configured.
It shells out to helper program when a credential store is
configured. Those programs can be implemented with any language as long as they
follow the convention to pass arguments and information.
There is an implementation for the OS X keychain in https://github.com/calavera/docker-credential-helpers.
That package also provides basic structure to create other helpers.
Signed-off-by: David Calavera <david.calavera@gmail.com>
During "COPY" or other tar unpack operations, a target/destination
parent dir might not exist and should be created with ownership of the
root in the right context (including remapped root when user namespaces
are enabled)
Docker-DCO-1.1-Signed-off-by: Phil Estes <estesp@linux.vnet.ibm.com> (github: estesp)
Moving all strings to the errors package wasn't a good idea after all.
Our custom implementation of Go errors predates everything that's nice
and good about working with errors in Go. Take as an example what we
have to do to get an error message:
```go
func GetErrorMessage(err error) string {
switch err.(type) {
case errcode.Error:
e, _ := err.(errcode.Error)
return e.Message
case errcode.ErrorCode:
ec, _ := err.(errcode.ErrorCode)
return ec.Message()
default:
return err.Error()
}
}
```
This goes against every good practice for Go development. The language already provides a simple, intuitive and standard way to get error messages, that is calling the `Error()` method from an error. Reinventing the error interface is a mistake.
Our custom implementation also makes very hard to reason about errors, another nice thing about Go. I found several (>10) error declarations that we don't use anywhere. This is a clear sign about how little we know about the errors we return. I also found several error usages where the number of arguments was different than the parameters declared in the error, another clear example of how difficult is to reason about errors.
Moreover, our custom implementation didn't really make easier for people to return custom HTTP status code depending on the errors. Again, it's hard to reason about when to set custom codes and how. Take an example what we have to do to extract the message and status code from an error before returning a response from the API:
```go
switch err.(type) {
case errcode.ErrorCode:
daError, _ := err.(errcode.ErrorCode)
statusCode = daError.Descriptor().HTTPStatusCode
errMsg = daError.Message()
case errcode.Error:
// For reference, if you're looking for a particular error
// then you can do something like :
// import ( derr "github.com/docker/docker/errors" )
// if daError.ErrorCode() == derr.ErrorCodeNoSuchContainer { ... }
daError, _ := err.(errcode.Error)
statusCode = daError.ErrorCode().Descriptor().HTTPStatusCode
errMsg = daError.Message
default:
// This part of will be removed once we've
// converted everything over to use the errcode package
// FIXME: this is brittle and should not be necessary.
// If we need to differentiate between different possible error types,
// we should create appropriate error types with clearly defined meaning
errStr := strings.ToLower(err.Error())
for keyword, status := range map[string]int{
"not found": http.StatusNotFound,
"no such": http.StatusNotFound,
"bad parameter": http.StatusBadRequest,
"conflict": http.StatusConflict,
"impossible": http.StatusNotAcceptable,
"wrong login/password": http.StatusUnauthorized,
"hasn't been activated": http.StatusForbidden,
} {
if strings.Contains(errStr, keyword) {
statusCode = status
break
}
}
}
```
You can notice two things in that code:
1. We have to explain how errors work, because our implementation goes against how easy to use Go errors are.
2. At no moment we arrived to remove that `switch` statement that was the original reason to use our custom implementation.
This change removes all our status errors from the errors package and puts them back in their specific contexts.
IT puts the messages back with their contexts. That way, we know right away when errors used and how to generate their messages.
It uses custom interfaces to reason about errors. Errors that need to response with a custom status code MUST implementent this simple interface:
```go
type errorWithStatus interface {
HTTPErrorStatusCode() int
}
```
This interface is very straightforward to implement. It also preserves Go errors real behavior, getting the message is as simple as using the `Error()` method.
I included helper functions to generate errors that use custom status code in `errors/errors.go`.
By doing this, we remove the hard dependency we have eeverywhere to our custom errors package. Yes, you can use it as a helper to generate error, but it's still very easy to generate errors without it.
Please, read this fantastic blog post about errors in Go: http://dave.cheney.net/2014/12/24/inspecting-errors
Signed-off-by: David Calavera <david.calavera@gmail.com>
This will allow us to have a windows-to-linux CI, where the linux host
can be anywhere, connecting with TLS.
Signed-off-by: Tibor Vass <tibor@docker.com>
This enhancement is to fix the wrong list results on
`docker ps` before and since filters specifying the non-running container.
Fixes issue #20431
Signed-off-by: Wen Cheng Ma <wenchma@cn.ibm.com>
Attempt layer mounts from up to 3 source repositories, possibly
falling back to a standard blob upload for cross repository pushes.
Addresses compatiblity issues with token servers which do not grant
multiple repository scopes, resulting in an authentication failure for
layer mounts, which would otherwise cause the push to terminate with an
error.
Signed-off-by: Brian Bland <brian.bland@docker.com>
In cases where the a plugin responds with both a null or empty volume
and a null or empty Err, the daemon would panic.
This is because we assumed the idiom if `err` is nil, then `v` must not
be but in reality the plugin may return whatever it wants and we want to
make sure it doesn't harm the daemon.
Signed-off-by: Brian Goff <cpuguy83@gmail.com>
If cgroup swap memory limit isn't enabled, then
the -m flag doesn't work and the container that is created for
both of these tests is very large. Because we are trying to run the
containers out of memory, this takes a very long time and causes the
tests to fail most of the time.
Follow-up to #17913
Signed-off-by: Christopher Jones <tophj@linux.vnet.ibm.com>
The container started with `-d` as part of the test requires a `waitRun`
to ensure it is actually running before attempting any other operation.
Signed-off-by: Arnaud Porterie <arnaud.porterie@docker.com>
- Allow to filter containers by volume with `--filter volume=name` and `filter volume=/dest`.
- Show their names in the list with the custom format `{{ .Mounts }}`.
Signed-off-by: David Calavera <david.calavera@gmail.com>
Now that the various fixes are all committed, let's see if this gets
less flaky now.
Docker-DCO-1.1-Signed-off-by: Phil Estes <estesp@linux.vnet.ibm.com> (github: estesp)
Add `--restart` flag for `update` command, so we can change restart
policy for a container no matter it's running or stopped.
Signed-off-by: Zhang Wei <zhangwei555@huawei.com>
The restarts in the authz plugin test suite seems to be causing
flakiness in CI, and can be avoided by separating the daemon start and
busybox image load.
Docker-DCO-1.1-Signed-off-by: Phil Estes <estesp@linux.vnet.ibm.com> (github: estesp)
- Order the flow of the handlers more cleanly--read req, do actions,
write response.
- Add "always allowed" endpoints to handle `/_ping` and `/info` usage
from the test framework/daemon start/restart management
Docker-DCO-1.1-Signed-off-by: Phil Estes <estesp@linux.vnet.ibm.com> (github: estesp)
This corrects `docker cp` behavior when user namespaces are enabled.
Instead of chown'ing copied-in files to real root (0,0), the code
queries for the remapped root uid & gid and sets the chown option
properly.
Docker-DCO-1.1-Signed-off-by: Phil Estes <estesp@linux.vnet.ibm.com> (github: estesp)
Since channel is getting a send instead of a close now, this can cause
random issues ranging through the list of channels if the channel is
unbuffered since the send may be blocked.
Signed-off-by: Brian Goff <cpuguy83@gmail.com>
Sometimes transient network issues will cause
TestPullFromCentralRegistryImplicitRefParts to end up pulling with the
v1 protocol. This violates the assumptions behind the test. To make the
test more robust, allow a few retries if any pull ends up using the v1
protocol.
Fixes#17214
Signed-off-by: Aaron Lehmann <aaron.lehmann@docker.com>
Fixes: #20328
We sort network ls output with incresing order,
it may make output more easy to consume for users.
Signed-off-by: Kai Qiang Wu(Kennan) <wkqwu@cn.ibm.com>
In Docker 1.10 and earlier, "docker build" can do a build FROM a private
repository that hasn't yet been pulled. This doesn't work on master. I
bisected this to https://github.com/docker/docker/pull/19414.
AuthConfigs is deserialized from the HTTP request, but not included in
the builder options.
Signed-off-by: Aaron Lehmann <aaron.lehmann@docker.com>
Progress toward being able to run integration-cli campaign using a
client hitting a remote host.
Most of these fixes imply flagging tests that assume they are running on
the same host than the Daemon. Also fixes the `contrib/httpserver` image
that couldn't run because of a dynamically linked Go binary inside the
busybox image.
Signed-off-by: Arnaud Porterie <arnaud.porterie@docker.com>
This fixes an issue where `docker run -v foo:/bar --volume-driver
<remote driver>` -> daemon restart -> `docker run -v foo:/bar` would
make a `local` volume after the restart instead of using the existing
volume from the remote driver.
Signed-off-by: Brian Goff <cpuguy83@gmail.com>
This PR makes restores the pre-Docker 1.10 behavior of allowing
uppercase characters in registry hostnames.
Note that this only applies to hostnames, not remote image names.
Previous versions also prohibited uppercase letters after the hostname,
but Docker 1.10 extended this to the hostname itself.
- Vendor updated docker/distribution.
- Add a check to "normalize" that rejects remote names with uppercase
letters.
- Add test cases to TestTagValidPrefixedRepo and
TestTagInvalidUnprefixedRepo
Fixes: #20056
Signed-off-by: Aaron Lehmann <aaron.lehmann@docker.com>
This removes two tests on ppc64le.
There is an old bug with a syscall on power #8653, that causes
logrus to default to using logfmt. These two tests look for
logrus format specific strings, and fail if they don't see it.
Signed-off-by: Christopher Jones <tophj@linux.vnet.ibm.com>
Since we now automatically mount the mqueue device inside the
container (instead of bind mounting the one from the host), when
trying to start a container with --ipc=host, the mount will fail with
EPERM.
Signed-off-by: Kenfe-Mickael Laventure <mickael.laventure@gmail.com>
Currently, when running a container with --ipc=host, if /dev/mqueue is
a standard directory on the hos the daemon will bind mount it allowing
the container to create/modify files on the host.
This commit forces /dev/mqueue to always be of type mqueue except when
the user explicitely requested something to be bind mounted to
/dev/mqueue.
Signed-off-by: Kenfe-Mickael Laventure <mickael.laventure@gmail.com>
Filters should not include stopped container if `-a` is not specified.
Right now, before and since filter are acting as --before and --since
deprecated flags. This commit is fixing that.
Signed-off-by: Vincent Demeester <vincent@sbr.pm>
To support the requirement of blocking the request after the daemon
responded the authorization plugin use a `response recorder` that replay
the response after the flow ends.
This commit adds support for commands that hijack the connection and
flushes data via the http.Flusher interface. This resolves the error
with the event endpoint.
Signed-off-by: Liron Levin <liron@twistlock.com>
This came in with the Windows CI work and I assume was meant for local
debug, but adds a bare printf line to the test output.
Docker-DCO-1.1-Signed-off-by: Phil Estes <estesp@linux.vnet.ibm.com> (github: estesp)
Currently some commands including `kill`, `pause`, `restart`, `rm`,
`rmi`, `stop`, `unpause`, `udpate`, `wait` will print a lot of error
message on client side, with a lot of redundant messages, this commit is
trying to remove the unuseful and redundant information for user.
Signed-off-by: Zhang Wei <zhangwei555@huawei.com>
This is add support for #19559
We tried sort it in client side, and it sort follow go
sort : sorts a slice of strings in increasing order.
Signed-off-by: Kai Qiang Wu(Kennan) <wkqwu@cn.ibm.com>
This PR moves the testRequires check from within setupRegistry
and setupNotary to when the Suite itself starts up.
This fixes a bug where testRequires() in setupRegistry() would
succeed and start registry, but testRequires() in setupNotary()
would fail. This immediately exits the goroutine with registry
still running.
Signed-off-by: Christopher Jones <tophj@linux.vnet.ibm.com>
1. Replace raw `docker inspect -f xxx` with `inspectField`, to make code
cleaner and more consistent
2. assert the error in function `inspectField*` so we don't need to
assert the return value of it every time, this will make inspect easier.
Signed-off-by: Zhang Wei <zhangwei555@huawei.com>
Add `runSleepingContainer` and `runSleepingContainerInImage` helper
functions to factor out the way to run system-specific idle containers.
Define a sleeping container as command `top` in image `busybox` for
Unix and as command `sleep 60` in image `busybox` for Windows. Provide a
single point of code to update those.
Signed-off-by: Arnaud Porterie <arnaud.porterie@docker.com>
Vincent Demeester
Aaron Lehmann
David Calavera
Derek McGowan
Brian Goff
Tibor Vass
Christy Perez
Yong Tang
Wen Cheng Ma
Ken Cochrane
Sebastiaan van Stijn
Phil Estes
Qiang Huang
Antonio Murdaca
Shijiang Wei
Arnaud Porterie
Lei Jitang
Riyaz Faizullabhoy
Tibor Vass
John Howard
Brian Bland
Christopher Jones
Stefan Weil
Antonio Murdaca
Zhang Wei
Madhu Venugopal
Santhosh Manohar
Alexander Morozov
Kai Qiang Wu(Kennan)
Phil Estes
Michael Crosby
Tonis Tiigi
Alexander Morozov
Darren Stahl
Jess Frazelle
Morgan Bauer
Jessica Frazelle
Victor Vieux
Aidan Hobson Sayers
Stefan Scherer
Kenfe-Mickael Laventure
Shishir Mahajan
Liron Levin
Stephen Rust
Chun Chen
Alessandro Boch
Federico Gimenez
Doug Davis
cyli
Bryan Boreham