The Container.State struct holds the container's state, and most of
its fields are expected to change dynamically. Some o these state-changes
are explicit, for example, setting the container to be "stopped". Other
state changes can be more explicit, for example due to the containers'
process exiting or being "OOM" killed by the kernel.
The distinction between explicit ("desired") state changes and "state"
("actual state") is sometimes vague; for some properties, we clearly
separated them, for example if a user requested the container to be
stopped or restarted, we store state in the Container object itself;
HasBeenManuallyStopped bool // used for unless-stopped restart policy
HasBeenManuallyRestarted bool `json:"-"` // used to distinguish restart caused by restart policy from the manual one
Other properties are more ambiguous. such as "HasBeenStartedBefore" and
"RestartCount", which are stored on the Container (and persisted to
disk), but may be more related to "actual" state, and likely should
not be persisted;
RestartCount int
HasBeenStartedBefore bool
Given that (per the above) concurrency must be taken into account, most
changes to the `container.State` struct should be protected; here's where
things get blurry. While the `State` type provides various accessor methods,
only some of them take concurrency into account; for example, [State.IsRunning]
and [State.GetPID] acquire a lock, whereas [State.ExitCodeValue] does not.
Even the (commonly used) [State.StateString] has no locking at all.
The way to handle this is error-prone; [container.State] contains a mutex,
and it's exported. Given that its embedded in the [container.Container]
struct, it's also exposed as an exported mutex for the container. The
assumption here is that by "merging" the two, the caller to acquire a lock
when either the container _or_ its state must be mutated. However, because
some methods on `container.State` handle their own locking, consumers must
be deeply familiar with the internals; if both changes to the `Container`
AND `Container.State` must be made. This gets amplified more as some
(exported!) methods, such as [container.SetRunning] mutate multiple fields,
but don't acquire a lock (so expect the caller to hold one), but their
(also exported) counterpart (e.g. [State.IsRunning]) do.
It should be clear from the above, that this needs some architectural
changes; a clearer separation between "desired" and "actual" state (opening
the potential to update the container's config without manually touching
its `State`), possibly a method to obtain a read-only copy of the current
state (for those querying state), and reviewing which fields belong where
(and should be persisted to disk, or only remain in memory).
This PR preserves the status quo; it makes no structural changes, other
than exposing where we access the container's state. Where previously the
State fields and methods were referred to as "part of the container"
(e.g. `ctr.IsRunning()` or `ctr.Running`), we now explicitly reference
the embedded `State` (`ctr.State.IsRunning`, `ctr.State.Running`).
The exception (for now) is the mutex, which is still referenced through
the embedded struct (`ctr.Lock()` instead of `ctr.State.Lock()`), as this
is (mostly) by design to protect the container, and what's in it (including
its `State`).
[State.IsRunning]: c4afa77157/daemon/container/state.go (L205-L209)
[State.GetPID]: c4afa77157/daemon/container/state.go (L211-L216)
[State.ExitCodeValue]: c4afa77157/daemon/container/state.go (L218-L228)
[State.StateString]: c4afa77157/daemon/container/state.go (L102-L131)
[container.State]: c4afa77157/daemon/container/state.go (L15-L23)
[container.Container]: c4afa77157/daemon/container/container.go (L67-L75)
[container.SetRunning]: c4afa77157/daemon/container/state.go (L230-L277)
Signed-off-by: Sebastiaan van Stijn <github@gone.nl>
The "backend" types in API were designed to decouple the API server
implementation from the daemon, or other parts of the code that
back the API server. This would allow the daemon to evolve (e.g.
functionality moved to different subsystems) without that impacting
the API server's implementation.
Now that the API server is no longer part of the API package (module),
there is no benefit to having it in the API module. The API server
may evolve (and require changes in the backend), which has no direct
relation with the API module (types, responses); the backend definition
is, however, coupled to the API server implementation.
It's worth noting that, while "technically" possible to use the API
server package, and implement an alternative backend implementation,
this has never been a prime objective. The backend definition was
never considered "stable", and we don't expect external users to
(attempt) to use it as such.
This patch moves the backend types to the daemon/server package,
so that they can evolve with the daemon and API server implementation
without that impacting the API module (which we intend to be stable,
following SemVer).
Signed-off-by: Sebastiaan van Stijn <github@gone.nl>
The stdcopy package is used to produce and read multiplexed streams for
"attach" and "logs". It is used both by the API server (to produce), and
the client (to read / de-multiplex).
Move it to the api package, so that it can be included in the api module.
Signed-off-by: Sebastiaan van Stijn <github@gone.nl>
These comments were added to enforce using the correct import path for
our packages ("github.com/docker/docker", not "github.com/moby/moby").
However, when working in go module mode (not GOPATH / vendor), they have
no effect, so their impact is limited.
Remove these imports in preparation of migrating our code to become an
actual go module.
Signed-off-by: Sebastiaan van Stijn <github@gone.nl>
Alias and deprecate the status types and constants from the root
container package. The root container package is intended for use
within the daemon and no the api package.
Signed-off-by: Derek McGowan <derek@mcg.dev>
Both these defers were defined in the closure, and would be executed
at the end of the goroutine; inline them to prevent them being confused
for being executed after the function starting the goroutine.
Signed-off-by: Sebastiaan van Stijn <github@gone.nl>
- rename var that shadowed argument
- use consistent name for container arg; We use `ctr` as short name in most
places; use the same name to make the code slightly more readable.
- rename "logs" argument to prevent it being confused with a logs import
and to more clearly indicate its a bool
Signed-off-by: Sebastiaan van Stijn <github@gone.nl>
This plumbs a context down the stack and handles cancellation as needed
so that we can have correlated traces from the API.
Signed-off-by: Brian Goff <cpuguy83@gmail.com>
This var for the incoming attach request.
Just within this one function we also have `cfg`, and `ctr` already, so
`c` just makes things more confusing.
Not to mention `c` is usually referencing a container object in other
parts of the code.
Signed-off-by: Brian Goff <cpuguy83@gmail.com>
In cases where the client disconnects and there is nothing to read from
a stdio stream after that disconnect, the copy goroutines and file
descriptors are leaked because `io.Copy` is just blocked waiting for
data from the container's I/O stream.
This fix only applies to Linux.
Windows will need a separate fix.
Signed-off-by: Brian Goff <cpuguy83@gmail.com>
Move the initialization logic to the attachContext itself, so that
the container doesn't have to be aware about mutexes and other logic.
Signed-off-by: Sebastiaan van Stijn <github@gone.nl>
The github.com/containerd/containerd/log package was moved to a separate
module, which will also be used by upcoming (patch) releases of containerd.
This patch moves our own uses of the package to use the new module.
Signed-off-by: Sebastiaan van Stijn <github@gone.nl>
Define consts for the Actions we use for events, instead of "ad-hoc" strings.
Having these consts makes it easier to find where specific events are triggered,
makes the events less error-prone, and allows documenting each Action (if needed).
Signed-off-by: Sebastiaan van Stijn <github@gone.nl>
When daemon.ContainerLogs() is called with options.follow=true
(as in "docker logs --follow"), the "loggerutils.followLogs()"
function never returns (even then the logs consumer is gone).
As a result, all the resources associated with it (including
an opened file descriptor for the log file being read, two FDs
for a pipe, and two FDs for inotify watch) are never released.
If this is repeated (such as by running "docker logs --follow"
and pressing Ctrl-C a few times), this results in DoS caused by
either hitting the limit of inotify watches, or the limit of
opened files. The only cure is daemon restart.
Apparently, what happens is:
1. logs producer (a container) is gone, calling (*LogWatcher).Close()
for all its readers (daemon/logger/jsonfilelog/jsonfilelog.go:175).
2. WatchClose() is properly handled by a dedicated goroutine in
followLogs(), cancelling the context.
3. Upon receiving the ctx.Done(), the code in followLogs()
(daemon/logger/loggerutils/logfile.go#L626-L638) keeps to
send messages _synchronously_ (which is OK for now).
4. Logs consumer is gone (Ctrl-C is pressed on a terminal running
"docker logs --follow"). Method (*LogWatcher).Close() is properly
called (see daemon/logs.go:114). Since it was called before and
due to to once.Do(), nothing happens (which is kinda good, as
otherwise it will panic on closing a closed channel).
5. A goroutine (see item 3 above) keeps sending log messages
synchronously to the logWatcher.Msg channel. Since the
channel reader is gone, the channel send operation blocks forever,
and resource cleanup set up in defer statements at the beginning
of followLogs() never happens.
Alas, the fix is somewhat complicated:
1. Distinguish between close from logs producer and logs consumer.
To that effect,
- yet another channel is added to LogWatcher();
- {Watch,}Close() are renamed to {Watch,}ProducerGone();
- {Watch,}ConsumerGone() are added;
*NOTE* that ProducerGone()/WatchProducerGone() pair is ONLY needed
in order to stop ConsumerLogs(follow=true) when a container is stopped;
otherwise we're not interested in it. In other words, we're only
using it in followLogs().
2. Code that was doing (logWatcher*).Close() is modified to either call
ProducerGone() or ConsumerGone(), depending on the context.
3. Code that was waiting for WatchClose() is modified to wait for
either ConsumerGone() or ProducerGone(), or both, depending on the
context.
4. followLogs() are modified accordingly:
- context cancellation is happening on WatchProducerGone(),
and once it's received the FileWatcher is closed and waitRead()
returns errDone on EOF (i.e. log rotation handling logic is disabled);
- due to this, code that was writing synchronously to logWatcher.Msg
can be and is removed as the code above it handles this case;
- function returns once ConsumerGone is received, freeing all the
resources -- this is the bugfix itself.
While at it,
1. Let's also remove the ctx usage to simplify the code a bit.
It was introduced by commit a69a59ffc7 ("Decouple removing the
fileWatcher from reading") in order to fix a bug. The bug was actually
a deadlock in fsnotify, and the fix was just a workaround. Since then
the fsnofify bug has been fixed, and a new fsnotify was vendored in.
For more details, please see
https://github.com/moby/moby/pull/27782#issuecomment-416794490
2. Since `(*filePoller).Close()` is fixed to remove all the files
being watched, there is no need to explicitly call
fileWatcher.Remove(name) anymore, so get rid of the extra code.
Should fix https://github.com/moby/moby/issues/37391
Signed-off-by: Kir Kolyshkin <kolyshkin@gmail.com>
With a full attach, each attach was leaking 4 goroutines.
This updates attach to use errgroup instead of the hodge-podge of
waitgroups and channels.
In addition, the detach event was never being sent.
Signed-off-by: Brian Goff <cpuguy83@gmail.com>
Instead of having to create a bunch of custom error types that are doing
nothing but wrapping another error in sub-packages, use a common helper
to create errors of the requested type.
e.g. instead of re-implementing this over and over:
```go
type notFoundError struct {
cause error
}
func(e notFoundError) Error() string {
return e.cause.Error()
}
func(e notFoundError) NotFound() {}
func(e notFoundError) Cause() error {
return e.cause
}
```
Packages can instead just do:
```
errdefs.NotFound(err)
```
Signed-off-by: Brian Goff <cpuguy83@gmail.com>
Use strongly typed errors to set HTTP status codes.
Error interfaces are defined in the api/errors package and errors
returned from controllers are checked against these interfaces.
Errors can be wraeped in a pkg/errors.Causer, as long as somewhere in the
line of causes one of the interfaces is implemented. The special error
interfaces take precedence over Causer, meaning if both Causer and one
of the new error interfaces are implemented, the Causer is not
traversed.
Signed-off-by: Brian Goff <cpuguy83@gmail.com>
This patch adds the untilRemoved option to the ContainerWait API which
allows the client to wait until the container is not only exited but
also removed.
This patch also adds some more CLI integration tests for waiting for a
created container and waiting with the new --until-removed flag.
Docker-DCO-1.1-Signed-off-by: Josh Hawn <josh.hawn@docker.com> (github: jlhawn)
Handle detach sequence in CLI
Docker-DCO-1.1-Signed-off-by: Josh Hawn <josh.hawn@docker.com> (github: jlhawn)
Update Container Wait Conditions
Docker-DCO-1.1-Signed-off-by: Josh Hawn <josh.hawn@docker.com> (github: jlhawn)
Apply container wait changes to API 1.30
The set of changes to the containerWait API missed the cut for the
Docker 17.05 release (API version 1.29). This patch bumps the version
checks to use 1.30 instead.
This patch also makes a minor update to a testfile which was added to
the builder/dockerfile package.
Docker-DCO-1.1-Signed-off-by: Josh Hawn <josh.hawn@docker.com> (github: jlhawn)
Remove wait changes from CLI
Docker-DCO-1.1-Signed-off-by: Josh Hawn <josh.hawn@docker.com> (github: jlhawn)
Address minor nits on wait changes
- Changed the name of the tty Proxy wrapper to `escapeProxy`
- Removed the unnecessary Error() method on container.State
- Fixes a typo in comment (repeated word)
Docker-DCO-1.1-Signed-off-by: Josh Hawn <josh.hawn@docker.com> (github: jlhawn)
Use router.WithCancel in the containerWait handler
This handler previously added this functionality manually but now uses
the existing wrapper which does it for us.
Docker-DCO-1.1-Signed-off-by: Josh Hawn <josh.hawn@docker.com> (github: jlhawn)
Add WaitCondition constants to api/types/container
Docker-DCO-1.1-Signed-off-by: Josh Hawn <josh.hawn@docker.com> (github: jlhawn)
Address more ContainerWait review comments
- Update ContainerWait backend interface to not return pointer values
for container.StateStatus type.
- Updated container state's Wait() method comments to clarify that a
context MUST be used for cancelling the request, setting timeouts,
and to avoid goroutine leaks.
- Removed unnecessary buffering when making channels in the client's
ContainerWait methods.
- Renamed result and error channels in client's ContainerWait methods
to clarify that only a single result or error value would be sent
on the channel.
Docker-DCO-1.1-Signed-off-by: Josh Hawn <josh.hawn@docker.com> (github: jlhawn)
Move container.WaitCondition type to separate file
... to avoid conflict with swagger-generated code for API response
Docker-DCO-1.1-Signed-off-by: Josh Hawn <josh.hawn@docker.com> (github: jlhawn)
Address more ContainerWait review comments
Docker-DCO-1.1-Signed-off-by: Josh Hawn <josh.hawn@docker.com> (github: jlhawn)
This patch consolidates the two WaitStop and WaitWithContext methods
on the container.State type. Now there is a single method, Wait, which
takes a context and a bool specifying whether to wait for not just a
container exit but also removal.
The behavior has been changed slightly so that a wait call during a
Created state will not return immediately but instead wait for the
container to be started and then exited.
The interface has been changed to no longer block, but instead returns
a channel on which the caller can receive a *StateStatus value which
indicates the ExitCode or an error if there was one (like a context
timeout or state transition error).
These changes have been propagated through the rest of the deamon to
preserve all other existing behavior.
Docker-DCO-1.1-Signed-off-by: Josh Hawn <josh.hawn@docker.com> (github: jlhawn)
This cleans up attach a little bit, and moves it out of the container
package.
Really `AttachStream` is a method on `*stream.Config`, so moved if from
a package level function to one bound to `Config`.
In addition, uses a config struct rather than passing around tons and
tons of arguments.
Signed-off-by: Brian Goff <cpuguy83@gmail.com>
This fix tries to fix logrus formatting by removing `f` from
`logrus.[Error|Warn|Debug|Fatal|Panic|Info]f` when formatting string
is not present.
This fix fixes#23459.
Signed-off-by: Yong Tang <yong.tang.github@outlook.com>
e.g.
```
$ docker run -i --restart always busybox sh
pwd
/
exit 11
<...hang...>
```
This is because Attach(daemon side) and Run(client side) both hangs on
WaitStop, if container is restarted too quickly, wait won't have chance
to get exit signal.
Signed-off-by: Zhang Wei <zhangwei555@huawei.com>
If we attach to a running container and stream is closed afterwards, we
can never be sure if the container is stopped or detached. Adding a new
type of `detach` event can explicitly notify client that container is
detached, so client will know that there's no need to wait for its exit
code and it can move forward to next step now.
Signed-off-by: Zhang Wei <zhangwei555@huawei.com>
