libcontainer does not guarantee a stable API, and is not intended
for external consumers.
this patch replaces some uses of libcontainer/cgroups with
containerd/cgroups.
Signed-off-by: Sebastiaan van Stijn <github@gone.nl>
Rather than bifurcate the test completely, this lets us keep the test
intact with a small function wrapper to allow the compiler to build the
code that'll never be called on Windows, on Windows.
Signed-off-by: Paul "TBBle" Hampson <Paul.Hampson@Pobox.com>
It turns out that the Finally block does not see the exit code from the
`exit` call that triggered it, but from an earlier state. And it seems
that actions take in the Finally block other than `exit` will not affect
the $LastErrorCode set by the `exit` that triggered the Finally block.
Signed-off-by: Paul "TBBle" Hampson <Paul.Hampson@Pobox.com>
Commit 6ca3ec88ae deprecated the experimental
"--stream" option for the legacy builder, adding an error message is a client
attempted to use this feature.
However, the detection used the session-ID (`session=xxx` query parameter),
which happens to be set automatically by the CLI if it detects that the daemon
has session support. Because of this, builds fail when trying to perform them
on a daemon with the `--experimental` flag set.
This patch changes the detection to look for the `remote` query parameter, which
is set to "client-session" when using the `--stream` option with the classic
(non-Buildkit) builder.
Before this change, running `docker build` with an older (19.03 or older) cli
against a daemon with `--experimental` enabled caused an error:
$ dockerd --experimental &
$ docker pull docker:18.09
$ docker run -it --rm -v /var/run/docker.sock:/var/run/docker.sock docker:18.09 sh -c 'echo "FROM scratch" | docker build -'
Sending build context to Docker daemon 2.048kB
Error response from daemon: experimental session with v1 builder is no longer supported, use builder version v2 (BuildKit) instead
docker run -it --rm -v /var/run/docker.sock:/var/run/docker.sock -w /foo docker:18.09 sh -c 'echo "FROM scratch" > Dockerfile && docker build --stream .'
Error response from daemon: experimental session with v1 builder is no longer supported, use builder version v2 (BuildKit) instead
With this patch, the error only occurs when trying to use the experimental
`--stream` option:
$ dockerd --experimental &
$ docker pull docker:18.09
$ docker run -it --rm -v /var/run/docker.sock:/var/run/docker.sock docker:18.09 sh -c 'echo "FROM scratch" | docker build -'
Step 1/1 : FROM scratch
--->
No image was generated. Is your Dockerfile empty?
$ docker run -it --rm -v /var/run/docker.sock:/var/run/docker.sock -w /foo docker:18.09 sh -c 'echo "FROM scratch" > Dockerfile && docker build --stream .'
Error response from daemon: experimental session with v1 builder is no longer supported, use builder version v2 (BuildKit) instead
Signed-off-by: Sebastiaan van Stijn <github@gone.nl>
We aren't using containerd's image store, so we shouldn't be setting
this value.
This fixes container checkpoints, where containerd attempts to
checkpoint the image since one is set, but the image does not exist in
containerd.
Signed-off-by: Brian Goff <cpuguy83@gmail.com>
The daemon expects to manage/handle everything going into the containerd namespace it uses, and things break when it shares that namespace with a second instance, so we should expose these flags and expect users that want two daemons sharing a containerd instance to use separate namespaces for them.
Signed-off-by: Tianon Gravi <admwiggin@gmail.com>
This allows us to cache manifests and avoid extra round trips to the
registry for content we already know about.
dockerd currently does not support containerd on Windows, so this does
not store manifests on Windows, yet.
Signed-off-by: Brian Goff <cpuguy83@gmail.com>
The current usage of reflect.SliceHeader violates rule 6th of
unsafe.Pointer conversion. In short, reflect.SliceHeader could not be
used as plain struct.
See https://golang.org/pkg/unsafe/#Pointer
Signed-off-by: Cuong Manh Le <cuong.manhle.vn@gmail.com>
This prevents consumers of the opts package to also having to
depend on daemon/network, and everything related.
We can probably change some of the other constants to strings,
for easier concatenating, and need to review the windows-specific
"127.0.0.1" (instead of "localhost"), which may no longer be
needed.
Signed-off-by: Sebastiaan van Stijn <github@gone.nl>
The tag service does a `HEAD` request to get the manifest digest, where
we can then do a `GET` against the digest.
The `GET` by tag is not cacheable, but the `GET` against the digest is.
This allows proxies to work way better.
Signed-off-by: Brian Goff <cpuguy83@gmail.com>
full diff: https://github.com/moby/sys/compare/mountinfo/v0.1.3...mountinfo/v0.4.0
> Note that this dependency uses submodules, providing "github.com/moby/sys/mount"
> and "github.com/moby/sys/mountinfo". Our vendoring tool (vndr) currently doesn't
> support submodules, so we vendor the top-level moby/sys repository (which contains
> both) and pick the most recent tag, which could be either `mountinfo/vXXX` or
> `mount/vXXX`.
github.com/moby/sys/mountinfo v0.4.0
--------------------------------------------------------------------------------
Breaking changes:
- `PidMountInfo` is now deprecated and will be removed before v1.0; users should switch to `GetMountsFromReader`
Fixes and improvements:
- run filter after all fields are parsed
- correct handling errors from bufio.Scan
- documentation formatting fixes
github.com/moby/sys/mountinfo v0.3.1
--------------------------------------------------------------------------------
- mount: use MNT_* flags from golang.org/x/sys/unix on freebsd
- various godoc and CI fixes
- mountinfo: make GetMountinfoFromReader Linux-specific
- Add support for OpenBSD in addition to FreeBSD
- mountinfo: use idiomatic naming for fields
github.com/moby/sys/mountinfo v0.2.0
--------------------------------------------------------------------------------
Bug fixes:
- Fix path unescaping for paths with double quotes
Improvements:
- Mounted: speed up by adding fast paths using openat2 (Linux-only) and stat
- Mounted: relax path requirements (allow relative, non-cleaned paths, symlinks)
- Unescape fstype and source fields
- Documentation improvements
Testing/CI:
- Unit tests: exclude darwin
- CI: run tests under Fedora 32 to test openat2
- TestGetMounts: fix for Ubuntu build system
- Makefile: fix ignoring test failures
- CI: add cross build
github.com/moby/sys/mount v0.1.1
--------------------------------------------------------------------------------
https://github.com/moby/sys/releases/tag/mount%2Fv0.1.1
Improvements:
- RecursiveUnmount: add a fast path (#26)
- Unmount: improve doc
- fix CI linter warning on Windows
Testing/CI:
- Unit tests: exclude darwin
- Makefile: fix ignoring test failures
- CI: add cross build
Signed-off-by: Sebastiaan van Stijn <github@gone.nl>
Signed-off-by: Sebastiaan van Stijn <github@gone.nl>
just some minor cleaning up; construct the endpoint-path once
instead of twice, and grouping all variable declarations at the
start.
Signed-off-by: Sebastiaan van Stijn <github@gone.nl>
This fixes the case where log rotation fails on Windows while there are
clients reading container logs.
Evicts readers if there is an error during rotation and try rotation again.
This is needed for Windows with this scenario:
1. `docker logs -f` is called
2. Log rotation occurs (log.txt -> log.txt.1, truncate and re-open
log.txt)
3. Log rotation occurs again (rm log.txt.1, log.txt -> log.txt.1)
On step 3, before this change, the log rotation will fail with `Access
is denied`.
In this case, what we have is a reader holding a file handle to the
primary log file. The log file is then rotated, but the reader still has
a the handle open. `FILE_SHARE_DELETE` allows this to happen... but then
we try to do it again for the next rotation and it blows up.
So when it blows up we force all the readers to disconnect, close the
log file, and try rotation again, which will succeed based on the added
tests.
Signed-off-by: Brian Goff <cpuguy83@gmail.com>
This makes sure, on Windows, that all files are opened with
FILE_SHARE_DELETE.
On non-Windows this just calls the same `os.Open()`.
Signed-off-by: Brian Goff <cpuguy83@gmail.com>
- Ignore some pointless errors (like not exist on remove)
- Consolidate error handling/logging
- Fix race condition reading last log timestamp in the compression
goroutine. This needs to be done while holding the write lock, which
is not (or may not be) locked while compressing a rotated log file.
- Remove some indentation and consolidate mutex unlocking in
`compressFile`
Signed-off-by: Brian Goff <cpuguy83@gmail.com>
I don't think we still had a use for this script, and if someone
needs this, it'll still be accessible in older tags/releases.
Signed-off-by: Sebastiaan van Stijn <github@gone.nl>
This test was added in 0fc11699ab as
a regression test for saving credentials by the CLI.
However, the CLI code is no longer in this repository, so it was
no longer testing anything.
Signed-off-by: Sebastiaan van Stijn <github@gone.nl>
