Use utils.RFC3339NanoFixed ("2006-01-02T15:04:05.000000000Z07:00")
instead of time.RFC3339Nano to format our log timestamps - this way
things are aligned, in particular the nano seconds are padded with zeros
Signed-off-by: Doug Davis <dug@us.ibm.com>
- Errors sent to the walker callback functions were ignored. This meant that
one could get a panic when calling methods on a nil FileInfo object. For
example when the file did not exists any more.
- Lstat calls inside walker callback are reduntant because walker already calls
Lstat and passes the result to the callback.
- Error returned from filepath.Rel() can never be EACCES because it compares
strings and does not care about actual files.
- If Matched() returns error then ValidateContextDirectory() must return error.
Currently it still kept walking although the outcome was already known.
- Function will now fail in case of unknown error(not EACCES nor ENOENT).
Previous implementation did not make a clear decision about this (but
panicked because of the issues above).
Signed-off-by: Tõnis Tiigi <tonistiigi@gmail.com> (github: tonistiigi)
* Events subsystem merged from `server/events.go` and
`utils/jsonmessagepublisher.go` and moved to `events/events.go`
* Only public interface for this subsystem is engine jobs
* There is two new engine jobs - `log_event` and `subscribers_count`
* There is auxiliary function `container.LogEvent` for logging events for
containers
Docker-DCO-1.1-Signed-off-by: Alexandr Morozov <lk4d4math@gmail.com> (github: LK4D4)
[solomon@docker.com: resolve merge conflicts]
Signed-off-by: Solomon Hykes <solomon@docker.com>
/tmp is often a tmpfs file system and large temporary files could cause
docker commands to fail. Also using /tmp potentially allows users on the
system to get access to content, or even attack the content. Moving the tmpdir to
/var/lib/container/tmp will protect the data.
Docker-DCO-1.1-Signed-off-by: Dan Walsh <dwalsh@redhat.com> (github: rhatdan)
Conflicts:
docker/docker.go
The initial `ValidateContextDirectory` implementation fails loudly when a file
lacks read permissions in the current context. However that situation is valid
if the file is included in the `.dockerignore` patterns.
Docker-DCO-1.1-Signed-off-by: Bruno Renié <brutasse@gmail.com> (github: brutasse)
functions to pkg/parsers/kernel, and parsing filters to
pkg/parsers/filter. Adjust imports and package references.
Docker-DCO-1.1-Signed-off-by: Erik Hollensbe <github@hollensbe.org> (github: erikh)
Make it possible to inspect an initial response and pass it to
ResumableRequestReader. This makes it possible to inspect an initial
response and passing it to ResumableRequestReader to avoid making an
extra request.
Docker-DCO-1.1-Signed-off-by: Cristian Staretu <cristian.staretu@gmail.com> (github: unclejack)
Currently the docker logs timestamp flag generates log entries like:
$ sudo docker logs -ft daemon_dave
[May 10 13:06:17.934] hello world
It uses Go's StampMilli timestamp to generate the timestamp. The entry
is also wrapped in [ ].
This is non-standard operational timestamp and one that will require
custom parsing.
The new timestamp is RFC3999Nano and generates entries like:
2014-05-10T17:42:14.999999999Z07:00 hello world
These are readily parsed by tools like ELK.
Docker-DCO-1.1-Signed-off-by: James Turnbull <james@lovedthanlost.net> (github: jamtur01)
Docker-DCO-1.1-Signed-off-by: Tibor Vass <teabee89@gmail.com> (github: tiborvass)
It became slightly faster and lighter
possibly fixes#5923 problems
Docker-DCO-1.1-Signed-off-by: Alexandr Morozov <lk4d4math@gmail.com> (github: LK4D4)
The JSONMessage.Display methods erases the entire line by outputting the <ESC>[2K\r sequence before displaying the message content. This is not necessary for messages other than progress indicators, and introduces unwanted characters in the /events output.
Fixes#6203.
Docker-DCO-1.1-Signed-off-by: Arnaud Porterie <icecrime@gmail.com> (github: icecrime)
Events subscription (/events API endpoint) attributes pseudo-unique identifiers to incoming subscribers: originally its host, then its subscription time. This is unecessary and leads to code complexity.
Introduce a JSONMessagePublisher to provide simple pub/sub mechanism for JSONMessage, and rely on this new type to publish events to all subscribed listeners. The original logic is kept for the 'since' and 'until' parameters, and for client disconnection handling.
Docker-DCO-1.1-Signed-off-by: Arnaud Porterie <icecrime@gmail.com> (github: icecrime)
adding tests and allowing for easy passing of filters.Args from client
to server.
Docker-DCO-1.1-Signed-off-by: Vincent Batts <vbatts@redhat.com> (github: vbatts)
This is a new feature and flag. (replaces the suggestion of a flag for
--untagged images).
The concept is to have a syntax to filter. This begins with this
filtering for the 'images' subcommand, and at that only filtering for
whether images are untagged.
example like: docker rmi $(docker images -q --filter 'untagged=true')
Docker-DCO-1.1-Signed-off-by: Vincent Batts <vbatts@redhat.com> (github: vbatts)
After removed, the User-Agent shows in log like this:
[debug] http.go:160 https://index.docker.io/v1/repositories/busybox/images --
HEADERS: map[User-Agent:[docker/0.11.1-dev go/go1.2.2 git-commit/8887e00-dirty kernel/3.14.3-n1 os/linux arch/amd64]]
The code also moved all validation work into validVersion,
to keep the main logic as clean.
Docker-DCO-1.1-Signed-off-by: Derek <crq@kernel.org> (github: crquan)
This commit adds a function which can be used to ensure all contents of
a directory can be accessed.
This function doesn't follow symlinks to check if they're pointing to
files which exist. Such symlinks can be useful later.
Docker-DCO-1.1-Signed-off-by: Cristian Staretu <cristian.staretu@gmail.com> (github: unclejack)
to cover a couple of use-cases:
* 1mb file, using no compression
* 1mb file, using compression
* 1024 1k files, using no compression
* 1024 1k files, using compression
Docker-DCO-1.1-Signed-off-by: Vincent Batts <vbatts@redhat.com> (github: vbatts)
This commit refactors TruncIndex to make it possible to add container
ids to the TruncIndex without updating the Suffixarray.
This is useful during the Docker daemon's startup when we don't want to
update the Suffixarray for every container we add.
Add continues to function like before.
Docker-DCO-1.1-Signed-off-by: Cristian Staretu <cristian.staretu@gmail.com> (github: unclejack)
fs_test.go doesn't finish if Docker's code is placed under a directory
which has symlinks between / and the directory.
For example, the below doesn't finish before the change.
/home -> usr/home
FollowSymlinkInScope("/home/bob/foo/bar", "/home/bob/foo")
Docker-DCO-1.1-Signed-off-by: Kato Kazuyoshi <kato.kazuyoshi@gmail.com> (github: kzys)
Fixes#5166
Current graph.restore is essentially O(n^2 log n) due to how
suffixarray creation works.
Rather than create/append/create new this supports creation from a seed
array of ids.
Functional testing shows this eliminates the hang on Creating image
graph reported on list.
Docker-DCO-1.1-Signed-off-by: Paul Nasrat <pnasrat@gmail.com> (github: pnasrat)
In order to handle special configuration for different drivers we
make the Config field a map to string array. This lets
us use it for lxc, by using the "lxc" key for those, and we can
later extend it easily for other backend-specific options.
Docker-DCO-1.1-Signed-off-by: Alexander Larsson <alexl@redhat.com> (github: alexlarsson)
Add a --dns-search parameter and a DnsSearch
configuration field for specifying dns search
domains.
Docker-DCO-1.1-Signed-off-by: Daniel Norberg <daniel.norberg@gmail.com> (github: danielnorberg)
This removes the incomplete symlink handling from engine.go and it adds
it one place in docker.go.
It also enables handling symlinks for TMPDIR.
Docker-DCO-1.1-Signed-off-by: Cristian Staretu <cristian.staretu@gmail.com> (github: unclejack)
This makes the JSON streams a _lot_ easier to parse in less well-baked JSON parsers, and no less so in better ones.
Line-based JSON streams are very, very common, where simply chunk-based is not very common at all.
Docker-DCO-1.1-Signed-off-by: Andrew Page <admwiggin@gmail.com> (github: tianon)
This stops docker from accepting tcp:// as a valid bind address.
Docker-DCO-1.1-Signed-off-by: Cristian Staretu <cristian.staretu@gmail.com> (github: unclejack)
This adds the ability to socket activate docker by passing in
`-H fd://*` along with examples systemd configuration files.
The fastest way to test this is to run:
```
/usr/lib/systemd/systemd-activate -l 127.0.0.1:2001 /usr/bin/docker -d -H 'fd://*'
docker -H tcp://127.0.0.1:2001 ps
```
Docker-DCO-1.1-Signed-off-by: Brandon Philips <brandon.philips@coreos.com> (github: philips)
The idx.index array is overwritten when a new value is inserted to the index.
When two containers are created concurrently, their ids are inserted to the
index and one can overwrite the other leaving one of ids missing from the
index. Adding a RWMutex lock around read/write operations fixes this.
Docker-DCO-1.1-Signed-off-by: James Allen <jamesallen0108@gmail.com> (github: jpallen)
To avoid unexpected results since docker was using http.
For instance, my broadband doesn't return not found when it's down but
a html page saying that the internet is down. Docker was showing that
html instead of ignoring it.
Fix#3570
Docker-DCO-1.1-Signed-off-by: Fabio Falci <fabiofalci@gmail.com> (github: fabiofalci)
Use mountinfo rather than for cgroups parsing.
Make helper method private and change name.
Makes method naming more explicit rather than GetThisCgroup.
Use upstream term subsystem rather than cgroupType.
Only necessary if distro policy dictates that the path deviate from the paths already listed in utils/utils.go - please refrain from using it otherwise.
