I'm fairly consistently seeing an error in
DockerSuite.TestContainerApiRestartNotimeoutParam:
docker_api_containers_test.go:969:
c.Assert(status, check.Equals, http.StatusNoContent)
... obtained int = 500
... expected int = 204
And in the daemon logs I see:
INFO[0003] Container 8cf77c20275586b36c5095613159cf73babf92ba42ed4a2954bd55dca6b08971 failed to exit within 0 seconds of SIGTERM - using the force
ERRO[0003] Handler for POST /containers/{name:.*}/restart returned error: Cannot restart container 8cf77c20275586b36c5095613159cf73babf92ba42ed4a2954bd55dca6b08971: [2] Container does not exist: container destroyed
ERRO[0003] HTTP Error err=Cannot restart container 8cf77c20275586b36c5095613159cf73babf92ba42ed4a2954bd55dca6b08971: [2] Container does not exist: container destroyed
statusCode=500
Note the "container destroyed" error message. This is being generatd by
the libcontainer code and bubbled up in container.Kill() as a result of the
call to `container.killPossiblyDeadProcess(9)` on line 439.
See the comment in the code, but what I think is going on is that because we
don't have any timeout on the Stop() call we immediate try to force things to
stop. And by the time we get into libcontainer code the process just finished
stopping due to the initial signal, so this secondary sig-9 fails due to the
container no longer running (ie. its 'destroyed').
Since we can't look for "container destroyed" to just ignore the error, because
some other driver might have different text, I opted to just ignore the error
and keep going - with the assumption that if it couldnt send a sig-9 to the
process then it MUST be because its already dead and not something else.
To reproduce this I just run:
curl -v -X POST http://127.0.0.1:2375/v1.19/containers/8cf77c20275586b36c5095613159cf73babf92ba42ed4a2954bd55dca6b08971/restart
a few times and then it fails with the HTTP 500.
Would like to hear some other ideas on to handle this since I'm not
thrilled with the proposed solution.
Signed-off-by: Doug Davis <dug@us.ibm.com>
* Don't AllocateNetwork when network is disabled
* Don't createNetwork in execdriver when network is disabled
Signed-off-by: Alexander Morozov <lk4d4@docker.com>
We should let user create container even if the container he wants
join is not running, that check should be done at start time.
In this case, the running check is done by getIpcContainer() when
we start container.
Signed-off-by: Qiang Huang <h.huangqiang@huawei.com>
Signed by all authors:
Signed-off-by: Michael Crosby <crosbymichael@gmail.com>
Signed-off-by: Arnaud Porterie <arnaud.porterie@docker.com>
Signed-off-by: David Calavera <david.calavera@gmail.com>
Signed-off-by: Jeff Lindsay <progrium@gmail.com>
Signed-off-by: Alexander Morozov <lk4d4@docker.com>
Signed-off-by: Luke Marsden <luke@clusterhq.com>
Signed-off-by: David Calavera <david.calavera@gmail.com>
Sometimes container.cleanup() can be called from multiple paths
for the same container during error conditions from monitor and
regular startup path. So if the container network has been already
released do not try to release it again.
Signed-off-by: Jana Radhakrishnan <mrjana@docker.com>
The DOCKER_EXPERIMENTAL environment variable drives the activation of
the 'experimental' build tag.
Signed-off-by: Arnaud Porterie <arnaud.porterie@docker.com>
As part of this some generic packages like iptables, etchosts and resolvconf
have also been moved to libnetwork. Even though they can still be
consumed in a generic fashion they will reside and be maintained
from within the libnetwork project.
Signed-off-by: Jana Radhakrishnan <mrjana@docker.com>
- Updated Dockerfile to satisfy libnetwork GOPATH requirements.
- Reworked daemon to allocate network resources using libnetwork.
- Reworked remove link code to also update network resources in libnetwork.
- Adjusted the exec driver command population to reflect libnetwork design.
- Adjusted the exec driver create command steps.
- Updated a few test cases to reflect the change in design.
- Removed the dns setup code from docker as resolv.conf is entirely managed
in libnetwork.
- Integrated with lxc exec driver.
Signed-off-by: Jana Radhakrishnan <mrjana@docker.com>
Prior to this patch, the response of
- GET /images/json
- GET /containers/json
- GET /images/(name)/history
display the Created Time as UNIX format which doesn't make sense.
These should be more readable as CLI command `docker inspect` shows.
Due to the case that an older client with a newer version daemon, we
need the version check for now.
Signed-off-by: Hu Keping <hukeping@huawei.com>
- noplog driver pkg for '--log-driver=none' (null object pattern)
- centralized factory for log drivers (instead of case/switch)
- logging drivers registers themselves to factory upon import
(easy plug/unplug of drivers in daemon/logdrivers.go)
- daemon now doesn't start with an invalid log driver
- Name() method of loggers is actually now their cli names (made it useful)
- generalized Read() logic, made it unsupported except json-file (preserves
existing behavior)
Spotted some duplication code around processing of legacy json-file
format, didn't touch that and refactored in both places.
Signed-off-by: Ahmet Alp Balkan <ahmetalpbalkan@gmail.com>
Previously the cache was only updated once on startup, because the graph
code only check for filesystems on startup. However this breaks the API as it
was supposed and so unit tests.
Fixes#13142
Signed-off-by: Jörg Thalheim <joerg@higgsboson.tk>
Add handler for SIGUSR1 based on feedback regarding when to dump
goroutine stacks. This will also dump goroutine stack traces on SIGQUIT
followed by a hard-exit from the daemon.
Docker-DCO-1.1-Signed-off-by: Phil Estes <estesp@linux.vnet.ibm.com> (github: estesp)
Added --since argument to `docker logs` command. Accept unix
timestamps and shows logs only created after the specified date.
Default value is 0 and passing default value or not specifying
the value in the request causes parameter to be ignored (behavior
prior to this change).
Signed-off-by: Ahmet Alp Balkan <ahmetalpbalkan@gmail.com>
If firewalld is not installed (or I suppose not running), firewalld was
producing an error in the daemon init logs, even though firewalld is not
required for iptables stuff to function.
The firewalld library code was also logging directly to logrus instead
of returning errors.
Moved logging code higher up in the stack and changed firewalld code to
return errors where appropriate.
Signed-off-by: Brian Goff <cpuguy83@gmail.com>
Generation based on CAP_LAST_CAP, I hardcoded
capability.CAP_BLOCK_SUSPEND as last for systems which has no
/proc/sys/kernel/cap_last_cap
Signed-off-by: Alexander Morozov <lk4d4@docker.com>
The docker graph call driver.Exists() on initialisation for each filesystem in
the graph. This results will results in a lot `zfs get all` commands. To reduce
this, retrieve all descend filesystem at startup and cache it for later checks
Signed-off-by: Jörg Thalheim <joerg@higgsboson.tk>
instead of let zfs automaticly mount datasets, mount them on demand using mount(2).
This speed up this graph driver in 2 ways:
- less zfs processes needed to start a container
- /proc/mounts get smaller, so zfs userspace tools has less to read (which can
a significant amount of data as the number of layer grows)
This ways it can be also ensured that the correct mountpoint is always used.
Signed-off-by: Jörg Thalheim <joerg@higgsboson.tk>
Signed-off-by: Michael Crosby <crosbymichael@gmail.com>
Add tests for mounting into /proc and /sys
These two locations should be prohibited from mounting volumes into
those destinations.
Signed-off-by: Michael Crosby <crosbymichael@gmail.com>
The lxc code here is doing the exact same thing on calling
execdriver.Terminate, so let's just use that.
Also removes some dead comments originally introduced
50144aeb42 but no longer relevant since we
have restart policies.
Signed-off-by: Brian Goff <cpuguy83@gmail.com>
`lxc-stop` does not support sending arbitrary signals.
By default, `lxc-stop -n <id>` would send `SIGPWR`.
The lxc driver was always sending `lxc-stop -n <id> -k`, which always
sends `SIGKILL`. In this case `lxc-start` returns an exit code of `0`,
regardless of what the container actually exited with.
Because of this we must send signals directly to the process when we
can.
Also need to set quiet mode on `lxc-start` otherwise it reports an error
on `stderr` when the container exits cleanly (ie, we didn't SIGKILL it),
this error is picked up in the container logs... and isn't really an
error.
Also cleaned up some potential races for waitblocked test.
Signed-off-by: Brian Goff <cpuguy83@gmail.com>
The `--userland-proxy` daemon flag makes it possible to rely on hairpin
NAT and additional iptables routes instead of userland proxy for port
publishing and inter-container communication.
Usage of the userland proxy remains the default as hairpin NAT is
unsupported by older kernels.
Signed-off-by: Arnaud Porterie <arnaud.porterie@docker.com>
To help avoid version mismatches between libcontainer and Docker, this updates libcontainer to be the source of truth for which version of logrus the project is using. This should help avoid potential incompatibilities in the future, too. 👍
Signed-off-by: Andrew "Tianon" Page <admwiggin@gmail.com>
Adds a `stream` query param to the stats API which allows API users to
only collect one stats entry and disconnect instead of keeping the
connection alive to stream more stats.
Also adds a `--no-stream` flag to `docker stats` which does the same
Signed-off-by: Brian Goff <cpuguy83@gmail.com>
Right now devicemapper mounts thin device using online discards by default
and passes mount option "discard". Generally people discourage usage of
online discards as they can be a drain on performance. Instead it is
recommended to use fstrim once in a while to reclaim the space.
In case of containers, we recommend to keep data volumes separate. So
there might not be lot of rm, unlink operations going on and there might
not be lot of space being freed by containers. So it might not matter
much if we don't reclaim that free space in pool.
User can still pass mount option explicitly using dm.mountopt=discard to
enable discards if they would like to.
So this is more like setting the containers by default for better performance
instead of better space efficiency in pool. And user can change the behavior
if they don't like default behavior.
Reported-by: Mike Snitzer <snitzer@redhat.com>
Signed-off-by: Vivek Goyal <vgoyal@redhat.com>
This patch modifies the journald log driver to store the container ID in
a field named CONTAINER_ID, rather than (ab)using the MESSAGE_ID field.
Additionally, this adds the CONTAINER_ID_FULL field containing the
complete container ID and CONTAINER_NAME, containing the container name.
When using the journald log driver, this permits you to see log messages
from a particular container like this:
# journalctl CONTAINER_ID=a9238443e193
Example output from "journalctl -o verbose" includes the following:
CONTAINER_ID=27aae7361e67
CONTAINER_ID_FULL=27aae7361e67e2b4d3864280acd2b80e78daf8ec73786d8b68f3afeeaabbd4c4
CONTAINER_NAME=web
Closes: #12864
Signed-off-by: Lars Kellogg-Stedman <lars@redhat.com>
prioritize the ports with static mapping before dynamic mapping. This removes
the port conflicts when we allocate static port in the reserved range
together with dynamic ones.
When static port is allocated first, Docker will skip those when determining
free ports for dynamic ones.
Signed-off-by: Daniel, Dao Quang Minh <dqminh89@gmail.com>
When a container has errors on removal, it gets flagged as dead.
If you `docker rm -f` a dead container the container is dereffed from
the daemon and doesn't show up on `docker ps` anymore... except that the
container JSON file may still be lingering around and becomes undead
when you restart the daemon.
Signed-off-by: Brian Goff <cpuguy83@gmail.com>
statsCollector.publishers must be protected to prevent
modifications during the iteration in run().
Being locked for a long time is bad, so pairs of containers &
publishers (pointers) are copied to release the lock fast.
Signed-off-by: Anton Tiurin <noxiouz@yandex.ru>
Due to the importance of path safety, the internal sanitisation wrappers
for volumes and containers should be exposed so other parts of Docker
can benefit from proper path sanitisation.
Signed-off-by: Aleksa Sarai <cyphar@cyphar.com> (github: cyphar)
Assuming that docker kill is trying to actually kill the container
is a mistake. If the container is not running we should report it
back to the caller as a error.
Docker-DCO-1.1-Signed-off-by: Dan Walsh <dwalsh@redhat.com> (github: rhatdan)
Docker-DCO-1.1-Signed-off-by: Regan McCooey <rmccooey27@aol.com> (github: rmccooey27)
Docker-DCO-1.1-Signed-off-by: Regan McCooey <rmccooey27@aol.com> (github: rhatdan)
This also moves `exec -i` test to _unix_test.go because it seems to need a
pty to reliably reproduce the behavior.
Signed-off-by: Daniel, Dao Quang Minh <dqminh89@gmail.com>
- Mount struct now called volumeMount
- Merged volume creation for each volume type (volumes-from, binds, normal
volumes) so this only happens in once place
- Simplified container copy of volumes (for when `docker cp` is a
volume)
Signed-off-by: Brian Goff <cpuguy83@gmail.com>
If device is being reactivated before it could go away and deferred
deactivation is scheduled on it, cancel it.
Signed-off-by: Vivek Goyal <vgoyal@redhat.com>
This will help with debugging as one could just do "docker info" and figure
out of deferred removal is enabled or not.
Signed-off-by: Vivek Goyal <vgoyal@redhat.com>
Provide a new command line knob dm.deferred_device_removal which will enable
deferred device deactivation if driver and library support it.
This patch also checks for library support and driver version.
Signed-off-by: Vivek Goyal <vgoyal@redhat.com>
John Howard
Antonio Murdaca
Darren Shepherd
Doug Davis
Alexander Morozov
Arnaud Porterie
Madhu Venugopal
Alexander Morozov
David Calavera
Qiang Huang
Brian Goff
Jana Radhakrishnan
Lei Jitang
Jessie Frazelle
Zhang Wei
Michael Crosby
Ahmet Alp Balkan
Shishir Mahajan
unclejack
Victor Vieux
Tibor Vass
Sebastiaan van Stijn
Hu Keping
Jessica Frazelle
Jörg Thalheim
Phil Estes
Gianluca Borello
Bharath Thiruveedula
Ma Shimiao
wlan0
Burke Libbey
Anthony Baire
Vincent Batts
mauriyouth
Phil Estes
Tianon Gravi
unclejack
Vivek Goyal
Michael Crosby
Lars Kellogg-Stedman
Daniel, Dao Quang Minh
Dan Walsh
Jessica Frazelle
David Mackey
Aidan Hobson Sayers
jmzwcn
Anton Tiurin
Aleksa Sarai
Regan McCooey
jianbosun
Gosuke Miyashita