bump containerd to v1.2.0-13-g8afcade1

[thaJeztah]

Follow-up to #37932

Update runc binary: removed this bump, as it was updated already in #38128

• opencontainers/runc@a00bf01...58592df
  • libcontainer: intelrdt: add support for Intel RDT/MBA in runc opencontainers/runc#1632 libcontainer: intelrdt: add support for Intel RDT/MBA in runc
  • libcontainer: CurrentGroupSubGIDs -> CurrentUserSubGIDs opencontainers/runc#1880 libcontainer: CurrentGroupSubGIDs -> CurrentUserSubGIDs

Update containerd; changes since v1.2.0 (#37932)

• containerd/containerd@v1.2.0...8afcade
  • Bump to Go 1.11.0 containerd/containerd#2435 Bump to Go 1.11.0
  • Change unsupported snapshot warnings to INFO containerd/containerd#2745 Change unsupported snapshot warnings to INFO
  • Bump aufs for unsupported errors containerd/containerd#2746 Bump aufs for unsupported errors
  • bugfix: optimize shim lock in runtime v1 avoid dead lock containerd/containerd#2743 bugfix: optimize shim lock in runtime v1 avoid dead lock
  • bugfix: CloseIO should return correct status code containerd/containerd#2747 bugfix: CloseIO should return correct status code
  • Increase reaper buffer size and non-blocking send containerd/containerd#2748 Increase reaper buffer size and non-blocking send

[thaJeztah]

Marked it "wip", because 1.2.0 hasn't been tagged yet

[codecov]

Codecov Report

❗ No coverage uploaded for pull request base (master@a5e2dd2). Click here to learn what that means.
The diff coverage is n/a.

@@            Coverage Diff            @@
##             master   #38062   +/-   ##
=========================================
  Coverage          ?   36.22%           
=========================================
  Files             ?      610           
  Lines             ?    45475           
  Branches          ?        0           
=========================================
  Hits              ?    16475           
  Misses            ?    26750           
  Partials          ?     2250

[AkihiroSuda]

weird

https://jenkins.dockerproject.org/job/Docker-PRs-experimental/42645/console

13:48:24 PASS: docker_cli_pull_test.go:184: DockerHubPullSuite.TestPullScratchNotAllowed	0.022s
13:48:24 OK: 1475 passed, 52 skipped
13:48:24 PASS
13:48:24 ---> Making bundle: .integration-daemon-stop (in bundles/test-integration)
13:48:26 /go/src/github.com/docker/docker/hack/make/.integration-daemon-stop: line 10: wait: pid 29687 is not a child of this shell
13:48:26 warning: PID 29687 from bundles/test-integration/dfb5d22ec8645/docker.pid had a nonzero exit code
13:48:26 Clearing AppArmor profiles cache:.
13:48:26 All profile caches have been cleared, but no profiles have been unloaded.
13:48:26 Unloading profiles will leave already running processes permanently
13:48:26 unconfined, which can lead to unexpected situations.
13:48:26 
13:48:26 To set a process to complain mode, use the command line tool
13:48:26 'aa-complain'. To really tear down all profiles, run the init script
13:48:26 with the 'teardown' option."
13:48:26 Removing test suite binaries
14:50:12 Build timed out (after 180 minutes). Marking the build as failed.
14:50:12 Build timed out (after 180 minutes). Marking the build as aborted.

[tonistiigi]

CI seems to be hanging because of leaked containerd-shim processes created while docker-py ran.

Logs and processes in: https://gist.github.com/tonistiigi/a5739d2c0f82442c0317bea044cc56e2

From the logs. The test ran a DELETE request for that failed with

time="2018-10-23T19:03:57.531600098Z" level=error msg="fc42a122b164349bced9e08588f6782f65de38e823b5649467533cd07d359464 cleanup: failed to delete container from containerd: cannot delete running task fc42a122b164349bced9e08588f6782f65de38e823b5649467533cd07d359464: failed precondition"

and leaked the shim process.

[AkihiroSuda]

Does the failure relate to containerd/containerd#2748 ?

[thaJeztah]

I added a commit to see if that fixes the issue, thanks!

Diff in that commit; containerd/containerd@c444666...8afcade

relevant changes;

• Increase reaper buffer size and non-blocking send containerd/containerd#2748 Increase reaper buffer size and non-blocking send
• bugfix: CloseIO should return correct status code containerd/containerd#2747 bugfix: CloseIO should return correct status code
• bugfix: optimize shim lock in runtime v1 avoid dead lock containerd/containerd#2743 bugfix: optimize shim lock in runtime v1 avoid dead lock
• Bump aufs for unsupported errors containerd/containerd#2746 Bump aufs for unsupported errors
• Change unsupported snapshot warnings to INFO containerd/containerd#2745 Change unsupported snapshot warnings to INFO

[AkihiroSuda]

Weird, even without this PR, docker-py fails and hangs on my laptop (b8e87cf)

$ make test-docker-py
...
 generated xml file: /go/src/github.com/docker/docker/bundles/test-docker-py/results.xml
=========================== short test summary info ============================
XFAIL ../../../../../docker-py/tests/integration/api_container_test.py::CreateContainerTest::test_create_with_cpu_rt_options
  CONFIG_RT_GROUP_SCHED isn't enabled
XFAIL ../../../../../docker-py/tests/integration/api_container_test.py::CreateContainerTest::test_create_with_init_path
  init-path removed in 17.05.0
XFAIL ../../../../../docker-py/tests/integration/api_container_test.py::CreateContainerTest::test_create_with_storage_opt
  Not supported on most drivers
XFAIL ../../../../../docker-py/tests/integration/api_swarm_test.py::SwarmTest::test_init_swarm_with_log_driver
  This doesn't seem to be taken into account by the engine
SKIP [1] tests/integration/api_image_test.py:289: Doesn't work inside a container - FIXME
SKIP [1] /docker-py/tests/integration/api_swarm_test.py:31: Test stalls the engine on 1.12.0
=================================== FAILURES ===================================
____________________ BuildTest.test_build_invalid_platform _____________________
/docker-py/tests/helpers.py:70: in wrapped
    return f(self, *args, **kwargs)
/docker-py/tests/integration/api_build_test.py:388: in test_build_invalid_platform
    assert excinfo.value.status_code == 400
E   AssertionError: assert 500 == 400
E    +  where 500 = APIError(HTTPError(u'500 Server Error: Internal Server Error for url: http+doc...ld?q=False&pull=False&nocache=False&forcerm=False&rm=False&platform=foobar',),).status_code
E    +    where APIError(HTTPError(u'500 Server Error: Internal Server Error for url: http+doc...ld?q=False&pull=False&nocache=False&forcerm=False&rm=False&platform=foobar',),) = <ExceptionInfo APIError tblen=5>.value
___________________ PullImageTest.test_pull_invalid_platform ___________________
/docker-py/tests/helpers.py:70: in wrapped
    return f(self, *args, **kwargs)
/docker-py/tests/integration/api_image_test.py:73: in test_pull_invalid_platform
    assert 'invalid platform' in excinfo.exconly()
E   AssertionError: assert 'invalid platform' in 'APIError: 500 Server Error: Internal Server Error (""foobar": unknown operating system or architecture: invalid argument")'
E    +  where 'APIError: 500 Server Error: Internal Server Error (""foobar": unknown operating system or architecture: invalid argument")' = <bound method ExceptionInfo.exconly of <ExceptionInfo APIError tblen=4>>()
E    +    where <bound method ExceptionInfo.exconly of <ExceptionInfo APIError tblen=4>> = <ExceptionInfo APIError tblen=4>.exconly
=== 2 failed, 329 passed, 2 skipped, 4 xfailed, 1 xpassed in 497.40 seconds ====
(hangs up)

[thaJeztah]

Interesting; https://jenkins.dockerproject.org/job/Docker-PRs/51642/console

18:59:40 Build timed out (after 200 minutes). Marking the build as failed.
18:59:40 Build timed out (after 200 minutes). Marking the build as aborted.

let me restart

edit: never mind, was confusing this for another PR 😅

[thaJeztah]

Rebased; removed the runc-bump as it was updated already in #38128

[thaJeztah]

closing this one, as this was included in #38168