| Commit message (Collapse) | Author | Age |
|
|
|
|
|
|
|
|
| |
This gives the service 5 seconds to digest the signal and 5 more seconds
to shutdown. Create a new variable to make bumping the timeout easier
in case we see re-flake in the future.
Fixes: #12167
Signed-off-by: Valentin Rothberg <rothberg@redhat.com>
|
|
|
|
|
|
|
|
|
|
|
|
| |
Fixes: https://bugzilla.redhat.com/show_bug.cgi?id=2030599
When you create pod, it shares the UTS namespace with Containers.
Currently the --hostname is not passed to the pod created when
you create a container and pod in the same command.
Also fix error message on supported --share flags
Signed-off-by: Daniel J Walsh <dwalsh@redhat.com>
|
|
|
|
|
|
|
|
| |
...it's not needed: teardown() already does it. Or, it would,
if it had been updated to deal with the new pause image naming
convention, which I've just done.
Signed-off-by: Ed Santiago <santiago@redhat.com>
|
|\
| |
| | |
generate systemd: support entrypoint JSON strings
|
| |
| |
| |
| |
| |
| |
| | |
Make sure to preserve the quoting of entrypoint JSON strings.
Fixes: #12477
Signed-off-by: Valentin Rothberg <rothberg@redhat.com>
|
|\ \
| | |
| | | |
specgen: honor empty args for entrypoint specified as `--entrypoint ""`
|
| |/
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
Users should be able to override containers entrypoint using
`--entrypoint ""` following works fine for podman but not for podman
remote.
Specgen ignores empty argument for entrypoint so make specgen honor
empty arguments.
Signed-off-by: Aditya Rajan <arajan@redhat.com>
|
|/
|
|
|
|
|
|
|
| |
GLOBAL_OPTS haven't been supported for at least two major versions of
Podman. The runlabel code is extremely fragile and I think it should
be rewritten before adding new features.
Fixes: #12436
Signed-off-by: Valentin Rothberg <rothberg@redhat.com>
|
|\
| |
| | |
Update vendor or containers/common moving pkg/cgroups there
|
| |
| |
| |
| |
| |
| |
| | |
[NO NEW TESTS NEEDED] This is just moving pkg/cgroups out so
existing tests should be fine.
Signed-off-by: Daniel J Walsh <dwalsh@redhat.com>
|
|/
|
|
|
|
|
|
| |
While trying to match permissions of target directory podman adds
extra `0111` which should not be needed if target path does not have
execute permission.
Signed-off-by: Aditya Rajan <arajan@redhat.com>
|
|\
| |
| | |
tests: clean up FIXMEs and noise
|
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
e2e tests:
* remove two FIXMEs:
* one of them is expected behavior, not a FIXME
* the other is easily fixed
* File issue #12521 for a real podman-remote bug, and
update the Skip() message
system tests:
* in command-completion test, clean up stray podman-pause image
(followup to #12322, in which I missed this instance). This
removes distracting warnings from test logs.
Signed-off-by: Ed Santiago <santiago@redhat.com>
|
|\ \
| | |
| | | |
fix remote run/start flake
|
| |/
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
Fix the flake reported in #9597 with a workaround to at least stop
wasting energy until the root cause has been found and fixed.
It seems that a remote run returns before the container has transitioned
into the `exited` state which ultimately breaks a subsequent remote
start with attach.
Signed-off-by: Valentin Rothberg <rothberg@redhat.com>
|
|/
|
|
|
|
|
|
|
| |
Give the services some more time after SIGINT to shutdown.
`Eventually(session, 5)` will poll the session for the specific
condition at a frequency of 10ms for at most 5 seconds.
Fixes: #12167
Signed-off-by: Valentin Rothberg <rothberg@redhat.com>
|
|\
| |
| | |
Refactor podman system to report.Formatter
|
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
[NO NEW TESTS NEEDED]
Support better compatibility output for podman system commands
* Format and content of output from podman version changed to
be more compatible
See #10974
Depends on containers/common#831
Signed-off-by: Jhon Honce <jhonce@redhat.com>
|
| |
| |
| |
| |
| |
| |
| |
| |
| | |
Update the `--filter reference=...` tests to reflect recent changes in
c/common. The reference values now match as specified without
implicitly adding wildcards arounds.
Fixes: #11905
Signed-off-by: Valentin Rothberg <rothberg@redhat.com>
|
|\ \
| |/
|/|
| |
| | |
dispensable/respect_pod_yaml_spec_hostname_when_play
Hostname in `spec.hostname` should be passed to infra ctr init opt
|
| |
| |
| |
| |
| |
| | |
Fixes https://github.com/containers/podman/issues/12393
Signed-off-by: Qiang Wang <sunsetmask@gmail.com>
|
|\ \
| | |
| | | |
Add support for configmap volumes to play kube
|
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | | |
If the k8s yaml has volumes from a configmap, play kube
will now create a volume based on the data from the
configmap and volume source and set it to the right path
in the container accordingly.
Add tests for this and update some test for ENV from configmap.
Signed-off-by: Urvashi Mohnani <umohnani@redhat.com>
|
| | |
| | |
| | |
| | | |
Signed-off-by: Ondra Machacek <omachace@redhat.com>
|
|\ \ \
| | | |
| | | | |
e2e tests: use HaveKey() and HaveLen() when possible
|
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | | |
Fix a handful of instances not covered by earlier automated
replacements. Found via:
ack 'Expect\(len' test/e2e
There are still a bunch of BeNumerically(">", ...) that cannot (yet)
be handled by HaveLen(). Leave those as they are.
Signed-off-by: Ed Santiago <santiago@redhat.com>
|
| | | |
| | | |
| | | |
| | | |
| | | |
| | | | |
sed -i -e 's/Expect(len(\(.*\)))\.To(BeNumerically("==", \(.*\)))/Expect(\1).To(HaveLen(\2))/' test/e2e/*.go
Signed-off-by: Ed Santiago <santiago@redhat.com>
|
| | | |
| | | |
| | | |
| | | |
| | | |
| | | | |
sed -i -e 's/Expect(len(\(.*\)))\.To(Equal(\(.*\)))/Expect(\1).To(HaveLen(\2))/' test/e2e/*.go
Signed-off-by: Ed Santiago <santiago@redhat.com>
|
| | | |
| | | |
| | | |
| | | |
| | | |
| | | | |
sed -i -e 's/Expect(len(\(.*\)))\.To(BeNumerically("==", 0))/Expect(\1).To(BeEmpty())/' test/e2e/*.go
Signed-off-by: Ed Santiago <santiago@redhat.com>
|
| | | |
| | | |
| | | |
| | | |
| | | |
| | | | |
sed -i -e 's/Expect(len(\(.*\)))\.To(Equal(0))/Expect(\1).To(BeEmpty())/' test/e2e/*.go
Signed-off-by: Ed Santiago <santiago@redhat.com>
|
| | | |
| | | |
| | | |
| | | |
| | | |
| | | | |
sed -i -e 's/Expect(\(.*\)\[\(\".*\"\)\])\.To(\(.*\)/Expect(\1).To(HaveKeyWithValue(\2, \3)/' test/e2e/*_test.go
Signed-off-by: Ed Santiago <santiago@redhat.com>
|
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | | |
sed -i -e 's/Expect(\(.*\)\[\(\".*\"\)\])\.To(Equal(/Expect(\1).To(HaveKeyWithValue(\2, /' test/e2e/*_test.go
...with two manual tweaks, because this converted:
Expect(foo["bar"]).To(Equal(""))
-> Expect(foo).To(HaveKeyWithValue("bar",""))
It looks like the intention of the test was, instead:
...To(Not(HaveKey("bar")))
Signed-off-by: Ed Santiago <santiago@redhat.com>
|
|\ \ \ \
| | | | |
| | | | | |
Refactor podman image command output
|
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | | |
Leverage new report.Formatter allowing better compatibility from
podman command output.
Follow on PR's will cover containers, etc.
See #10974
Signed-off-by: Jhon Honce <jhonce@redhat.com>
|
|\ \ \ \ \
| |_|/ / /
|/| | | | |
compat: images/json
|
| |/ / /
| | | |
| | | |
| | | |
| | | |
| | | |
| | | | |
Do not list manifest lists. Docker doesn't either.
Fixes: #12453
Signed-off-by: Valentin Rothberg <rothberg@redhat.com>
|
|\ \ \ \
| |/ / /
|/| | | |
e2e tmpdir cleanup
|
| | | |
| | | |
| | | |
| | | |
| | | |
| | | | |
...they're not actually used for anything
Signed-off-by: Ed Santiago <santiago@redhat.com>
|
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | | |
Images were being cached in /tmp, with no option to
override. Now $PODMAN_TEST_IMAGE_CACHE_DIR can be
used to point to a user-preferred location. If unset,
try $TMPDIR before settling on /tmp.
Also: refactor the logic for determining the tarball name.
Also: include registry name in tarball name.
Also: clean up unused/unnecessary code
Also: do not echo "Restoring..." if we're not actually restoring.
Signed-off-by: Ed Santiago <santiago@redhat.com>
|
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | | |
...and remove other uses of "crio". They're confusing
and misleading. (I'm sure it made sense at one time)
Signed-off-by: Ed Santiago <santiago@redhat.com>
|
|\ \ \ \
| | | | |
| | | | | |
e2e: yet more cleanup of BeTrue/BeFalse
|
| | |_|/
| |/| |
| | | |
| | | |
| | | |
| | | | |
Thanks to Paul for teaching me about HaveKey()
Signed-off-by: Ed Santiago <santiago@redhat.com>
|
|\ \ \ \
| | | | |
| | | | | |
Implement 'podman run --blkio-weight-device'
|
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | | |
`--blkio-weight-device` is not fully implemented and this causes an
unexpected panic when specified because an entry is put into an
uninitialized map at parsing.
This fix implements the `--blkio-weight-device` and adds a system test.
When creating a spec generator on a client, a major number and a minor
number of a device cannot be set. So, these numbers are inspected on a
server and set to a runtime spec.
Signed-off-by: Hironori Shiina <shiina.hironori@jp.fujitsu.com>
|
|\ \ \ \ \
| |_|_|/ /
|/| | | | |
podman, push: expose --compression-format
|
| |/ / /
| | | |
| | | |
| | | |
| | | |
| | | | |
support overriding the compression format at push time.
Signed-off-by: Giuseppe Scrivano <gscrivan@redhat.com>
|
|\ \ \ \
| |_|/ /
|/| | | |
top: parse ps(1) args correctly
|
| |/ /
| | |
| | |
| | |
| | |
| | |
| | | |
The arguments of ps(1) should be shlexed.
Fixes: #12452
Signed-off-by: Valentin Rothberg <rothberg@redhat.com>
|
|\ \ \
| | | |
| | | | |
More BeTrue cleanup
|
| | | |
| | | |
| | | |
| | | | |
Signed-off-by: Ed Santiago <santiago@redhat.com>
|