Improve the error handling of `container inspect` to properly handle
when the container has been removed _between_ the lookup and the
inspect. That will yield the correct "no such object" error message in
`inspect`.
[NO TESTS NEEDED] since I do not know have a reliable and cheap
reproducer. It's fixing a CI flake, so there's already an indicator.
Fixes: #11392
Signed-off-by: Valentin Rothberg <rothberg@redhat.com>
Commit 092902b455 introduced advanced network options for podman play
kube. However this never worked because it unconditionally set the
network mode to bridge after it parsed the network option.
Added a test to ensure the correct mode is set.
Truly fixes#10807
Signed-off-by: Paul Holzinger <pholzing@redhat.com>
When attempting to run remote builds, users with UID/GID values that
were high enough that they wouldn't be mapped into their default user
namespace configurations would see their builds fail when the server
attempted to extract the build contexts that they supplied, and failed
to set ownership of the build context content to the UID/GID that were
originally assigned to them.
When archiving the build context at the client, set ownership of
everything to 0:0, which we know is always mapped. Both ADD and COPY
require that we set the ownership of newly-added content to 0:0 (unless
the --chown flag is used), so throwing away the original ownership
information doesn't hurt, anyway. As usual, tarballs that we extract
as part of ADD aren't going to be affected.
Signed-off-by: Nalin Dahyabhai <nalin@redhat.com>
QEmu normally install BIOS images under `/usr/local` prefix, but
Homebrew installs them under `/opt/homebrew`. This change searches both
locations and then puts back to an unpathed name if it doesn't find the
BIOS. (I imitated other architectures' implemenations in that failback
behavior.)
[NO TESTS NEEDED]
Signed-off-by: Jonathan Springer <jonpspri@gmail.com>
Handle custom restart policies of containers when generating the unit
files; those should be set on the unit level and removed from ExecStart
flags.
Fixes: #11438
Signed-off-by: Valentin Rothberg <rothberg@redhat.com>
Prevent hitting `panic: runtime error: index out of range [1] with length 1`
while performing `podman info` when unexpected values for user.slice is found.
[NO TESTS NEEDED]
Signed-off-by: Aditya Rajan <arajan@redhat.com>
Adding the `-cpu host` option to the `addArchOptions` function for
darwin removes the warning message, "host doesn't support requested
feature: CPUID.80000001H:ECX.svm [bit 2]" by qemu-system-x86_64 when
using the `podman machine start` command on MacOS
Closes#11421
[NO TESTS NEEDED]
Signed-off-by: Scott Schreckengaust <scottschreckengaust@users.noreply.github.com>
Cause qemu to fall back to using TCG acceleration when HVP acceleration
is not available on Darwin Aarch64. Qemu prints a warning which it is
desirable to leave to embarrass the upstream Qemu into approving the HVF
patches.
[NO TESTS NEEDED]
Signed-off-by: Jonathan Springer <jspringer@us.ibm.com>
Signed-off-by: Jonathan Springer <jonpspri@gmail.com>
Following PR adds support for `kubectl` like `pod logs` to podman.
Usage `podman pod logs <podIDorName` gives a stream of logs for all
the containers within the pod with **containername** as a field.
Just like **`kubectl`** also supports `podman pod logs -c ctrIDorName podIDorName`
to limit the log stream to any of the specificied container which belongs to pod.
Signed-off-by: Aditya Rajan <arajan@redhat.com>
Equivalent for supported platforms, and makes it easier to support
additional unix-like OSes.
[NO TESTS NEEDED]
Signed-off-by: Maya Rashish <maya@NetBSD.org>
Filtering is missing in both compat API and libpod API, while docker
has filtering functinality. This commit enables filtering option using
name and id in both libpod and http API.
Signed-off-by: Jakub Guzik <jakubmguzik@gmail.com>
When net.Dial always fail in the above loop, the code following the loop
is executed. This error check prevents this.
[NO TESTS NEEDED]
Signed-off-by: Guillaume Rose <gurose@redhat.com>
Following PR makes sure that `podman manifest rm <list>` only removes
the named manifest list and not referenced images.
Bumping and squashing c/common to v0.43.3-0.20210902095222-a7acc160fb25
in same commit in order to make sure build commit test passes.
Signed-off-by: Aditya Rajan <arajan@redhat.com>
Creating the rootlessport socket can fail with `bind: invalid argument`
when the socket path is longer than 108 chars. This is the case for
users with a long runtime directory.
Since the kernel does not allow to use socket paths with more then 108
chars use a workaround to open the socket path.
[NO TESTS NEEDED]
Signed-off-by: Paul Holzinger <pholzing@redhat.com>
When a container is configured for auto removal podman stop should still
do cleanup, there is no guarantee the the cleanup process spawned by
conmon will be successful. Also a user expects after podman stop that
the network/mounts are cleaned up. Therefore podman stop should not return
early and instead do the cleanup and ignore errors if the container was
already removed.
[NO TESTS NEEDED] I don't know how to test this.
Fixes#11384
Signed-off-by: Paul Holzinger <pholzing@redhat.com>
`generate systemd --new` is looking at the "create command" of the
container/pod which is simply the os.Args at creation time.
It does not work on containers or pods created via the REST API since
the create command is not set. `--new` does work on such containers and
pods since there is no reliable way to reverse-map their configs to
command-line arguments of podman.
Fixes: #11370
Signed-off-by: Valentin Rothberg <rothberg@redhat.com>
to avoid segvs, we should clean up as much of the socket and regular
files from podman machine as possible on stop. also, on start, we
should add logic to remove these files before starting in case the start
process is stopped prematurely (due to an error for example).
[NO TESTS NEEDED]
Signed-off-by: Brent Baude <bbaude@redhat.com>
Make sure that the container's authfile label is used when pulling down
a new image.
[NO TESTS NEEDED] since it would require some larger rewrite of the
auto-update system tests that I currently have no time for. I added a
reminder to have some breadcrumbs when there is more time.
Fixes: #11171
Signed-off-by: Valentin Rothberg <rothberg@redhat.com>
InfraContainer should go through the same creation process as regular containers. This change was from the cmd level
down, involving new container CLI opts and specgen creating functions. What now happens is that both container and pod
cli options are populated in cmd and used to create a podSpecgen and a containerSpecgen. The process then goes as follows
FillOutSpecGen (infra) -> MapSpec (podOpts -> infraOpts) -> PodCreate -> MakePod -> createPodOptions -> NewPod -> CompleteSpec (infra) -> MakeContainer -> NewContainer -> newContainer -> AddInfra (to pod state)
Signed-off-by: cdoern <cdoern@redhat.com>
* Add response.Body.Close() where needed to release HTTP
connections to API server.
* Add tests to ensure no general leaks occur. 100% coverage would be
required to ensure no leaks on any call.
* Update code comments to be godoc correct
Signed-off-by: Jhon Honce <jhonce@redhat.com>
add the ability for play kube to tear down based on the yaml used to
play it. it is indicated by --down in the play kube command. volumes
are NOT deleted during the teardown. pods and their containers are
stopped and removed.
Signed-off-by: Brent Baude <bbaude@redhat.com>
