we need to allow users to expose ports to the host for the purposes
of networking, like a webserver. the port exposure must be done at
the time the pod is created.
strictly speaking, the port exposure occurs on the infra container.
Signed-off-by: baude <bbaude@redhat.com>
scope out new kube subcommand where we can add generate. you can now generate kubernetes
YAML that will allow you to run the container in a kubernetes environment. When
The YAML description will always "wrap" a container in a simple v1.Pod description.
Tests and further documentation will be added in additional PRs.
This function should be considered very much "under heavy development" at
this point.
Signed-off-by: baude <bbaude@redhat.com>
Use github.com/google/shlex for splitting commands instead of splitting
at whitespaces. This way, we avoid accidentally splitting single string
arguments into mutliple ones.
Signed-off-by: Valentin Rothberg <vrothberg@suse.com>
We now can remove a paused container by sending it a kill signal while it
is paused. We then unpause the container and it is immediately killed.
Also, reworked how the parallelWorker results are handled to provide a
more consistent approach to how each subcommand implements it. It also
fixes a bug where if one container errors, the error message is duplicated
when printed out.
Signed-off-by: baude <bbaude@redhat.com>
Since we use buildah containers for the build process, the
user will not know if we have any buildah containers lingering
due to a failed build. Setting this to true by default till
we figure out a better way to solve this.
Signed-off-by: Urvashi Mohnani <umohnani@redhat.com>
Addressing:
podman run -it -a STDERR --rm alpine /bin/ash
hanging. As we droped stdin as soon as -a was used. Notice this is contrary to
what D-tool does and contrary to what podman help implies:
podman run --help | grep interact
--interactive, -i Keep STDIN open even if not attached
Signed-off-by: Šimon Lukašík <slukasik@redhat.com>
Operations like kill, pause, and unpause -- which can operation on one or
more containers -- can greatly benefit from parallizing its main job (eq kill).
In the case of pauseand unpause, an --all option as was added. pause --all will
pause all **running** containers. And unpause --all will unpause all **paused**
containers.
Signed-off-by: baude <bbaude@redhat.com>
When attempting to restart many containers, we can benefit from making
the restarts parallel. For convenience, two new options are added:
--all attempts to restart all containers
--run-only when used with --all will attempt to restart only running containers
Signed-off-by: baude <bbaude@redhat.com>
It was setting the wrong variable (CamelCase)
in the wrong module ("main", not "libpod")...
Signed-off-by: Anders F Björklund <anders.f.bjorklund@gmail.com>
when the PS command was reworked for performance and formatting improvements,
i forgot to truncate the command field. Long container commands was throwing
the formatting off. we now truncated to 17 characters plus the elipses.
Signed-off-by: baude <bbaude@redhat.com>
for backwards compatibility and auto-test, we needed a few changes
that slipped in when i reworked ps to be faster to be reverted. the
follow behaviours were reverted:
1. the is_infra column was redacted. that appears to be a mistake on my
part.
2. a newline after ps prints its format was added
3. a newline prior to printing the headers was removed.
Signed-off-by: baude <bbaude@redhat.com>
Ed has asked that we revert to using two spaces for padding between PS fields. I assume
this is for docker autotests.
Signed-off-by: baude <bbaude@redhat.com>
add a global flag for --max-workers so users can limit the number
of parallel operations for a given function. also, when not limited
by max-workers, we implement a heuristic function that returns the
number of preferred parallel workers based on the number of CPUs and
the given operation.
Signed-off-by: baude <bbaude@redhat.com>
As discussed [1], the runlabel command should execute any command
specified in a label. The reasoning behind is that we cannot restrict
which options are passed to Podman which thereby has full access to the
host (runlabels must be used with care).
With the updated semantics, runlabel will substitute the commands with a
basepath equal to "docker" or "podman" with "/proc/self/exe", and
otherwise leave the command unchanged to execute any other command on
the host.
[1] https://github.com/containers/libpod/pull/1607#issuecomment-428321382
Signed-off-by: Valentin Rothberg <vrothberg@suse.com>
prevent opening the same file twice, since we re-exec podman in
rootless mode. While at it, also solve a possible race between the
check for the file and writing to it. Another process could have
created the file in the meanwhile and we would just end up overwriting
it.
Signed-off-by: Giuseppe Scrivano <gscrivan@redhat.com>
This removes duplicate code paths which has been previously factored out
as getAllOrLatestContainers().
Signed-off-by: Adrian Reber <areber@redhat.com>
Instead of duplicating the same code in multiple commands this uses the
newly added function checkAllAndLatest() instead.
Signed-off-by: Adrian Reber <areber@redhat.com>
Just as the checkAllAndLatest() function the new code in
getAllOrLatestContainers() is used in some commands and duplicated. This
factors out this code to be used in other places without duplicating it.
Signed-off-by: Adrian Reber <areber@redhat.com>
The check about the --all and --latest option is used and repeated and
some commands. Factor it out and put it into common.
Signed-off-by: Adrian Reber <areber@redhat.com>
Fix the parsing of environment variables to catch invalid ones, such as
`-e = ` or `-e =A`, early in the stack to return meaningful error
messages. Also, instead of erroring out, set unspecified env variables
as empty (e.g., `-e FOO`) to remain compatible with Docker.
Fixes: #1663
Signed-off-by: Valentin Rothberg <vrothberg@suse.com>
when doing stats -a|--all, if you have non-running containers, we should
not error on not being able to get information like PID, etc on them.
Signed-off-by: baude <bbaude@redhat.com>
When doing rm, we now parallelize the actual conainter deletions so they
can complete faster. This speeds up operations like rm -a.
Signed-off-by: baude <bbaude@redhat.com>
Running 'podman port -l' on a system without any containers created
gives:
$ podman port -l
panic: runtime error: invalid memory address or nil pointer dereference
[signal SIGSEGV: segmentation violation code=0x1 addr=0x10 pc=0xf3cef1]
goroutine 1 [running]:
github.com/containers/libpod/libpod.(*Container).State(0x0, 0x0, 0x0, 0x0)
/share/go/src/github.com/containers/libpod/libpod/container.go:658 +0x41
main.portCmd(0xc420094580, 0x0, 0x0)
/share/go/src/github.com/containers/libpod/cmd/podman/port.go:118 +0x406
This fixes it by making sure the variable 'containers' is nil and not [<nil>].
Signed-off-by: Adrian Reber <areber@redhat.com>
We were encountering sync issues with the map, so swap to a
thread-safe channel and convert into a map when we output
Signed-off-by: Matthew Heon <matthew.heon@gmail.com>
