Often distributions to not have newuidmap and netgidmap configured
to be setuid. If Podman fails to setup the user namespace, check to
see if these files doe not have the proper protection and tell the user.
[NO NEW TESTS NEEDED]
Signed-off-by: Daniel J Walsh <dwalsh@redhat.com>
Update the login tests to reflect the latest changes to allow http{s}
prefixes (again) to address bugzilla.redhat.com/show_bug.cgi?id=2062072.
Signed-off-by: Valentin Rothberg <vrothberg@redhat.com>
Bumps
c/buildah to -> `v1.24.3-0.20220310160415-5ec70bf01ea5`
c/storage to -> `v1.38.3-0.20220308085612-93ce26691863`
c/image to -> `v5.20.1-0.20220310094651-0d8056ee346f`
Signed-off-by: Aditya R <arajan@redhat.com>
overlayfs -- the kernel's version, not fuse-overlayfs -- recently learned
(as of linux 5.16.0, I believe) how to support rootless users. Previously,
rootless users had to use these storage.conf(5) settings:
* storage.driver=vfs (aka STORAGE_DRIVER=vfs), or
* storage.driver=overlay (aka STORAGE_DRIVER=overlay),
storage.options.overlay.mount_program=/usr/bin/fuse-overlayfs
(aka STORAGE_OPTS=/usr/bin/fuse-overlayfs)
Now that a third backend is available, setting only:
* storage.driver=overlay (aka STORAGE_DRIVER=overlay)
https://github.com/containers/podman/issues/13123 reported EXDEV errors
during the normal operation of their container. Tracing it out, the
problem turned out to be that their container was being mounted without
'userxattr'; I don't fully understand why, but mount(8) mentions this is
needed for rootless users:
> userxattr
>
> Use the "user.overlay." xattr namespace instead of "trusted.overlay.".
> This is useful for unprivileged mounting of overlayfs.
https://github.com/containers/storage/pull/1156 found and fixed the issue
in podman, and this just pulls in that via
go get github.com/containers/storage@ebc90ab
go mod vendor
make vendor
Closes https://github.com/containers/podman/issues/13123
Signed-off-by: Nick Guenther <nick.guenther@polymtl.ca>
Since containers-common package is tied to specific versions
of Podman, add tools to build the package into the contrib directory
This should help other distributions to figure out which commont
package to ship.
Signed-off-by: Daniel J Walsh <dwalsh@redhat.com>
update c/storage to commit 26c561f9a64585d9a25d340e1ae5479eca8008a1.
It contains an important fix for partial pulls.
[NO NEW TESTS NEEDED]
Signed-off-by: Giuseppe Scrivano <gscrivan@redhat.com>
Bumps Buildah to v1.24.0 and adopts the new values for pull:
true, false, never, and always. The pull-never and pull-always options
for the build command are still usable, but they have been removed from
the man page documentation with this change.
Signed-off-by: tomsweeneyredhat <tsweeney@redhat.com>
This commit bumps majorly c/common so netavark features could be synced
with podman.
But there are some other vendor bumps as well
[NO NEW TESTS NEEDED]
[NO TESTS NEEDED]
Signed-off-by: Aditya R <arajan@redhat.com>
The libpod/network packages were moved to c/common so that buildah can
use it as well. To prevent duplication use it in podman as well and
remove it from here.
Signed-off-by: Paul Holzinger <pholzing@redhat.com>
Also update the e2e pull test to account for the changes when pulling
from the dir transport. Images pulled via the dir transport are not
tagged anymore; the path is not a reliable source.
Signed-off-by: Valentin Rothberg <rothberg@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>
The Docker-compatible REST API has historically behaved just as the rest
of Podman and Buildah (and the atomic Docker in older RHEL/Fedora) where
`containers-registries.conf` is centrally controlling which registries
a short name may resolve to during pull or local image lookups. Please
refer to a blog for more details [1].
Docker, however, is only resolving short names to docker.io which has
been reported (see #12320) to break certain clients who rely on this
behavior. In order to support this scenario, `containers.conf(5)`
received a new option to control whether Podman's compat API resolves
to docker.io only or behaves as before.
Most endpoints allow for directly normalizing parameters that represent
an image. If set in containers.conf, Podman will then normalize the
references directly to docker.io. The build endpoint is an outlier
since images are also referenced in Dockerfiles. The Buildah API,
however, supports specifying a custom `types.SystemContext` in which
we can set a field that enforces short-name resolution to docker.io
in `c/image/pkg/shortnames`.
Notice that this a "hybrid" approach of doing the normalization directly
in the compat endpoints *and* in `pkg/shortnames` by passing a system
context. Doing such a hybrid approach is neccessary since the compat
and the libpod endpoints share the same `libimage.Runtime` which makes
a global enforcement via the `libimage.Runtime.systemContext`
impossible. Having two separate runtimes for the compat and the libpod
endpoints seems risky and not generally applicable to all endpoints.
[1] https://www.redhat.com/sysadmin/container-image-short-namesFixes: #12320
Signed-off-by: Valentin Rothberg <rothberg@redhat.com>
For reasons buried in the history of Podman, looking up an untagged
image would match any tag of matching image. For instance, looking up
centos would match a local image centos:foobar. Change that behavior
to only match the latest tag.
Fix: #11964
Signed-off-by: Valentin Rothberg <rothberg@redhat.com>
Support downloading files, for instance via
`podman load -i server.com/image.tar`. The specified URL is downloaded
in the frontend and stored as a temp file that gets passed down to the
backend.
Also vendor in c/common@main to use the new `pkg/download`.
Fixes: #11970
Signed-off-by: Valentin Rothberg <rothberg@redhat.com>
Health checks may be defined in the container config or the config of an
image. So far, Podman only looked at the container config.
The plumbing happened in libimage but add a regression test to Podman as
well to make sure the glue code will not regress.
Note that I am pinning github.com/onsi/gomega to v1.16.0 since v1.17.0
requires go 1.16 which in turn is breaking CI.
Fixes: #12226
Signed-off-by: Valentin Rothberg <rothberg@redhat.com>
Make sure that the value is only set if specified on the CLI. c/image
already defaults to true but if set in the system context, we'd skip
settings in the registries.conf.
Fixes: #11933
Signed-off-by: Valentin Rothberg <rothberg@redhat.com>
