Unfortunately --pre-checkpointing never worked as intended and recent
changes to runc have shown that it is broken.
To create a pre-checkpoint CRIU expects the paths between the
pre-checkpoints to be a relative path. If having a previous checkpoint
it needs the be referenced like this: --prev-images-dir ../parent
Unfortunately Podman was giving runc (and CRIU) an absolute path.
Unfortunately, again, until March 2021 CRIU silently ignored if
the path was not relative and switch back to normal checkpointing.
This has been now fixed in CRIU and runc and running pre-checkpoint
with the latest runc fails, because runc already sees that the path is
absolute and returns an error.
This commit fixes this by giving runc a relative path.
This commit also fixes a second pre-checkpointing error which was just
recently introduced.
So summarizing: pre-checkpointing never worked correctly because CRIU
ignored wrong parameters and recent changes broke it even more.
Now both errors should be fixed.
[NO TESTS NEEDED]
Signed-off-by: Adrian Reber <areber@redhat.com>
Signed-off-by: Adrian Reber <adrian@lisas.de>
When 127.0.0.53 is the only nameserver in /etc/resolv.conf assume
systemd-resolved is used. This is better because /etc/resolv.conf does
not have to be symlinked to /run/systemd/resolve/stub-resolv.conf in
order to use systemd-resolved.
[NO TESTS NEEDED]
Fixes: #10570
Signed-off-by: Paul Holzinger <pholzing@redhat.com>
The checkpoint archive compression was hardcoded to `archive.Gzip`.
There have been requests to make the used compression algorithm
selectable. There was especially the request to not compress the
checkpoint archive to be able to create faster checkpoints when not
compressing it.
This also changes the default from `gzip` to `zstd`. This change should
not break anything as the restore code path automatically handles
whatever compression the user provides during restore.
Signed-off-by: Adrian Reber <areber@redhat.com>
The containers /etc/resolv.conf allways preserved the ipv6 nameserves
from the host even when the container did not supported ipv6. Check
if the cni result contains an ipv6 address or slirp4netns has ipv6
support enabled and only add the ipv6 nameservers when this is the case.
The test needs to have an ipv6 nameserver in the hosts /etc/hosts but we
should never mess with this file on the host. Therefore the test is
skipped when no ipv6 is detected.
Fixes#10158
Signed-off-by: Paul Holzinger <pholzing@redhat.com>
Support UID, GID, Mode options for mount type secrets. Also, change
default secret permissions to 444 so all users can read secret.
Signed-off-by: Ashley Cui <acui@redhat.com>
This change adds the entry `host.containers.internal` to the `/etc/hosts`
file within a new containers filesystem. The ip address is determined by
the containers networking configuration and points to the gateway address
for the containers networking namespace.
Closes#5651
Signed-off-by: Baron Lenardson <lenardson.baron@gmail.com>
Docker allows relabeling of any volume passed in via -v, even
including named volumes. This normally isn't an issue at all,
given named volumes get the right label for container access
automatically, but this becomes an issue when volume plugins are
involved - these aren't managed by Podman, and may well be
unaware of SELinux labelling. We could automatically relabel
these volumes on creation, but I'm still reluctant to do that
(feels like it could break things). Instead, let's allow :z and
:Z to be used with named volumes, so users can explicitly request
relabel of a volume plugin-backed volume.
We also get :U at the same time. I don't see any real need for it
but it also doesn't seem to hurt, so I didn't bother disabling
it.
Fixes#10273
Signed-off-by: Matthew Heon <mheon@redhat.com>
Revert : https://github.com/containers/podman/pull/9895
Turns out that if Docker is in --selinux-enabeled, it still relabels if
the user tells the system to, even if running a --privileged container
or if the selinux separation is disabled --security-opt label=disable.
Signed-off-by: Daniel J Walsh <dwalsh@redhat.com>
Env var secrets are env vars that are set inside the container but not
commited to and image. Also support reading from env var when creating a
secret.
Signed-off-by: Ashley Cui <acui@redhat.com>
extend to pods the existing check whether the cgroup is usable when
running as rootless with cgroupfs.
commit 17ce567c68 introduced the
regression.
Signed-off-by: Giuseppe Scrivano <gscrivan@redhat.com>
Migrate the Podman code base over to `common/libimage` which replaces
`libpod/image` and a lot of glue code entirely.
Note that I tried to leave bread crumbs for changed tests.
Miscellaneous changes:
* Some errors yield different messages which required to alter some
tests.
* I fixed some pre-existing issues in the code. Others were marked as
`//TODO`s to prevent the PR from exploding.
* The `NamesHistory` of an image is returned as is from the storage.
Previously, we did some filtering which I think is undesirable.
Instead we should return the data as stored in the storage.
* Touched handlers use the ABI interfaces where possible.
* Local image resolution: previously Podman would match "foo" on
"myfoo". This behaviour has been changed and Podman will now
only match on repository boundaries such that "foo" would match
"my/foo" but not "myfoo". I consider the old behaviour to be a
bug, at the very least an exotic corner case.
* Futhermore, "foo:none" does *not* resolve to a local image "foo"
without tag anymore. It's a hill I am (almost) willing to die on.
* `image prune` prints the IDs of pruned images. Previously, in some
cases, the names were printed instead. The API clearly states ID,
so we should stick to it.
* Compat endpoint image removal with _force_ deletes the entire not
only the specified tag.
Signed-off-by: Valentin Rothberg <rothberg@redhat.com>
if --cgroup-parent is specified, always honor it without doing any
detection whether cgroups are supported or not.
Closes: https://github.com/containers/podman/issues/10173
Signed-off-by: Giuseppe Scrivano <gscrivan@redhat.com>
- Persist CDIDevices in container config
- Add e2e test
- Log HasDevice error and add additional condition for safety
Signed-off-by: Sebastian Jug <seb@stianj.ug>
do not set the cgroup parent when running as rootless with cgroupfs,
even if cgroup v2 is used.
Closes: https://bugzilla.redhat.com/show_bug.cgi?id=1947999
Signed-off-by: Giuseppe Scrivano <gscrivan@redhat.com>
Docker does not relabel this content, and openstack is running
containers in this manner. There is a penalty for doing this
on each container, that is not worth taking on a disable SELinux
container.
Signed-off-by: Daniel J Walsh <dwalsh@redhat.com>
Instead of creating an extra container create a network and mount
namespace inside the podman user namespace. This ns is used to
for rootless cni operations.
This helps to align the rootless and rootful network code path.
If we run as rootless we just have to set up a extra net ns and
initialize slirp4netns in it. The ocicni lib will be called in
that net ns.
This design allows allows easier maintenance, no extra container
with pause processes, support for rootless cni with --uidmap
and possibly more.
The biggest problem is backwards compatibility. I don't think
live migration can be possible. If the user reboots or restart
all cni containers everything should work as expected again.
The user is left with the rootless-cni-infa container and image
but this can safely be removed.
To make the existing cni configs work we need execute the cni plugins
in a extra mount namespace. This ensures that we can safely mount over
/run and /var which have to be writeable for the cni plugins without
removing access to these files by the main podman process. One caveat
is that we need to keep the netns files at `XDG_RUNTIME_DIR/netns`
accessible.
`XDG_RUNTIME_DIR/rootless-cni/{run,var}` will be mounted to `/{run,var}`.
To ensure that we keep the netns directory we bind mount this relative
to the new root location, e.g. XDG_RUNTIME_DIR/rootless-cni/run/user/1000/netns
before we mount the run directory. The run directory is mounted recursive,
this makes the netns directory at the same path accessible as before.
This also allows iptables-legacy to work because /run/xtables.lock is
now writeable.
Signed-off-by: Paul Holzinger <paul.holzinger@web.de>
Erik Sjolund reported an issue where a badly formated file
could be passed into the `--tz` option and then the date in the container
would be badly messed up:
```
erik@laptop:~$ echo Hello > file.txt
erik@laptop:~$ podman run --tz=../../../home/erik/file.txt --rm -ti
docker.io/library/alpine cat /etc/localtime
Hello
erik@laptop:~$ podman --version
podman version 3.0.0-rc1
erik@laptop:~$
```
This fix checks to make sure the TZ passed in is a valid
value and then proceeds with the rest of the processing.
This was first reported as a potential security issue, but it
was thought not to be. However, I thought closing the hole
sooner rather than later would be good.
Signed-off-by: TomSweeneyRedHat <tsweeney@redhat.com>
The --trace has helped in early stages analyze Podman code. However,
it's contributing to dependency and binary bloat. The standard go
tooling can also help in profiling, so let's turn `--trace` into a NOP.
[NO TESTS NEEDED]
Signed-off-by: Valentin Rothberg <rothberg@redhat.com>
To be able to reuse common checkpoint/restore functions this commit
moves code to pkg/checkpoint/crutils.
This commit has not functional changes. It only moves code around.
[NO TESTS NEEDED] - only moving code around
Signed-off-by: Adrian Reber <areber@redhat.com>
We missed bumping the go module, so let's do it now :)
* Automated go code with github.com/sirkon/go-imports-rename
* Manually via `vgrep podman/v2` the rest
Signed-off-by: Valentin Rothberg <rothberg@redhat.com>
Currently if the host shares container storage with a container
running podman, the podman inside of the container resets the
storage on the host. This can cause issues on the host, as
well as causes the podman command running the container, to
fail to unmount /dev/shm.
podman run -ti --rm --privileged -v /var/lib/containers:/var/lib/containers quay.io/podman/stable podman run alpine echo hello
* unlinkat /var/lib/containers/storage/overlay-containers/a7f3c9deb0656f8de1d107e7ddff2d3c3c279c11c1635f233a0bffb16051fb2c/userdata/shm: device or resource busy
* unlinkat /var/lib/containers/storage/overlay-containers/a7f3c9deb0656f8de1d107e7ddff2d3c3c279c11c1635f233a0bffb16051fb2c/userdata/shm: device or resource busy
Since podman is volume mounting in the graphroot, it will add a flag to
/run/.containerenv to tell podman inside of container whether to reset storage or not.
Since the inner podman is running inside of the container, no reason to assume this is a fresh reboot, so if "container" environment variable is set then skip
reset of storage.
Also added tests to make sure /run/.containerenv is runnig correctly.
Fixes: https://github.com/containers/podman/issues/9191
Signed-off-by: Daniel J Walsh <dwalsh@redhat.com>
Make sure to not set an empty $HOME for containers and let it default to
"/".
https://github.com/containers/crun/pull/599 is required to fully
address #9378.
Partially-Fixes: #9378
Signed-off-by: Valentin Rothberg <rothberg@redhat.com>
Currently podman is always chowning the WORKDIR to root:root
This PR will return if the WORKDIR already exists.
Fixes: https://github.com/containers/podman/issues/9387
Signed-off-by: Daniel J Walsh <dwalsh@redhat.com>
The old copy-up implementation was very unhappy with symlinks,
which could cause containers to fail to start for unclear reasons
when a directory we wanted to copy-up contained one. Rewrite to
use the Buildah Copier, which is more recent and should be both
safer and less likely to blow up over links.
At the same time, fix a deadlock in copy-up for volumes requiring
mounting - the Mountpoint() function tried to take the
already-acquired volume lock.
Fixes#6003
Signed-off-by: Matthew Heon <mheon@redhat.com>
Implement podman secret create, inspect, ls, rm
Implement podman run/create --secret
Secrets are blobs of data that are sensitive.
Currently, the only secret driver supported is filedriver, which means creating a secret stores it in base64 unencrypted in a file.
After creating a secret, a user can use the --secret flag to expose the secret inside the container at /run/secrets/[secretname]
This secret will not be commited to an image on a podman commit
Signed-off-by: Ashley Cui <acui@redhat.com>
When resolving the workdir of a container, we may need to create unless
the user set it explicitly on the command line. Otherwise, we just do a
presence check. Unfortunately, there was a missing return that lead us
to fall through into attempting to create and chown the workdir. That
caused a regression when running on a read-only root fs.
Fixes: #9230
Signed-off-by: Valentin Rothberg <rothberg@redhat.com>
A container's workdir can be specified via the CLI via `--workdir` and
via an image config with the CLI having precedence.
Since images have a tendency to specify workdirs without necessarily
shipping the paths with the root FS, make sure that Podman creates the
workdir. When specified via the CLI, do not create the path, but check
for its existence and return a human-friendly error.
NOTE: `crun` is performing a similar check that would yield exit code
127. With this change, however, Podman performs the check and yields
exit code 126. Since this is specific to `crun`, I do not consider it
to be a breaking change of Podman.
Fixes: #9040
Signed-off-by: Valentin Rothberg <rothberg@redhat.com>
This implements support for mounting and unmounting volumes
backed by volume plugins. Support for actually retrieving
plugins requires a pull request to land in containers.conf and
then that to be vendored, and as such is not yet ready. Given
this, this code is only compile tested. However, the code for
everything past retrieving the plugin has been written - there is
support for creating, removing, mounting, and unmounting volumes,
which should allow full functionality once the c/common PR is
merged.
A major change is the signature of the MountPoint function for
volumes, which now, by necessity, returns an error. Named volumes
managed by a plugin do not have a mountpoint we control; instead,
it is managed entirely by the plugin. As such, we need to cache
the path in the DB, and calls to retrieve it now need to access
the DB (and may fail as such).
Notably absent is support for SELinux relabelling and chowning
these volumes. Given that we don't manage the mountpoint for
these volumes, I am extremely reluctant to try and modify it - we
could easily break the plugin trying to chown or relabel it.
Also, we had no less than *5* separate implementations of
inspecting a volume floating around in pkg/infra/abi and
pkg/api/handlers/libpod. And none of them used volume.Inspect(),
the only correct way of inspecting volumes. Remove them all and
consolidate to using the correct way. Compat API is likely still
doing things the wrong way, but that is an issue for another day.
Fixes#4304
Signed-off-by: Matthew Heon <matthew.heon@pm.me>
`staticcheck` is a golang code analysis tool. https://staticcheck.io/
This commit fixes a lot of problems found in our code. Common problems are:
- unnecessary use of fmt.Sprintf
- duplicated imports with different names
- unnecessary check that a key exists before a delete call
There are still a lot of reported problems in the test files but I have
not looked at those.
Signed-off-by: Paul Holzinger <paul.holzinger@web.de>
writing to the id map fails when an extent overlaps multiple mappings
in the parent user namespace:
$ cat /proc/self/uid_map
0 1000 1
1 100000 65536
$ unshare -U sleep 100 &
[1] 1029703
$ printf "0 0 100\n" | tee /proc/$!/uid_map
0 0 100
tee: /proc/1029703/uid_map: Operation not permitted
This limitation is particularly annoying when working with rootless
containers as each container runs in the rootless user namespace, so a
command like:
$ podman run --uidmap 0:0:2 --rm fedora echo hi
Error: writing file `/proc/664087/gid_map`: Operation not permitted: OCI permission denied
would fail since the specified mapping overlaps the first
mapping (where the user id is mapped to root) and the second extent
with the additional IDs available.
Detect such cases and automatically split the specified mapping with
the equivalent of:
$ podman run --uidmap 0:0:1 --uidmap 1:1:1 --rm fedora echo hi
hi
A fix has already been proposed for the kernel[1], but even if it
accepted it will take time until it is available in a released kernel,
so fix it also in pkg/rootless.
[1] https://lkml.kernel.org/lkml/20201203150252.1229077-1-gscrivan@redhat.com/
Signed-off-by: Giuseppe Scrivano <gscrivan@redhat.com>
When migrating a container with associated volumes, the content of
these volumes should be made available on the destination machine.
This patch enables container checkpoint/restore with named volumes
by including the content of volumes in checkpoint file. On restore,
volumes associated with container are created and their content is
restored.
The --ignore-volumes option is introduced to disable this feature.
Example:
# podman container checkpoint --export checkpoint.tar.gz <container>
The content of all volumes associated with the container are included
in `checkpoint.tar.gz`
# podman container checkpoint --export checkpoint.tar.gz --ignore-volumes <container>
The content of volumes is not included in `checkpoint.tar.gz`. This is
useful, for example, when the checkpoint/restore is performed on the
same machine.
# podman container restore --import checkpoint.tar.gz
The associated volumes will be created and their content will be
restored. Podman will exit with an error if volumes with the same
name already exist on the system or the content of volumes is not
included in checkpoint.tar.gz
# podman container restore --ignore-volumes --import checkpoint.tar.gz
Volumes associated with container must already exist. Podman will not
create them or restore their content.
Signed-off-by: Radostin Stoyanov <rstoyanov@fedoraproject.org>
Instead of individual values from ContainerCheckpointOptions,
provide the options object.
This is a preparation for the next patch where one more value
of the options object is required in exportCheckpoint().
Signed-off-by: Radostin Stoyanov <rstoyanov@fedoraproject.org>
When adding the HOSTNAME environment variable, only do so if it
is not already present in the spec. If it is already present, it
was likely added by the user, and we should honor their requested
value.
Fixes#8886
Signed-off-by: Matthew Heon <mheon@redhat.com>
do not check whether the specified ID is valid in the user namespace.
crun handles this case[1], so the check in Podman prevents to get to
the OCI runtime at all.
$ podman run --user 10:0 --uidmap 0:0:1 --rm -ti fedora:33 sh -c 'id; cat /proc/self/uid_map'
uid=10(10) gid=0(root) groups=0(root),65534(nobody)
10 0 1
[1] https://github.com/containers/crun/pull/556
Signed-off-by: Giuseppe Scrivano <gscrivan@redhat.com>
This adds a new command, 'podman network reload', to reload the
networks of existing containers, forcing recreation of firewall
rules after e.g. `firewall-cmd --reload` wipes them out.
Under the hood, this works by calling CNI to tear down the
existing network, then recreate it using identical settings. We
request that CNI preserve the old IP and MAC address in most
cases (where the container only had 1 IP/MAC), but there will be
some downtime inherent to the teardown/bring-up approach. The
architecture of CNI doesn't really make doing this without
downtime easy (or maybe even possible...).
At present, this only works for root Podman, and only locally.
I don't think there is much of a point to adding remote support
(this is very much a local debugging command), but I think adding
rootless support (to kill/recreate slirp4netns) could be
valuable.
Signed-off-by: Matthew Heon <matthew.heon@pm.me>
Signed-off-by: Paul Holzinger <paul.holzinger@web.de>
We have been asked to leak some information into the container
to indicate:
* The name and id of the container
* The version of podman used to launch the container
* The image name and ID the container is based on.
* Whether the container engine is running in rootless mode.
Fixes: https://github.com/containers/podman/issues/6192
Signed-off-by: Daniel J Walsh <dwalsh@redhat.com>
Our users are missing certain warning messages that would
make debugging issues with Podman easier.
For example if you do a podman build with a Containerfile
that contains the SHELL directive, the Derective is silently
ignored.
If you run with the log-level warn you get a warning message explainging
what happened.
$ podman build --no-cache -f /tmp/Containerfile1 /tmp/
STEP 1: FROM ubi8
STEP 2: SHELL ["/bin/bash", "-c"]
STEP 3: COMMIT
--> 7a207be102a
7a207be102aa8993eceb32802e6ceb9d2603ceed9dee0fee341df63e6300882e
$ podman --log-level=warn build --no-cache -f /tmp/Containerfile1 /tmp/
STEP 1: FROM ubi8
STEP 2: SHELL ["/bin/bash", "-c"]
STEP 3: COMMIT
WARN[0000] SHELL is not supported for OCI image format, [/bin/bash -c] will be ignored. Must use `docker` format
--> 7bd96fd25b9
7bd96fd25b9f755d8a045e31187e406cf889dcf3799357ec906e90767613e95f
These messages will no longer be lost, when we default to WARNing level.
Signed-off-by: Daniel J Walsh <dwalsh@redhat.com>
The buildah/pkg/secrts package was move to
containers/common/pkg/subscriptions.
Switch to using this by default.
Signed-off-by: Daniel J Walsh <dwalsh@redhat.com>
When looking up a container's cgroup path, parse /proc/[PID]/cgroup.
This will work across all cgroup managers and configurations and is
supported on cgroups v1 and v2.
Fixes: #8265
Signed-off-by: Valentin Rothberg <rothberg@redhat.com>
The --hostname and containername should always be added to containers.
Added some tests to make sure you can always ping the hostname and container
name from within the container.
Fixes: https://github.com/containers/podman/issues/8095
Signed-off-by: Daniel J Walsh <dwalsh@redhat.com>
Allow users to specify unbindable on volume command line
Switch internal mounts to rprivate to help prevent leaks.
Signed-off-by: Daniel J Walsh <dwalsh@redhat.com>
Most of the builtin golang functions like os.Stat and
os.Open report errors including the file system object
path. We should not wrap these errors and put the file path
in a second time, causing stuttering of errors when they
get presented to the user.
This patch tries to cleanup a bunch of these errors.
Signed-off-by: Daniel J Walsh <dwalsh@redhat.com>
Add a new "image" mount type to `--mount`. The source of the mount is
the name or ID of an image. The destination is the path inside the
container. Image mounts further support an optional `rw,readwrite`
parameter which if set to "true" will yield the mount writable inside
the container. Note that no changes are propagated to the image mount
on the host (which in any case is read only).
Mounts are overlay mounts. To support read-only overlay mounts, vendor
a non-release version of Buildah.
Signed-off-by: Valentin Rothberg <rothberg@redhat.com>
If the resolv.conf file is empty we provide default dns servers.
If the file does not exists we error and don't create the
container. We should also provide the default entries in this
case. This is also what docker does.
Fixes#8089
Signed-off-by: Paul Holzinger <paul.holzinger@web.de>
This does not match Docker, which does not add hostname in this
case, but it seems harmless enough.
Fixes#8095
Signed-off-by: Matthew Heon <matthew.heon@pm.me>
When a container uses --net=host the default hostname is set to
the host's hostname. However, we were not creating any entries
in `/etc/hosts` despite having a hostname, which is incorrect.
This hostname, for Docker compat, will always be the hostname of
the host system, not the container, and will be assigned to IP
127.0.1.1 (not the standard localhost address).
Also, when `--hostname` and `--net=host` are both passed, still
use the hostname from `--hostname`, not the host's hostname (we
still use the host's hostname by default in this case if the
`--hostname` flag is not passed).
Fixes#8054
Signed-off-by: Matthew Heon <matthew.heon@pm.me>
We need to do a length check before we can access the
networkStatus slice by index to prevent a runtime panic.
Fixes#8026
Signed-off-by: Paul Holzinger <paul.holzinger@web.de>
Currently the HOME environment is set to /root if
the user does not override it.
Also walk the parent directories of users homedir
to see if it is volume mounted into the container,
if yes, then set it correctly.
Signed-off-by: Daniel J Walsh <dwalsh@redhat.com>
When we create a container, we assign a cgroup parent based on
the current cgroup manager in use. This parent is only usable
with the cgroup manager the container is created with, so if the
default cgroup manager is later changed or overridden, the
container will not be able to start.
To solve this, store the cgroup manager that created the
container in container configuration, so we can guarantee a
container with a systemd cgroup parent will always be started
with systemd cgroups.
Unfortunately, this is very difficult to test in CI, due to the
fact that we hard-code cgroup manager on all invocations of
Podman in CI.
Fixes#7830
Signed-off-by: Matthew Heon <matthew.heon@pm.me>
We do not populate the hostname field with the IP Address
when running within a user namespace.
Fixes https://github.com/containers/podman/issues/7490
Signed-off-by: Daniel J Walsh <dwalsh@redhat.com>
This commit is courtesy of
```
for f in $(git ls-files *.go | grep -v ^vendor/); do \
sed -i 's/\(errors\..*\)"Error /\1"error /' $f;
done
for f in $(git ls-files *.go | grep -v ^vendor/); do \
sed -i 's/\(errors\..*\)"Failed to /\1"failed to /' $f;
done
```
etc.
Self-reviewed using `git diff --word-diff`, found no issues.
Signed-off-by: Kir Kolyshkin <kolyshkin@gmail.com>
In case os.Open[File], os.Mkdir[All], ioutil.ReadFile and the like
fails, the error message already contains the file name and the
operation that fails, so there is no need to wrap the error with
something like "open %s failed".
While at it
- replace a few places with os.Open, ioutil.ReadAll with
ioutil.ReadFile.
- replace errors.Wrapf with errors.Wrap for cases where there
are no %-style arguments.
Signed-off-by: Kir Kolyshkin <kolyshkin@gmail.com>
check there are enough gids in the user namespace before adding
supplementary gids from /etc/group.
Follow-up for baede7cd27
Signed-off-by: Giuseppe Scrivano <gscrivan@redhat.com>
There is a risk here, that if the GID does not exists
within the User Namespace the container will fail to start.
This is only likely to happen in HPC Envioronments, and I think
we should add a field to disable it for this environment,
Added a FIXME for this issue.
We currently have this problem with running a rootfull container within
a user namespace, it will fail if the GID is not available.
I looked at potentially checking the usernamespace that you are assigned
to, but I believe this will be very difficult to code up and to figure out.
Signed-off-by: Daniel J Walsh <dwalsh@redhat.com>
The kernel will not allow you to modify existing mount flags on a volume
when bind mounting it to another place. Since /sys/fs/cgroup/systemd is
mounted noexec on the host, it needs to be mounted with the same flags
in the rootless container.
Signed-off-by: Daniel J Walsh <dwalsh@redhat.com>
To ensure that the user running in the container ahs a valid
entry in /etc/passwd so lookup functions for the current user
will not error, Podman previously began adding entries to the
passwd file. We did not, however, add entries to the group file,
and this created problems - our passwd entries included the group
the user is in, but said group might not exist. The solution is
to mirror our logic for /etc/passwd modifications to also edit
/etc/group in the container.
Unfortunately, this is not a catch-all solution. Our logic here
is only advanced enough to *add* to the group file - so if the
group already exists but we add a user not a part of it, we will
not modify that existing entry, and things remain inconsistent.
We can look into adding this later if we absolutely need to, but
it would involve adding significant complexity to this already
massively complicated function.
While we're here, address an edge case where Podman could add a
user or group whose UID overlapped with an existing user or
group.
Also, let's make users able to log into users we added. Instead
of generating user entries with an 'x' in the password field,
indicating they have an entry in /etc/shadow, generate a '*'
indicating the user has no password but can be logged into by
other means e.g. ssh key, su.
Fixes#7503Fixes#7389Fixes#7499
Signed-off-by: Matthew Heon <matthew.heon@pm.me>
Usage:
```
$ podman network create foo
$ podman run -d --name web --hostname web --network foo nginx:alpine
$ podman run --rm --network foo alpine wget -O - http://web.dns.podman
Connecting to web.dns.podman (10.88.4.6:80)
...
<h1>Welcome to nginx!</h1>
...
```
See contrib/rootless-cni-infra for the design.
Signed-off-by: Akihiro Suda <akihiro.suda.cz@hco.ntt.co.jp>
The system defaults /run to "exec" mode, and we default --read-only
mounts on /run to "exec", so --systemd should follow suit.
Signed-off-by: Daniel J Walsh <dwalsh@redhat.com>
We want to modify /etc/passwd to add an entry for the user in
question, but at the same time we don't want to require the
container provide a /etc/passwd (a container with a single,
statically linked binary and nothing else is perfectly fine and
should be allowed, for example). We could create the passwd file
if it does not exist, but if the container doesn't provide one,
it's probably better not to make one at all. Gate changes to
/etc/passwd behind a stat() of the file in the container
returning cleanly.
Fixes#7515
Signed-off-by: Matthew Heon <mheon@redhat.com>
We had a customer incident where they ran out of space on /run.
If you don't specify size, it will be still limited to 50% or memory
available in the cgroup the container is running in. If the cgroup is
unlimited then the /run will be limited to 50% of the total memory
on the system.
Also /run is mounted on the host as exec, so no reason for us to mount
it noexec.
Signed-off-by: Daniel J Walsh <dwalsh@redhat.com>
I used the wrong propagation first time around because I forgot
that rprivate is the default propagation. Oops. Switch to
rprivate so we're using the default.
Signed-off-by: Matthew Heon <mheon@redhat.com>
On cgroups v1 systems, we need to mount /sys/fs/cgroup/systemd
into the container. We were doing this with no explicit mount
propagation tag, which means that, under some circumstances, the
shared mount propagation could be chosen - which, combined with
the fact that we need a mount to mask
/sys/fs/cgroup/systemd/release_agent in the container, means we
would leak a never-ending set of mounts under
/sys/fs/cgroup/systemd/ on container restart.
Fortunately, the fix is very simple - hardcode mount propagation
to something that won't leak.
Signed-off-by: Matthew Heon <mheon@redhat.com>
A recent crun change stopped the creation of the container's
working directory if it does not exist. This is arguably correct
for user-specified directories, to protect against typos; it is
definitely not correct for image WORKDIR, where the image author
definitely intended for the directory to be used.
This makes Podman create the working directory and chown it to
container root, if it does not already exist, and only if it was
specified by an image, not the user.
Signed-off-by: Matthew Heon <matthew.heon@pm.me>
Bind-mounting /etc/passwd into the container is problematic
becuase of how system utilities like `useradd` work. They want
to make a copy and then rename to try to prevent breakage; this
is, unfortunately, impossible when the file they want to rename
is a bind mount. The current behavior is fine for read-only
containers, though, because we expect useradd to fail in those
cases.
Instead of bind-mounting, we can edit /etc/passwd in the
container's rootfs. This is kind of gross, because the change
will show up in `podman diff` and similar tools, and will be
included in images made by `podman commit`. However, it's a lot
better than breaking important system tools.
Fixes#6953
Signed-off-by: Matthew Heon <matthew.heon@pm.me>
Currently you can not apply an ApparmorProfile if you specify
--privileged. This patch will allow both to be specified
simultaniosly.
By default Apparmor should be disabled if the user
specifies --privileged, but if the user specifies --security apparmor:PROFILE,
with --privileged, we should do both.
Added e2e run_apparmor_test.go
Signed-off-by: Daniel J Walsh <dwalsh@redhat.com>
--umask sets the umask inside the container
Defaults to 0022
Co-authored-by: Daniel J Walsh <dwalsh@redhat.com>
Signed-off-by: Ashley Cui <acui@redhat.com>
We added code to create a `/etc/passwd` file that we bind-mount
into the container in some cases (most notably,
`--userns=keep-id` containers). This, unfortunately, was not
persistent, so user-added users would be dropped on container
restart. Changing where we store the file should fix this.
Further, we want to ensure that lookups of users in the container
use the right /etc/passwd if we replaced it. There was already
logic to do this, but it only worked for user-added mounts; it's
easy enough to alter it to use our mounts as well.
Signed-off-by: Matthew Heon <matthew.heon@pm.me>
This was inspired by https://github.com/cri-o/cri-o/pull/3934 and
much of the logic for it is contained there. However, in brief,
a named return called "err" can cause lots of code confusion and
encourages using the wrong err variable in defer statements,
which can make them work incorrectly. Using a separate name which
is not used elsewhere makes it very clear what the defer should
be doing.
As part of this, remove a large number of named returns that were
not used anywhere. Most of them were once needed, but are no
longer necessary after previous refactors (but were accidentally
retained).
Signed-off-by: Matthew Heon <matthew.heon@pm.me>
If I enter a continer with --userns keep-id, my UID will be present
inside of the container, but most likely my user will not be defined.
This patch will take information about the user and stick it into the
container.
Signed-off-by: Daniel J Walsh <dwalsh@redhat.com>
With the advent of Podman 2.0.0 we crossed the magical barrier of go
modules. While we were able to continue importing all packages inside
of the project, the project could not be vendored anymore from the
outside.
Move the go module to new major version and change all imports to
`github.com/containers/libpod/v2`. The renaming of the imports
was done via `gomove` [1].
[1] https://github.com/KSubedi/gomove
Signed-off-by: Valentin Rothberg <rothberg@redhat.com>
--tz flag sets timezone inside container
Can be set to IANA timezone as well as `local` to match host machine
Signed-off-by: Ashley Cui <acui@redhat.com>
When running under systemd there is no need to create yet another
cgroup for the container.
With conmon-delegated the current cgroup will be split in two sub
cgroups:
- supervisor
- container
The supervisor cgroup will hold conmon and the podman process, while
the container cgroup is used by the OCI runtime (using the cgroupfs
backend).
Closes: https://github.com/containers/libpod/issues/6400
Signed-off-by: Giuseppe Scrivano <gscrivan@redhat.com>
This will allow containers that connect to the network namespace be
able to use the container name directly.
For example you can do something like
podman run -ti --name foobar fedora ping foobar
While we can do this with hostname now, this seems more natural.
Also if another container connects on the network to this container it
can do
podman run --network container:foobar fedora ping foobar
And connect to the original container,without having to discover the name.
Signed-off-by: Daniel J Walsh <dwalsh@redhat.com>
fix the check for c.state.NetNS == nil. Its value is changed in the
first code block, so the condition is always true in the second one
and we end up running slirp4netns twice.
Closes: https://github.com/containers/libpod/issues/6538
Signed-off-by: Giuseppe Scrivano <gscrivan@redhat.com>
do not set the hostname when joining an UTS namespace, as it could be
owned by a different userns.
Signed-off-by: Giuseppe Scrivano <gscrivan@redhat.com>
when running in a new userns, make sure the resolv.conf and hosts
files bind mounted from another container are accessible to root in
the userns.
Signed-off-by: Giuseppe Scrivano <gscrivan@redhat.com>
In FIPS Mode we expect to work off of the Mountpath not the Rundir path.
This is causing FIPS Mode checks to fail.
Signed-off-by: Daniel J Walsh <dwalsh@redhat.com>
Add more default options parsing
Switch to using --time as opposed to --timeout to better match Docker.
Signed-off-by: Daniel J Walsh <dwalsh@redhat.com>
We previously attempted to work within CNI to do this, without
success. So let's do it manually, instead. We know where the
files should live, so we can remove them ourselves instead. This
solves issues around sudden reboots where containers do not have
time to fully tear themselves down, and leave IP address
allocations which, for various reasons, are not stored in tmpfs
and persist through reboot.
Fixes#5433
Signed-off-by: Matthew Heon <matthew.heon@pm.me>
This corrects a regression from Podman 1.4.x where container exec
sessions inherited supplemental groups from the container, iff
the exec session did not specify a user.
Signed-off-by: Matthew Heon <matthew.heon@pm.me>
Enables most of the network-related functionality from
`podman run` in `podman pod create`. Custom CNI networks can be
specified, host networking is supported, DNS options can be
configured.
Also enables host networking in `podman play kube`.
Fixes#2808Fixes#3837Fixes#4432Fixes#4718Fixes#4770
Signed-off-by: Matthew Heon <matthew.heon@pm.me>
`gocritic` is a powerful linter that helps in preventing certain kinds
of errors as well as enforcing a coding style.
Signed-off-by: Valentin Rothberg <rothberg@redhat.com>
When doing a checkpoint with --export the root file-system diff was not
working as expected. Instead of getting the changes from the running
container to the highest storage layer it got the changes from the
highest layer to that parent's layer. For a one layer container this
could mean that the complete root file-system is part of the checkpoint.
With this commit this changes to use the same functionality as 'podman
diff'. This actually enables to correctly diff the root file-system
including tracking deleted files.
This also removes the non-working helper functions from libpod/diff.go.
Signed-off-by: Adrian Reber <areber@redhat.com>
The code currently assumes that the container we delegate network
namespace to will never further delegate to another container, so
when looking up things like /etc/hosts and /etc/resolv.conf we
won't pull the correct files from the chained dependency. The
changes to resolve this are relatively simple - just need to keep
looking until we find a container without NetNsCtr set.
Fixes#4626
Signed-off-by: Matthew Heon <matthew.heon@pm.me>
Trying to checkpoint a container started with --rm works, but it makes
no sense as the container, including the checkpoint, will be deleted
after writing the checkpoint. This commit inhibits checkpointing
containers started with '--rm' unless '--export' is used. If the
checkpoint is exported it can easily be restored from the exported
checkpoint, even if '--rm' is used. To restore a container from a
checkpoint it is even necessary to manually run 'podman rm' if the
container is not started with '--rm'.
Signed-off-by: Adrian Reber <areber@redhat.com>
When Libpod removes a container, there is the possibility that
removal will not fully succeed. The most notable problems are
storage issues, where the container cannot be removed from
c/storage.
When this occurs, we were faced with a choice. We can keep the
container in the state, appearing in `podman ps` and available for
other API operations, but likely unable to do any of them as it's
been partially removed. Or we can remove it very early and clean
up after it's already gone. We have, until now, used the second
approach.
The problem that arises is intermittent problems removing
storage. We end up removing a container, failing to remove its
storage, and ending up with a container permanently stuck in
c/storage that we can't remove with the normal Podman CLI, can't
use the name of, and generally can't interact with. A notable
cause is when Podman is hit by a SIGKILL midway through removal,
which can consistently cause `podman rm` to fail to remove
storage.
We now add a new state for containers that are in the process of
being removed, ContainerStateRemoving. We set this at the
beginning of the removal process. It notifies Podman that the
container cannot be used anymore, but preserves it in the DB
until it is fully removed. This will allow Remove to be run on
these containers again, which should successfully remove storage
if it fails.
Fixes#3906
Signed-off-by: Matthew Heon <mheon@redhat.com>
When restoring a container with user namespace, the user namespace is
created by the OCI runtime, and the network namespace is created after
the user namespace to ensure correct ownership.
In this case PostConfigureNetNS will be set and the value of
c.state.NetNS would be nil. Hence, the following error occurs:
$ sudo podman run --name cr \
--uidmap 0:1000:500 \
-d docker.io/library/alpine \
/bin/sh -c 'i=0; while true; do echo $i; i=$(expr $i + 1); sleep 1; done'
$ sudo podman container checkpoint cr
$ sudo podman container restore cr
...
panic: runtime error: invalid memory address or nil pointer dereference
[signal SIGSEGV: segmentation violation code=0x1 addr=0x30 pc=0x13a5e3c]
Signed-off-by: Radostin Stoyanov <rstoyanov1@gmail.com>
Pull in changes to pkg/secrets/secrets.go that adds the
logic to disable fips mode if a pod/container has a
label set.
Signed-off-by: Urvashi Mohnani <umohnani@redhat.com>
Refactor the `RuntimeConfig` along with related code from libpod into
libpod/config. Note that this is a first step of consolidating code
into more coherent packages to make the code more maintainable and less
prone to regressions on the long runs.
Some libpod definitions were moved to `libpod/define` to resolve
circular dependencies.
Signed-off-by: Valentin Rothberg <rothberg@redhat.com>
when running in systemd mode on cgroups v1, make sure the
/sys/fs/cgroup/systemd/release_agent is masked otherwise the container
is able to modify it and execute scripts on the host.
Signed-off-by: Giuseppe Scrivano <gscrivan@redhat.com>
Previously, `podman checkport restore` with exported containers,
when told to create a new container based on the exported
checkpoint, would create a new container, with a new container
ID, but not reset CGroup path - which contained the ID of the
original container.
If this was done multiple times, the result was two containers
with the same cgroup paths. Operations on these containers would
this have a chance of crossing over to affect the other one; the
most notable was `podman rm` once it was changed to use the --all
flag when stopping the container; all processes in the cgroup,
including the ones in the other container, would be stopped.
Reset cgroups on restore to ensure that the path matches the ID
of the container actually being run.
Signed-off-by: Matthew Heon <matthew.heon@pm.me>
For future work, we need multiple implementations of the OCI
runtime, not just a Conmon-wrapped runtime matching the runc CLI.
As part of this, do some refactoring on the interface for exec
(move to a struct, not a massive list of arguments). Also, add
'all' support to Kill and Stop (supported by runc and used a bit
internally for removing containers).
Signed-off-by: Matthew Heon <matthew.heon@pm.me>
CNI expects that a DELETE be run before re-creating container
networks. If a reboot occurs quickly enough that containers can't
stop and clean up, that DELETE never happens, and Podman
currently wipes the old network info and thinks the state has
been entirely cleared. Unfortunately, that may not be the case on
the CNI side. Some things - like IP address reservations - may
not have been cleared.
To solve this, manually re-run CNI Delete on refresh. If the
container has already been deleted this seems harmless. If not,
it should clear lingering state.
Fixes: #3759
Signed-off-by: Matthew Heon <matthew.heon@pm.me>
In order to run Podman with VM-based runtimes unprivileged, the
network must be set up prior to the container creation. Therefore
this commit modifies Podman to run rootless containers by:
1. create a network namespace
2. pass the netns persistent mount path to the slirp4netns
to create the tap inferface
3. pass the netns path to the OCI spec, so the runtime can
enter the netns
Closes#2897
Signed-off-by: Gabi Beyer <gabrielle.n.beyer@intel.com>
If the HOME environment variable is not set, make sure it is set to
the configuration found in the container /etc/passwd file.
It was previously depending on a runc behavior that always set HOME
when it is not set. The OCI runtime specifications do not require
HOME to be set so move the logic to libpod.
Closes: https://github.com/debarshiray/toolbox/issues/266
Signed-off-by: Giuseppe Scrivano <gscrivan@redhat.com>
When --cgroupns=private is used we need to mount a new cgroup file
system so that it points to the correct namespace.
Needs: https://github.com/containers/crun/pull/88
Signed-off-by: Giuseppe Scrivano <gscrivan@redhat.com>
This is mostly used with Systemd, which really wants to manage
CGroups itself when managing containers via unit file.
Signed-off-by: Matthew Heon <matthew.heon@pm.me>
This change adds the following annotation to every container created by
podman:
```json
"Annotations": {
"io.containers.manager": "libpod"
}
```
Target of this annotaions is to indicate which project in the containers
ecosystem is the major manager of a container when applications share
the same storage paths. This way projects can decide if they want to
manipulate the container or not. For example, since CRI-O and podman are
not using the same container library (libpod), CRI-O can skip podman
containers and provide the end user more useful information.
A corresponding end-to-end test has been adapted as well.
Relates to: https://github.com/cri-o/cri-o/pull/2761
Signed-off-by: Sascha Grunert <sgrunert@suse.com>
Previously, we only did this for volumes created at the same time
as the container. However, this is not correct behavior - Docker
does so for all named volumes, even those made with
'podman volume create' and mounted into a container later.
Fixes#3945
Signed-off-by: Matthew Heon <matthew.heon@pm.me>
When we fail to remove a container's SHM, that's an error, and we
need to report it as such. This may be part of our lingering
storage woes.
Also, remove MNT_DETACH. It may be another cause of the storage
removal failures.
Signed-off-by: Matthew Heon <matthew.heon@pm.me>
When volume options and the local volume driver are specified,
the volume is intended to be mounted using the 'mount' command.
Supported options will be used to volume the volume before the
first container using it starts, and unmount the volume after the
last container using it dies.
This should work for any local filesystem, though at present I've
only tested with tmpfs and btrfs.
Signed-off-by: Matthew Heon <matthew.heon@pm.me>
when cni returns a list of dns servers, we should add them under the
right conditions. the defined conditions are as follows:
- if the user provides dns, it and only it are added.
- if not above and you get a cni name server, it is added and a
forwarding dns instance is created for what was in resolv.conf.
- if not either above, the entries from the host's resolv.conf are used.
Signed-off-by: baude <bbaude@redhat.com>
Signed-off-by: baude <bbaude@redhat.com>
commit 223fe64dc0 introduced the
regression.
When running on cgroups v1, bind mount only /sys/fs/cgroup/systemd as
rw, as the code did earlier.
Also, simplify the rootless code as it doesn't require any special
handling when using --systemd.
Closes: https://bugzilla.redhat.com/show_bug.cgi?id=1737554
Signed-off-by: Giuseppe Scrivano <gscrivan@redhat.com>
If a container is restored multiple times from an exported checkpoint
with the help of '--import --name', the restore will fail if during
'podman run' a static container IP was set with '--ip'. The user can
tell the restore process to ignore the static IP with
'--ignore-static-ip'.
Signed-off-by: Adrian Reber <areber@redhat.com>
when running on a cgroups v2 system, do not bind mount
the named hierarchy /sys/fs/cgroup/systemd as it doesn't exist
anymore. Instead bind mount the entire /sys/fs/cgroup.
Signed-off-by: Giuseppe Scrivano <gscrivan@redhat.com>
Daniel J Walsh
Adrian Reber
Paul Holzinger
OpenShift Merge Robot
Ashley Cui
Baron Lenardson
Matthew Heon
Giuseppe Scrivano
Valentin Rothberg
Sebastian Jug
Paul Holzinger
なつき
TomSweeneyRedHat
Eduardo Vega
Matthew Heon
zhangguanzhang
unknown
Radostin Stoyanov
Josh Soref
Andy Librian
Kir Kolyshkin
Akihiro Suda
Qi Wang
Radostin Stoyanov
Jakub Filak
Urvashi Mohnani
Gabi Beyer
Sascha Grunert
baude