We were never passing the arguments to add the PAUSE_IMAGE to the
rootfs, leading to it never being present in the confidential image /
initrd.
Fixes: #9032 -- part II
Signed-off-by: Fabiano Fidêncio <fabiano.fidencio@intel.com>
Avoid error when building for amd64 using:
USE_CACHE=no AGENT_POLICY=yes DEBUG=1 \
tools/packaging/kata-deploy/local-build/kata-deploy-binaries.sh \
--build=rootfs-image
Fixes: #9067
Signed-off-by: Dan Mihai <dmihai@microsoft.com>
This env ver will serve us to pass the pause image tarball to the rootfs builder, which will then just
unpack the content into the rootfs.
Signed-off-by: ChengyuZhu6 <chengyu.zhu@intel.com>
Co-authored-by: Fabiano Fidêncio <fabiano.fidencio@intel.com>
Co-authored-by: Wang, Arron <arron.wang@intel.com>
Co-authored-by: stevenhorsman <steven@uk.ibm.com>
Co-authored-by: Jakob Naucke <jakob.naucke@ibm.com>
This is required as the tdx-attest-rs crate, which is used as part of
the guest components, has a runtime dependency on libattest-tdx.
Fixes: #9021 -- part II
Signed-off-by: Fabiano Fidêncio <fabiano.fidencio@intel.com>
This will make our lives easier to figure out whether the components are
being installed or not.
Signed-off-by: Fabiano Fidêncio <fabiano.fidencio@intel.com>
We need to pass this to the container where the rootfs is built, so it
can actually be unpacked inside the rootfs.
Signed-off-by: Fabiano Fidêncio <fabiano.fidencio@intel.com>
We need to actually check that the env var is not empty. :-)
This was introduced by 8307718842.
Signed-off-by: Fabiano Fidêncio <fabiano.fidencio@intel.com>
This env ver will serve us to pass the Confidential Containers
guest-components tarball to the rootfs builder, which will then just
unpack the content into the rootfs.
Fixes: #8848 -- part I
Signed-off-by: Fabiano Fidêncio <fabiano.fidencio@intel.com>
Signed-off-by: Linda Yu <linda.yu@intel.com>
Co-authored-by: stevenhorsman <steven@uk.ibm.com>
Co-authored-by: Jakob Naucke <jakob.naucke@ibm.com>
Co-authored-by: Wang, Arron <arron.wang@intel.com>
Co-authored-by: zhouliang121 <liang.a.zhou@linux.alibaba.com>
Co-authored-by: Alex Carter <alex.carter@ibm.com>
Co-authored-by: Suraj Deshmukh <suraj.deshmukh@microsoft.com>
Co-authored-by: Xynnn007 <xynnn@linux.alibaba.com>
As we'll be untarring the agent tarball (and any other component that
may be part of the rootfs) into the rootfs, we have to have xz
installed.
For debian and ubuntu the package is called xz-utils; for centos,
alpine and cbl-mariner the package is called xz.
Signed-off-by: Fabiano Fidêncio <fabiano.fidencio@intel.com>
This env var will serve us to pass the agent tarball to the rootfs
builder, which will then just unpack the content into the rootfs instead
of building the agent again.
AGENT_TARBALL and AGENT_SOURCE_BIN should never be used together.
Signed-off-by: Fabiano Fidêncio <fabiano.fidencio@intel.com>
The versions.yaml has a default for the amd64 binary, but there is no
code to actually build the arm64 binary, which seems an overlook.
Let's simplify the OPA logic by removing the direct link to the binary,
and construct that link as part of the checks we do to decide whether we
need to build OPA or not.
Fixes: #8373
Signed-off-by: Zvonko Kaiser <zkaiser@nvidia.com>
Signed-off-by: Fabiano Fidêncio <fabiano.fidencio@intel.com>
When building with AGENT_POLICY=yes and AGENT_INIT=yes:
1. Include OPA and the Policy settings in rootfs.
2. Start OPA from the kata agent.
Before these changes, building with both AGENT_POLICY=yes and
AGENT_INIT=yes was unsupported.
Starting OPA from systemd (when AGENT_INIT=no) was already supported.
Fixes: #7615
Signed-off-by: Dan Mihai <dmihai@microsoft.com>
Fixes: #7573
To enable this feature, build your rootfs using AGENT_POLICY=yes. The
default is AGENT_POLICY=no.
Building rootfs using AGENT_POLICY=yes has the following effects:
1. The kata-opa service gets included in the Guest image.
2. The agent gets built using AGENT_POLICY=yes.
After this patch, the shim calls SetPolicy if and only if a Policy
annotation is attached to the sandbox/pod. When creating a sandbox/pod
that doesn't have an attached Policy annotation:
1. If the agent was built using AGENT_POLICY=yes, the new sandbox uses
the default agent settings, that might include a default Policy too.
2. If the agent was built using AGENT_POLICY=no, the new sandbox is
executed the same way as before this patch.
Any SetPolicy calls from the shim to the agent fail if the agent was
built using AGENT_POLICY=no.
If the agent was built using AGENT_POLICY=yes:
1. The agent reads the contents of a default policy file during sandbox
start-up.
2. The agent then connects to the OPA service on localhost and sends
the default policy to OPA.
3. If the shim calls SetPolicy:
a. The agent checks if SetPolicy is allowed by the current
policy (the current policy is typically the default policy
mentioned above).
b. If SetPolicy is allowed, the agent deletes the current policy
from OPA and replaces it with the new policy it received from
the shim.
A typical new policy from the shim doesn't allow any future SetPolicy
calls.
4. For every agent rpc API call, the agent asks OPA if that call
should be allowed. OPA allows or not a call based on the current
policy, the name of the agent API, and the API call's inputs. The
agent rejects any calls that are rejected by OPA.
When building using AGENT_POLICY_DEBUG=yes, additional Policy logging
gets enabled in the agent. In particular, information about the inputs
for agent rpc API calls is logged in /tmp/policy.txt, on the Guest VM.
These inputs can be useful for investigating API calls that might have
been rejected by the Policy. Examples:
1. Load a failing policy file test1.rego on a different machine:
opa run --server --addr 127.0.0.1:8181 test1.rego
2. Collect the API inputs from Guest's /tmp/policy.txt and test on the
machine where the failing policy has been loaded:
curl -X POST http://localhost:8181/v1/data/agent_policy/CreateContainerRequest \
--data-binary @test1-inputs.json
Signed-off-by: Dan Mihai <dmihai@microsoft.com>
It's too long a time to cross build agent based on docker buildx, thus
we cross build rootfs based on a container with cross compile toolchain
of gcc and rust with musl libc. Then we get fast build just like native
build.
rootfs initrd cross build is disabled as no cross compile tolchain for
rust with musl lib if found for alpine and based on docker buildx takes
too long a time.
Fixes: #6557
Signed-off-by: Jianyong Wu <jianyong.wu@arm.com>
Add required kernel config for dm-crypt/dm-integrity/dm-verity
and related crypto config.
Add userspace command line tools for disk encryption support
and ext4 file system utilities.
Fixes: #6674
Signed-off-by: Arron Wang <arron.wang@intel.com>
This PR replaces single spaces to tabs in order to fix the
indentation of the rootfs script.
Fixes#6848
Signed-off-by: Gabriela Cervantes <gabriela.cervantes.tellez@intel.com>
- D-Bus enabling now occurs only in setup_rootfs (instead of
prepare_overlay and setup_rootfs)
- Adjust permissions of / so dbus-broker will be able to traverse FS
These changes enables kata-agent to successfully communicate with D-Bus.
Fixes#6677
Signed-off-by: Vladimir <amigo.elite@gmail.com>
The agent now offloads cgroup configuration to systemd when
possible. This requires to enable D-Bus in order to communicate
with systemd.
Fixes#6657
Signed-off-by: Greg Kurz <groug@kaod.org>
Add osbuilder support to build a rootfs and image
based on the CBL-Mariner Linux distro
Fixes: #6462
Signed-off-by: Dallas Delaney <dadelan@microsoft.com>
When starting an initrd the kernel expects to find /dev/console in the initrd,
so that it can connect it as stdin/stdout/stderr to the /init process. If the
device node is missing the kernel will complain that it was unable to open an
initial console. If kata-agent is the initrd init process, it will also result
in log messages not being logged to console and thus not forwarded to host
syslog.
Add a set of standard device nodes for completeness, so that console logging
works. To do that we install the makedev packge which provides a MAKEDEV helper
that knows the major/minor numbers. Unfortunately the debian package tries to
create devnodes from postinst, which can be suppressed if systemd-detect-virt
is present. That's why we create a small dummy script that matches what
systemd-detect-virt would output (anything is enough to suppress mknod).
Fixes: #6261
Signed-off-by: Jeremi Piotrowski <jpiotrowski@microsoft.com>
Create a guest image to support SELinux for containers inside the guest
if `SELINUX=yes` is specified. This works only if the guest rootfs is
CentOS and the init service is systemd, not the agent init. To enable
labeling the guest image on the host, selinuxfs must be mounted on the
host. The kata-agent will be labeled as `container_runtime_exec_t` type.
Fixes: #4812
Signed-off-by: Manabu Sugimoto <Manabu.Sugimoto@sony.com>
1. Implemented a rust module for operating cgroups through systemd with the help of zbus (src/agent/rustjail/src/cgroups/systemd).
2. Add support for optional cgroup configuration through fs and systemd at agent (src/agent/rustjail/src/container.rs).
3. Described the usage and supported properties of the agent systemd cgroup (docs/design/agent-systemd-cgroup.md).
Fixes: #4336
Signed-off-by: Yuan-Zhuo <yuanzhuo0118@outlook.com>
To avoid the random failures when we are building the rootfs as it seems
that it does not find the value for the libseccomp and gperf directory,
this PR export these variables.
Fixes#5232
Signed-off-by: Gabriela Cervantes <gabriela.cervantes.tellez@intel.com>
Guest log is showing a hang on systemd getty start.
Adding symlink for /dev/ttyS0 resolves issue.
Fixes: #4932
Signed-Off-By: Ryan Savino <ryan.savino@amd.com>
It would be nice to use `versions.yaml` for the maintainability.
Previously, we have been specified the `libseccomp` and the `gperf` version
directly in this script without using the `versions.yaml` because the current
snap workflow is incomplete and fails.
This is because snap CI environment does not have kata-cotnainers repository
under ${GOPATH}. To avoid the failure, the `rootfs.sh` extracts the libseccomp
version and url in advance and pass them to the `install_libseccomp.sh` as
environment variables.
Fixes: #4941
Signed-off-by: Manabu Sugimoto <Manabu.Sugimoto@sony.com>
Since we are introducing an agent API for interacting with guest
iptables, let's ensure that our example rootfs' have iptables-save/restore
installed.
Fixes: #4356
Signed-off-by: Eric Ernst <eric_ernst@apple.com>
This PR removes the clear containers reference as this is not longer
being used and is deprecated at the rootfs builder README.
Fixes#4278
Signed-off-by: Gabriela Cervantes <gabriela.cervantes.tellez@intel.com>
In at least kata versions 2.3.3 and 2.4.0 it was noticed that the guest
operating system's clock would drift out of sync slowly over time
whilst the pod was running.
This had previously been raised and fixed in the old reposity via [1].
In essence kvm_ptp and chrony were paired together in order to
keep the system clock up to date with the host.
In the recent versions of kata metioned above,
the chronyd.service fails upon boot with status `266/NAMESPACE`
which seems to be due to the fact that the `/var/lib/chrony`
directory no longer exists.
This change sets the `/var/lib/chrony` directory for the `ReadWritePaths`
to be ignored when the directory does not exist, as per [2].
[1] https://github.com/kata-containers/runtime/issues/1279
[2] https://www.freedesktop.org/software/systemd
/man/systemd.exec.html#ReadWritePaths=
Fixes: #4167
Signed-off-by: Champ-Goblem <cameron_mcdermott@yahoo.co.uk>
Use `multistrap` for building Ubuntu rootfs. Adds support for building
for foreign architectures using the `ARCH` environment variable.
In the process, the Ubuntu rootfs workflow is vastly simplified.
Signed-off-by: Jakob Naucke <jakob.naucke@ibm.com>
- Add a doc comment
- Pass to build container, e.g. to build x86_64 with glibc (would
always use musl)
Signed-off-by: Jakob Naucke <jakob.naucke@ibm.com>
Remove a lot of cruft of musl installations -- we needed those for the
Go agent, but Rustup just takes care of everything. aarch64 on
Debian-based & Alpine is an exception -- create a symlink
`aarch64-linux-musl-gcc` to `musl-tools`'s `musl-gcc` or `gcc` on
Alpine. This is unified -- arch-specific Dockerfiles are removed.
Signed-off-by: Jakob Naucke <jakob.naucke@ibm.com>
Hadolint DL3019. If you're wondering why this is in this PR, that's
because I touch the file later, and we're only triggering the lints for
changed files.
Signed-off-by: Jakob Naucke <jakob.naucke@ibm.com>
to cover a Red Hat (adjacent) rootfs with great cross-platform compatibility
and a workable release cadence. The previous CentOS & Fedora workflows are
simplified.
Also remove unnecessary `/usr/share` files as on Ubuntu and mark Alpine
as unuspported on ppc64le (due to musl, for a while already).
Fixes: #3340
Signed-off-by: Jakob Naucke <jakob.naucke@ibm.com>
We install Rust in the build containers, but we also install Rust in
`rootfs.sh` if it is missing. It makes sense to install Rust in the build
containers so it does not have to be installed every time, but for that check
to work on non-login shells, we should source `.cargo/env` before running it.
Signed-off-by: Jakob Naucke <jakob.naucke@ibm.com>
Relative links within this repository allow for easier navigation to
the corresponding file / directory in the current commit / for the
selected version.
Link text was slightly changed / fixed in
- docs/Unit-Test-Advice.md
- docs/how-to/how-to-run-docker-with-kata.md
Fixes#3045
Signed-off-by: Daniel Höxtermann <daniel@hxtm.dev>
Remove the libseccomp package from Dockerfile of `alpine` and `clearlinux`
because the libseccomp library is installed by the `ci/install_libseccomp.sh`
script when building the kata-agent.
Fixes: #3508
Signed-off-by: Manabu Sugimoto <Manabu.Sugimoto@sony.com>
As the gentoo rootfs is not tested in our CI, we can't guarantee it
actually works as expected.
Whenever we have someone willing to maintain this rootfs we can have it
added back, and also add a CI job to test it altogether, avoiding then
any possible regression.
Fixes: #2144
Signed-off-by: Fabiano Fidêncio <fabiano.fidencio@intel.com>
As the suse rootfs is not tested in our CI, we can't guarantee it
actually works as expected.
Whenver we have someone willing to maintain this rootfs we can have it
added back, and also add a CI job to test it altogether, avoiding then
any possible regression.
Fixes: #2145
Signed-off-by: Fabiano Fidêncio <fabiano.fidencio@intel.com>
As the fedora rootfs is not tested in our CI, we can't guarantee it
actually works as expected.
Whenever we have someone willing to maintain the rootfs we can have it
added back, and also add a CI job to test it altogether, avoiding then
any possible regression.
Fixes: #2143
Signed-off-by: Fabiano Fidêncio <fabiano.fidencio@intel.com>
As the centos rootfs is not tested in our CI, we can't guarantee it
actually works as expected.
Whenever we have someone willing to maintain the rootfs we can have it
added back, and also add a CI job to test it altogether, avoiding then
any possible regression.
Fixes: #2140
Signed-off-by: Fabiano Fidêncio <fabiano.fidencio@intel.com>
Restore Debian as a rootfs.
1. revert of #3154, but some change
2. update debian version to 10.11
3. update `libstdc++-6-dev` to `libstdc++-8-dev`
4. changes discarded in QAT are not restored
Fixes: #3372
Signed-off-by: zhaojizhuang <571130360@qq.com>
Removed all errors/warnings pointed out by hadolint version 2.7.0, except for the following
ignored rules:
- "DL3008 warning: Pin versions in apt get install"
- "DL3041 warning: Specify version with `dnf install -y <package>-<version>`"
- "DL3033 warning: Specify version with `yum install -y <package>-<version>`"
- "DL3048 style: Invalid label key"
- "DL3003 warning: Use WORKDIR to switch to a directory"
- "DL3018 warning: Pin versions in apk add. Instead of apk add <package> use apk add <package>=<version>"
- "DL3037 warning: Specify version with zypper install -y <package>[=]<version>"
Fixes#3107
Signed-off-by: Wainer dos Santos Moschetta <wainersm@redhat.com>
It seems the lack of protoc in the alpine containers is causing issues
with some of our CIs, such as the VFIO one.
Fixes: #3323
Signed-off-by: Fabiano Fidêncio <fabiano.fidencio@intel.com>
Now if no options/arguments specified, the shell scripts will return an error:
ERROR: Invalid rootfs directory: ''
This commit will show usage if no options/arguments specified.
Fixes: #3256
Signed-off-by: bin <bin@hyper.sh>
Use the same runtime used for podman run also for the podman build cmd
Additionally remove "docker" from the docker_run_args variable
Fixes: #3239
Signed-off-by: Snir Sheriber <ssheribe@redhat.com>
- Upgrade Alpine guest rootfs to 3.15
- Specify a minor version rather than patch level as the Alpine
repositories use that.
Signed-off-by: Jakob Naucke <jakob.naucke@ibm.com>
#2399 partially reverted #418, missing on returning to bootstrapping a
rootfs with `apk.static` instead of copying the entire root, which can
result in drastically larger (more than 10x) images. Revert this as well
(requires some updates to URL building).
Fixes: #3216
Signed-off-by: Jakob Naucke <jakob.naucke@ibm.com>
Currently we do not have debian as part of the kata CI as we
do not have a mantainer, this PR removes debian as a supported
rootfs in order to have only the distros that we are supporting
and mantainining.
Fixes#3153
Signed-off-by: Gabriela Cervantes <gabriela.cervantes.tellez@intel.com>
So that the debug console is more useful. In the meantime, remove
iptables as it is not used by kata-agent any more.
Fixes: #3138
Signed-off-by: Peng Tao <bergwolf@hyper.sh>
The osbuilder needs to set up libseccomp library to build the kata-agent
because the kata-agent supports seccomp currently.
The library is built from the sources to create a static library for musl libc.
In addition, environment variables for the libseccomp crate are set to
link the library statically.
Fixes: #1476
Signed-off-by: Manabu Sugimoto <Manabu.Sugimoto@sony.com>
When building with dracut method the build_rootfs_distro() is not called, in turn
detect_rust_version() isn't either, so the install_rust.sh script is gave a null
rust version. This changed the script to call detect_rust_version() right before
install_rust.sh.
Related to commit: f34f67d610Fixes#2862
Signed-off-by: Wainer dos Santos Moschetta <wainersm@redhat.com>
and update the script in `ci/` accordingly.
When only parts of the Kata Containers repositories are checked out
(e.g. when building with Snap) and no Rust version is provided in
calling `install_rust.sh`, the scripts will attempt to clone the
appropriate repos to read the version, which will fail because the
directories already exist. Since we have read the version already, we
can just specify it.
Signed-off-by: Jakob Naucke <jakob.naucke@ibm.com>
The agent build inside a Docker or Podman container has been re-enabled,
but we have since introduced the `$CI` environment variable. Pass it to
avoid checking out the tests repo to main when there is a dependency.
Signed-off-by: Jakob Naucke <jakob.naucke@ibm.com>
or Podman. This is a partial revert of
76c18aa345. The rationale behind that
commit was the fact that the agent could not be built on Alpine, and
then this capability was removed altogether. The issue in Alpine has
since been resolved (see
https://github.com/kata-containers/osbuilder/issues/386). At the same
time, this ensures being able to run a glibc agent on hosts with distros
more recent than the osbuilder distro used (i.e. as of now, when you
build the agent on the host, and its glibc is newer than the one used in
the guest, the agent may encounter unresolved symbols).
Fixes#2398
Signed-off-by: Jakob Naucke <jakob.naucke@ibm.com>
The centos ppc64le gpg key at mirror.centos.org doesn't exist (link rot?).
Replacing it with url from CentOS/sig-core-AltArch on github.
Fixes: #2676
Signed-off-by: Aaron Simmons <paleozogt@gmail.com>
zsh doesn't support "==" as equal comparison operator, so
replace "==" with "=" to make the script more portable
Fixes: #2584
Signed-off-by: Binbin Zhang <binbin36520@gmail.com>
The script runs apt sync at some point which scans all possible fds
in order to close them. The operation is incredibly slow on VMs
and may lead to build timeouts.
Fix it by limiting the container runtime fds to a sane limit.
Fixes: #2510
Signed-off-by: Marcel Apfelbaum <marcel@redhat.com>
Once the ${ROOTFS_DIR} is created, the tool can't run the second
time since the directory is populated and the debootstrap tool
will fail.
Fix by deleting the contents of ${ROOTFS_DIR} if the directory exists.
Note that running make clean will also allow the re-run, it
is only an optimization for some cases the build fails in the middle.
Signed-off-by: Marcel Apfelbaum <marcel@redhat.com>
- no need to create `/usr/lib/systemd/systemd` link any more
- install `chrony` as extra package and install extra packages in chroot
rather than `debootstrap`, because `chrony` provides `time-daemon`,
which under 20.04 is provided by `systemd-timesyncd`, which is
required by `systemd`, and `debootstrap`'s conflict resolvement can't
handle this, but `apt`'s can.
Fixes: #2147
Depends-on: github.com/kata-containers/tests#3636
Signed-off-by: Jakob Naucke <jakob.naucke@ibm.com>
With lines like
0a2e2c6038/tools/osbuilder/rootfs-builder/fedora/config.sh (L8)
we imply that one can set another OS_VERSION and it will get picked up.
This is not the case when building inside Docker/Podman because the
variable is not passed to the container, which can lead to confusion.
Forward this env.
Fixes: #2378
Signed-off-by: Jakob Naucke <jakob.naucke@ibm.com>
Removed all TOCs now that GitHub auto-generates them.
Also updated the documentation requirements doc removing the requirement
to add a TOC.
Fixes: #2022.
Signed-off-by: James O. D. Hunt <james.o.hunt@intel.com>
If alpine image is created inside a container,
it does not get any golang version data. It will try
to get it by installing yq. To install yq curl is used.
Signed-off-by: Carlos Venegas <jos.c.venegas.munoz@intel.com>
Building rootfs does not depend on golang, delete intalling
golang may save build time.
And there is only rust agent now, the code for golang agent should
be deleted too.
Fixes: #2170
Signed-off-by: bin <bin@hyper.sh>
- Merge codepath in lib.sh with ppc64le -- do not install CMake
- Like ppc64le, do not install musl rather than just not using it
Fixes: #1975
Signed-off-by: Jakob Naucke <jakob.naucke@ibm.com>
While we introduced IMAGE_REGISTRY, we didn't actually update the
corresponding Dockerfiles to utilize it. Let's add
Fixes: #1622
Signed-off-by: Eric Ernst <eric.g.ernst@gmail.com>
For building rootfs with docker, glibc based rust target should be installed on ppc64le.
Additionally, protobuf-compiler would be required on ppc64le as it is not present by default.
Fixes: #1417
Signed-off-by: Amulya Meka <amulmek1@in.ibm.com>
Give the user chance to specify their own registry in event the default
provided are not accessible, desirable.
Fixes: #1393
Signed-off-by: Eric Ernst <eric.g.ernst@gmail.com>
- Install the required protobuf-compiler on Ubuntu
- Install correct libc Rust target (glibc on s390x)
- Do not skip Rust installation on s390x
Fixes: #1266
Signed-off-by: Jakob Naucke <jakob.naucke@ibm.com>
Since there is no Rust target for musl on s390x, builds on s390x should use
glibc. This commit removes the "Cannot build Rust agent on s390x" restriction
in rootfs.sh and only installs musl in the build environment when it is
required.
Fixes: #1262
Signed-off-by: Jakob Naucke <jakob.naucke@ibm.com>
