NEMU deprecation was announced in 1.8 of Kata. Removing from tree.
Thanks for all the fish!
Fixes: #2195
Signed-off-by: Eric Ernst <eric.ernst@intel.com>
For shimv2 case, when hypervisor's debug option set, log out
the firecracker's console output which contains the kernel boot
logs; thus it would be easy for system panic debugging.
When agent debug was enabled by passing "agent.log=debug" to
kernel parameter, it will also log out the agent logs from
the console output.
Fixes: #2201
Signed-off-by: lifupan <lifupan@gmail.com>
Updated to better clarify this configuration option. Currently in QEMU
pre 1.9 release of Kata, vhost-net is used for the virtio-net backend.
This results in efficient network I/O performance, but does rely on the
backend running in ring0 (host kernel). Update comment to clarify this
trade-off for end-users.
Fixes: #2198
Signed-off-by: Eric Ernst <eric.ernst@intel.com>
vhost-net backend is not supported by the Firecracker VMM. It doesn't
make sense to have this in the configuration, and we should explicitly
disable it.
Fixes: #2192
Signed-off-by: Eric Ernst <eric.ernst@intel.com>
Firecracker have its own logging scheme, providing two fifo files with log
and metrics info.
We should extract error info for better debugging.
Fixes: #2072
Signed-off-by: Penny Zheng <penny.zheng@arm.com>
Signed-off-by: Haibo Xu <haibo.xu@arm.com>
firecracker 0.19.0 API is not backward compatible, hence we need
to bump the firecracker minimum supported version to 0.19.0
Signed-off-by: Julio Montes <julio.montes@intel.com>
Prior to the addition of tcMirroring support kata-runtime had
compatibility issues with some CNI plugins some of which were addressed
by the bridged model. With the addition of tc mode there are no gaps in
networking that can be filled by the bridged mode or enlightened mode
(which was never implemented).
Eliminate both of these options to simplify the setup.
Fixes: #1213
Signed-off-by: Manohar Castelino <manohar.r.castelino@intel.com>
We do not want to create cgroups in case of rootless.
Fix the logic to implement this.
Fixes#2177
Signed-off-by: Archana Shinde <archana.m.shinde@intel.com>
FIPS are a set of security standards for encryption algorithms
in user and kernel space among others.
Have Kata support this by starting the VM for a container
in FIPS mode on detecting that the host is running in FIPS mode.
Depends-on: github.com/kata-containers/packaging#788
Fixes#2170
Signed-off-by: Archana Shinde <archana.m.shinde@intel.com>
cri-tools version was managed in the tests repository, but as
we define here cri-o, containerd and kubernetes versions, it
make sense to have the cri-tools version defined in this repo.
conmon has now to be installed/built separately. So add it
to the list.
Depends-on: github.com/kata-containers/tests#2057
Signed-off-by: Salvador Fuentes <salvador.fuentes@intel.com>
Update k8s supported version from 1.15.3 to 1.16.2
and cri-o from 1.15.0 to 1.16.0
Fixes: #2166.
Signed-off-by: Salvador Fuentes <salvador.fuentes@intel.com>
- sandbox/cgroups: don't constrain if using SandboxCgroupsOnly
- cli: add kata-overhead subcommand
- versions: support specify version in versions.yaml for rust agent.
- virtcontainers/sandbox: calculate container's CPU from sandbox.contai…
- ci: Fix versions_checker.sh
- ci: Fix versions_checker.sh
- virtcontainers: unmount host mounts if container can't be created
- virtcontainers/store: make VCStoreUUIDPath rootless
- virtcontainers/annotations: use right domain name for kata annotations
- v2: Change the event and error behavior of pause/resume
- Update cni plugin version
- Load state early so that hypervisor can store the correct state
e4c816b versions: support specify version in versions.yaml for rust agent.
691a6a7 sandbox/cgroups: don't constrain if using SandboxCgroupsOnly
7fe0100 cli: add kata-overhead command
1bbc1d5 virtcontainers: add StatsSandbox to vc API
569bd78 virtcontainers: change pass by value to pass by reference
5b226d0 ci: Fix versions_checker.sh
f8b84d7 ci: Fix versions_checker.sh
24d7aff virtcontainers: change pass by value to pass by reference
abec17f virtcontainers/store: make VCStoreUUIDPath rootless
eca7bd2 virtcontainers: unmount host mounts if container can't be created
91bd095 virtcontainers/annotations: use right domain name for kata annotations
c1060a3 v2: Change the event and error behavior of pause/resume
f6a10bc state: Refactor code to move all the state load code
fa4acad state: Load the state from storage early on
929c4e7 network: Change NewNS() call
c0995c6 vendor: Vendor the latest CNI plugins
86d8346 version: Update the version for cni plugins
Signed-off-by: Eric Ernst <eric.ernst@intel.com>
When SandboxCgroupsOnly is set, we are expected to just inherit our parent's
cgroup settings and to move all Kata threads within that sandbox cgroup. The
initial implementation still adjusted the size of this cgroup. This commit
fixes this.
This commit makes a couple of functional changes, small refactors, and
adds clarifying comments for some functions.
Fixes: #2090
Signed-off-by: Eric Ernst <eric.ernst@intel.com>
Introduce kata-overhead command to kata-runtime CLI, to help
with calculating sandbox overhead.
Fixes: #2096
Signed-off-by: Jose Carlos Venegas Munoz <jose.carlos.venegas.munoz@intel.com>
Signed-off-by: Eric Ernst <eric.ernst@intel.com>
StatsSandbox is used to gather metrics for the sandbox (host cgroup) as
well as from the individual containers (from the guest cgroups). This is
intended to be used for easily calculating Kata sandbox overheads.
Fixes: #2096
Signed-off-by: Jose Carlos Venegas Munoz <jose.carlos.venegas.munoz@intel.com>
Signed-off-by: Eric Ernst <eric.ernst@intel.com>
container.config does not point to sandbox.config.Containers.ContainerConfig
which caused the ContainerConfig not sync.
Fixes: #2129
Signed-off-by: Wang Liang <wangliangzz@inspur.com>
These include features like privileged containers without host devices
and support for per runtime annotations.
Depends-on: github.com/kata-containers/tests#2029
Fixes#2099
Signed-off-by: Archana Shinde <archana.m.shinde@intel.com>
ACRN doesn't support configuring number of guest vcpu option ('-c') anymore.
Number of guest vcpus will be defined in the hypervisor scenario
configuration file instead.
Removed the -c option from the acrn-dm parameters when launching VMs and
also trimmed configuration.toml file accordingly.
fixes#2136
Signed-off-by: Vijay Dhanraj <vijay.dhanraj@intel.com>
