inotify/watchable-mount changes...
- Allow up to 16 files. It isn't that uncommon to have 3 files in a secret.
In Kubernetes, this results in 9 files in the mount (the presented files,
which are symlinks to the latest files, which are symlinks to actual files
which are in a seperate hidden directoy on the mount). Bumping from eight to 16 will
help ensure we can support "most" secret/tokens, and is still a pretty
small number to scan...
- Now we will only replace the watched storage with a bindmount if we observe
that there are too many files or if its too large. Since the scanning/updating is racy,
we should expect that we'll occassionally run into errors (ie, a file
deleted between scan / update). Rather than stopping and making a bind
mount, continue updating, as the changes will be updated the next time
check is called for that entry (every 2 seconds today).
To facilitate the 'oversized' handling, we create specific errors for too large
or too many files, and handle these specific errors when scanning the storage entry.
- When handling an oversided mount, do not remove the prior files -- we'll just
overwrite them with the bindmount. This'll help avoid the files
disappearing from the user, avoid racy cleanup and simplifies the flow.
Similarly, only mark it as a non-watched storage device after the
bindmount is created successfully.
- When creating bind mount, make sure destination exists. If we hadn't
had a successful scan before, this wouldn't exist and the mount would
fail. Update logic and unit test to cover this.
- In several spots, we were returning when there was an error (both in
scan and update). For update case, let's just log an warning and continue;
since the scan/update is racy, we should expect that we'll have
transient errors which should resolve the next time the watcher runs.
Fixes: #2402
Signed-off-by: Eric Ernst <eric_ernst@apple.com>
changed the document project url in the using-vpp-and-kata.md and
runtime experimental README.md files.
Fixes: #2418
Signed-off-by: wangyongchao.bj <wangyongchao.bj@inspur.com>
When running a firecracker instance jailed, block devices
are not removed correctly, as the jailerRoot path is not
stripped from the PATCH command sent to the FC API.
This patch differentiates the jailed case from the non-jailed
one and allows the firecracker instance to be properly
terminated.
Fixes#2387
Signed-off-by: Anastassios Nanos <ananos@nubificus.co.uk>
Update to commit 3c64244cbb, in particular to get these fixes which
are needed to work with qemu-6.0 and later:
https://github.com/kata-containers/govmm/pull/192https://github.com/kata-containers/govmm/pull/194
Git log
d27256f (qmp: Don't use deprecated 'props' field for object-add, 2021-08-03)
d8cdf9a (qemu: Drop support for versions older than 5.0, 2021-08-03)
1b02192 (Use 'host_device' driver for blockdev backends, 2021-07-29)
9518675 (add support for "sandbox" feature to qemu, 2021-07-20)
335fa81 (qemu: fix golangci-lint errors, 2021-07-21)
61b6378 (.github/workflows: reimplement github actions CI, 2021-07-21)
9d6e797 (go: support go modules, 2021-07-21)
0d21263 (qemu: support read-only nvdimm, 2021-07-21)
Signed-off-by: David Gibson <david@gibson.dropbear.id.au>
Kata uses the 'kernel_irqchip' machine option to qemu. By default it
uses it in what qemu calls the "short-form boolean" with no parameter.
That style was deprecated by qemu between 5.2 and 6.0 (commit
ccd3b3b8112b) and effectively removed entirely between 6.0 and 6.1
(commit d8fb7d0969d5).
Update ourselves for newer qemus by using an explicit
"kernel_irqchip=on".
Signed-off-by: David Gibson <david@gibson.dropbear.id.au>
The swappiness is not right if just set
io.katacontainers.container.resource.swappiness:
$ pod_yaml=pod.yaml
$ container_yaml=container.yaml
$ image="quay.io/prometheus/busybox:latest"
$ cat << EOF > "${pod_yaml}"
metadata:
name: busybox-sandbox1
EOF
$ cat << EOF > "${container_yaml}"
metadata:
name: busybox-killed-vmm
annotations:
io.katacontainers.container.resource.swappiness: "100"
image:
image: "$image"
command:
- top
EOF
$ sudo crictl pull $image
$ podid=$(sudo crictl runp $pod_yaml)
$ cid=$(sudo crictl create $podid $container_yaml $pod_yaml)
$ sudo crictl start $cid
crictl exec $cid cat /sys/fs/cgroup/memory/memory.swappiness
60
The cause of this issue is there are two elements store the resources
infomation. They are c.config.Resources for calculateSandboxMemory and
c.GetPatchedOCISpec() for agent.
This add initConfigResourcesMemory to Container and call it in
newContainer to handle the issue.
Fixes: #2372
Signed-off-by: Hui Zhu <teawater@antfin.com>
Although the OCI specification does not explictly requires that, we
should create the process CWD if it does not exist, before chdir'ing
to it. Without that fizx, the kata-agent fails to create a container
and returns a grpc error when it's trying to change the containerd
working directory to an non existing folder.
runc, the OCI runtime reference implementation, also creates the process
CWD when it's not part of the container rootfs.
Fixes#2374
Signed-off-by: Samuel Ortiz <samuel.e.ortiz@protonmail.com>
When enabling tracing with Cloud Hypervisor, we end up establishing 2
connections to 2 different HTTP servers: The Cloud Hypervisor API one
that runs over a UNIX socket and the Jaeger endpoint running over UDP.
Both connections use the default HTTP golang client instance, and thus
share the same transport layer. As the Cloud Hypervisor implementation
sets it up to be over a Unix socket, the jaeger uploader ends up going
through that transport as well, and sending its spans to the Cloud
Hypervisor API server.
We fix that by giving the Cloud Hypervisor implementation its own HTTP
client instance and we avoid sharing it with anything else in the shim.
Fixes#2364
Signed-off-by: Samuel Ortiz <samuel.e.ortiz@protonmail.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>
The old ones are carrying CVEs, do not use them.
PS: In order to update the modules, we're running `make handle_vendor`
target from the runtime's Makefile. This is now part of the CI and
ensures that the vendored code is up-to-date. It's important to note
that older versions of golang may generate different results for those,
but those versions are not supported anymore, so we're good to go with
what we have in the CI (1.15 and 1.16).
Signed-off-by: Peng Tao <bergwolf@hyper.sh>
Signed-off-by: Fabiano Fidêncio <fidencio@redhat.com>
Since the old ones are carrying CVEs. Do not use them.
PS: In order to update the modules, we're running `make handle_vendor`
target from the runtime's Makefile. This is now part of the CI and
ensures that the vendored code is up-to-date. It's important to note
that older versions of golang may generate different results for those,
but those versions are not supported anymore, so we're good to go with
what we have in the CI (1.15 and 1.16).
Fixes: #2338
Signed-off-by: Peng Tao <bergwolf@hyper.sh>
Signed-off-by: Fabiano Fidêncio <fidencio@redhat.com>
There is a new "ReadOnly" option added to nvdimm device in qemu
and now added to kata. However, qemu used for arm64 is a little
old and has no this feature. Here we remove this feature for arm.
Fixes: #2320
Signed-off-by: Jianyong Wu <jianyong.wu@arm.com>
This PR updates the experimental documentation with the proper reference
to kata 2.x
Fixes#2317
Signed-off-by: Gabriela Cervantes <gabriela.cervantes.tellez@intel.com>
Pior our bump to runc 1.0.1 the manager's Set() would take a Config as
its parameter. Now it takes the Resources directly.
Signed-off-by: Fabiano Fidêncio <fidencio@redhat.com>
Previously part of the "system" namespace, the RunningInUserNS() has
been moved to the "userns" namespace.
Signed-off-by: Fabiano Fidêncio <fidencio@redhat.com>
Dependabot brought to us attention that we were still vendoring the runc
code which was affected by CVE-2021-30465.
Although the vulnerability doesn't seem to affect kata-containers, we
better keep our dependencies up-to-date anyways. With this in mind,
let's bump our runc dependency to the latest release.
Fixes: #2309
Signed-off-by: Fabiano Fidêncio <fidencio@redhat.com>
Improve security by making rootfs image read-only, nobody
will be able to modify it from the guest.
fixes#1916
Signed-off-by: Julio Montes <julio.montes@intel.com>
Bring read-only nvdimm support
Shortlog:
335fa81 qemu: fix golangci-lint errors
61b6378 .github/workflows: reimplement github actions CI
9d6e797 go: support go modules
0d21263 qemu: support read-only nvdimm
ff34d28 qemu: Consistent parameter building
Signed-off-by: Julio Montes <julio.montes@intel.com>
Since the monitor socket used the unix socket path file,
which needed to be cleaned after the pod terminated,
thus put it into the sandbox data directory, and it
would be cleaned up once the sandbox termianted.
Fixes: #2269
Signed-off-by: fupan.lfp <fupan.lfp@antgroup.com>
This commit add option "enable_guest_swap" to config hypervisor.qemu.
It will enable swap in the guest. Default false.
When enable_guest_swap is enabled, insert a raw file to the guest as the
swap device if the swappiness of a container (set by annotation
"io.katacontainers.container.resource.swappiness") is bigger than 0.
The size of the swap device should be
swap_in_bytes (set by annotation
"io.katacontainers.container.resource.swap_in_bytes") - memory_limit_in_bytes.
If swap_in_bytes is not set, the size should be memory_limit_in_bytes.
If swap_in_bytes and memory_limit_in_bytes is not set, the size should be
default_memory.
Fixes: #2201
Signed-off-by: Hui Zhu <teawater@antfin.com>
