Make getStore() take a config struct from which it pulls the store
options, then update the kpod commands so that they call getConfig()
and pass the config into getStore()
Signed-off-by: Ryan Cole <rcyoalne@gmail.com>
Avoid parsing metadata that the image library keeps in order to find an
image's top layer and creation date; instead, use the values which the
storage library now makes available, which will be correct once we merge
PR #654 or something like it.
Instead of assuming the last blob which was added for the image was the
manifest, read it directly and compute its digest ourselves.
Signed-off-by: Nalin Dahyabhai <nalin@redhat.com>
Switch from using the lower-level storage APIs (accessing LayerStore,
ImageStore, and ContainerStore types directly) in favor of the
higher-level ones that take care of synchronization and locking for us.
Signed-off-by: Nalin Dahyabhai <nalin@redhat.com>
Bump containers/image (pulling in its new dependency on ostree-go),
containers/storage, and updated image-spec.
This pulls in the OCI v1.0 specifications and code that allows us to
support 1.0 images.
Signed-off-by: Dan Walsh <dwalsh@redhat.com>
Signed-off-by: Nalin Dahyabhai <nalin@redhat.com>
Design: The output of the `info` subcommand ought to be directly
consumable in a format like JSON or yaml.
The structure being a map of sorts.
Each subsection of information being an individual cluster under the
top-level, like platform info, debug, storage, etc.
Even if there are errors under the top level key, the value will be a
map with the key of "error" and the value as the message of the
`err.Error()`. In this way, the command always returns usable output.
Ideally there will be a means for anything that can register info to do
so independently from it being in the single info.go, so this approach
is having a typed signature for the function that gives info, but i'm
sure it could be better.
Current iteration of this outputs the following as a limited user:
```yaml
host:
MemFree: 711307264
MemTotal: 2096222208
SwapFree: 2147479552
SwapTotal: 2147479552
arch: amd64
cpus: 1
os: linux
store:
error: 'mkdir /var/run/containers/storage: permission denied'
```
and as root (`sudo kpod info -D`):
```yaml
debug:
compiler: gc
go version: go1.7.6
goroutines: 3
host:
MemFree: 717795328
MemTotal: 2096222208
SwapFree: 2147479552
SwapTotal: 2147479552
arch: amd64
cpus: 1
os: linux
store:
ContainerStore:
number: 1
GraphDriverName: overlay2
GraphRoot: /var/lib/containers/storage
ImageStore:
number: 1
```
And with the `--json --debug` flag:
```json
{
"debug": {
"compiler": "gc",
"go version": "go1.7.6",
"goroutines": 3
},
"host": {
"MemFree": 709402624,
"MemTotal": 2096222208,
"SwapFree": 2147479552,
"SwapTotal": 2147479552,
"arch": "amd64",
"cpus": 1,
"os": "linux"
},
"store": {
"ContainerStore": {
"number": 1
},
"GraphDriverName": "overlay2",
"GraphRoot": "/var/lib/containers/storage",
"ImageStore": {
"number": 1
}
}
}
```
Signed-off-by: Vincent Batts <vbatts@hashbangbash.com>
We now have actual kpod code, so no reason to have a not implemented feature.
Especially when we don't intend to create kpod launch.
Signed-off-by: Daniel J Walsh <dwalsh@redhat.com>
The syscall package is locked down and the comment in [1] advises to
switch code to use the corresponding package from golang.org/x/sys. Do
so and replace usage of package syscall where possible (leave
syscall.SysProcAttr and syscall.Stat_t).
[1] https://github.com/golang/go/blob/master/src/syscall/syscall.go#L21-L24
This will also allow to get updates and fixes just by re-vendoring
golang.org/x/sys/unix instead of having to update to a new go version.
Signed-off-by: Tobias Klauser <tklauser@distanz.ch>
'kpod images' lists all images on a system. 'kpod rmi' removes
one or more images from a system. The images will not be removed
if they are associated with a running container, unless the -f
option is used
Signed-off-by: Ryan Cole <rcyoalne@gmail.com>
This patch also hides the profile under the debug flag as there's
runtime cost to enable the profiler.
This removes the old way of profiling (CPU) as that's not really
needed.
Signed-off-by: Antonio Murdaca <runcom@redhat.com>
Container runtimes provide different levels of isolation, from kernel
namespaces to hardware virtualization. When starting a specific
container, one may want to decide which level of isolation to use
depending on how much we trust the container workload. Fully verified
and signed containers may not need the hardware isolation layer but e.g.
CI jobs pulling packages from many untrusted sources should probably not
run only on a kernel namespace isolation layer.
Here we allow CRI-O users to define a container runtime for trusted
containers and another one for untrusted containers, and also to define
a general, default trust level. This anticipates future kubelet
implementations that would be able to tag containers as trusted or
untrusted. When missing a kubelet hint, containers are trusted by
default.
A container becomes untrusted if we get a hint in that direction from
kubelet or if the default trust level is set to "untrusted" and the
container is not privileged. In both cases CRI-O will try to use the
untrusted container runtime. For any other cases, it will switch to the
trusted one.
Signed-off-by: Samuel Ortiz <sameo@linux.intel.com>
If we create a container using the image ID like
771cd5947d5ea4bf8e8f4900dd357dbb67e7b16486c270f8274087d182d457c6, then
a call to container_status will return that same ID for the "Image"
field in ContainerStatusResponse.
This patch matches dockershim behavior and return the first tagged name
if available from the image store.
This is also needed to fix a failure in k8s e2d tests.
Reference:
https://github.com/kubernetes/kubernetes/pull/39298/files#diff-c7dd39479fd733354254e70845075db5R369
Reference:
67a5bf8454/test/e2e/framework/util.go (L1941)
Signed-off-by: Antonio Murdaca <runcom@redhat.com>
The ocid project was renamed to CRI-O, months ago, it is time that we moved
all of the code to the new name. We want to elminate the name ocid from use.
Move fully to crio.
Also cric is being renamed to crioctl for the time being.
Signed-off-by: Dan Walsh <dwalsh@redhat.com>
To collect CPU profile information added a flag `--cpu-profile`
which is a path to file where this collected information will be
dumped.
Fixes#464
Signed-off-by: Suraj Deshmukh <surajssd009005@gmail.com>
We use the k8s remotecommand client API to create a
streaming executor, and then stream the executed process
into stdout/stderr.
Signed-off-by: Samuel Ortiz <sameo@linux.intel.com>
We have moved selinux support out of opencontainers/runc into its
own package. This patch moves to using the new selinux go bindings.
Signed-off-by: Daniel J Walsh <dwalsh@redhat.com>
Not all runtimes are able to handle some of the kubelet
security context options, in particular the ones granting
host privileges to containers.
By adding a host privileged runtime path configuration, we
allow ocid to use a different runtime for host privileged
operations like e.g. host namespaces access.
Signed-off-by: Samuel Ortiz <sameo@linux.intel.com>
container-storage-setup (Formerly docker-storage-setup) is being converted to
run with container runtimes outside of docker. Specifically we want to use it
with CRI-O/ocid. It does not know anything about the container runtimes it
is generating options for, so it generates them based on the storage CLI of
docker. I see no reason to have the storage option for ocid to be different
and we can just depracate the option for now.
Signed-off-by: Daniel J Walsh <dwalsh@redhat.com>
Use containers/storage to store images, pod sandboxes, and containers.
A pod sandbox's infrastructure container has the same ID as the pod to
which it belongs, and all containers also keep track of their pod's ID.
The container configuration that we build using the data in a
CreateContainerRequest is stored in the container's ContainerDirectory
and ContainerRunDirectory.
We catch SIGTERM and SIGINT, and when we receive either, we gracefully
exit the grpc loop. If we also think that there aren't any container
filesystems in use, we attempt to do a clean shutdown of the storage
driver.
The test harness now waits for ocid to exit before attempting to delete
the storage root directory.
Signed-off-by: Nalin Dahyabhai <nalin@redhat.com>
Vendor updated containers/image and containers/storage, along
with any new dependencies they drag in, and updated versions of other
dependencies that happen to get pulled in.
github.com/coreos/go-systemd/daemon/SdNotify() now takes a boolean to
control whether or not it unsets the NOTIFY_SOCKET variable from the
calling process's environment. Adapt.
github.com/opencontainers/runtime-tools/generate/Generator.AddProcessEnv()
now takes the environment variable name and value as two arguments, not
one. Adapt.
Signed-off-by: Nalin Dahyabhai <nalin@redhat.com>
* Rename 'vendor/src' -> 'vendor'
* Ignore vendor/ instead of vendor/src/ for lint
* Rename 'cmd/client' -> 'cmd/ocic' to make it 'go install'able
* Rename 'cmd/server' -> 'cmd/ocid' to make it 'go install'able
* Update Makefile to build and install from GOPATH
* Update tests to locate ocid/ocic in GOPATH/bin
* Search for binaries in GOPATH/bin instead of PATH
* Install tools using `go get -u`, so they are updated on each run
Signed-off-by: Jonathan Yu <jawnsy@redhat.com>
The default configuration can only be accessed from the cmd/server
package, which cannot be imported (since it's a "package main").
This change promotes DefaultConfig() to the "server" package.
Closes: #315
Signed-off-by: Jonathan Yu <jawnsy@redhat.com>
The client size field that we get back when we inspect an image is a
pointer to a number, not just a number, so we need to dereference it for
display.
Signed-off-by: Nalin Dahyabhai <nalin@redhat.com>
We add 2 ocid options for choosing the CNI configuration and plugin
binaries directories: --cni-config-dir and --cni-plugin-dir.
Signed-off-by: Samuel Ortiz <sameo@linux.intel.com>
Any binary that will be managing storage needs to initialize the reexec
package in order to be able to apply or read image layers.
Signed-off-by: Nalin Dahyabhai <nalin@redhat.com>
While logDir isn't currently used (until the conmon implementation
lands) it's probably not a great idea to hardcode our defaults. The main
issue with this setting is that the kubelet can override it at will.
Signed-off-by: Aleksa Sarai <asarai@suse.de>
It's a bit odd to have ImageStore be part of the config and yet we don't
allow people to modify it. However, leave it out of the commented
version because it's currently unused.
Signed-off-by: Aleksa Sarai <asarai@suse.de>
This subcommand is so that users can get a nice commented version of the
ocid configuration file. This comes from the "current" version of the
configuration (allowing somone to get their custom configuration as a
file). It also has a --default option.
In addition, update the tests to use `ocid config` so that we test this
setup (the loading and saving of the options).
Signed-off-by: Aleksa Sarai <asarai@suse.de>
The --config is merged with the default, and then is overridden by any
command-line options. Everything is organised to be in sub-tables so
that the sections are more clear.
Signed-off-by: Aleksa Sarai <asarai@suse.de>
This paves the way for having a configuration file that is loaded rather
than everything being set via the command-line.
Signed-off-by: Aleksa Sarai <asarai@suse.de>
This fixes a bug where --conmon wouldn't actually set the conmon binary
path, and also where we weren't setting CFLAGS while compiling conmon.
Signed-off-by: Aleksa Sarai <asarai@suse.de>
This lessens the Docker requirement for creating sandboxes (with the
requirement only existing for the actual image pulling that is done when
adding a container to a pod). The interface was chosen to match the
--conmon interface, so that the location of the pause binary can be
chosen by a user.
Signed-off-by: Aleksa Sarai <asarai@suse.de>
