After previous commit, found that kata-proxy is not quit
when vmcache server is stopped by ctrl-c.
The cause is current kata-proxy is setsid when it exec. It will
not get the signal ctrl-c.
Call vm.Disconnect() when close vm in cache factory to handle
this issue.
Fixes: #1726
Signed-off-by: Hui Zhu <teawater@hyper.sh>
Now, template feature can't work on arm64. When using bypass-shared-memory
cap to saving sandbox as a template, new sandbox cloning from the template
will fail. From initial investigation, it maybe a qemu issue. Further
research should be done.
To enable template on arm64, this patch adds a switch to offer the capability
to open or close the bypass-shared-memory cap.
While closing bypass-shared-memory cap, saving vm will occupy more space
and consume more time. So increase 300M for mount size and bump the time
waiting for migration to 10 seconds.
Fixes: #1336
Signed-off-by: Jianyong Wu <jianyong.wu@arm.com>
This commit adds a new gRPC function Status to CacheService. VMCache
server will reply the status of VMCache server.
Factory destroy will call gRPC Status to get the status of VMCache
server and output it when VMCache is enabled.
Fixes: #1395
Signed-off-by: Hui Zhu <teawater@hyper.sh>
Now, function NewFactory will return nil even create template
does't complete. As for this, it will tell user that factory
has been initialized no matter whether the template is created
or not. This patch correct it by adding another return value
of error in NewFactory.
Testing initFactoryCommand when enable template will need root
privilege to mount tmpfs. So skip it for no-root user.
Testing initFactoryCommand func will create template, but no
proxy type assigned to VMconfig which will using katabuiltinProxy
instead. this will lead to failure for this type of proxy will
check proxyparams which contains many null value. This commit
fix it by substitute katabuiltinProxy as noopProxy when for test
purpose.
Fixes: #1333
Signed-off-by: Jianyong Wu <jianyong.wu@arm.com>
VMCache is a new function that creates VMs as caches before using it.
It helps speed up new container creation.
The function consists of a server and some clients communicating
through Unix socket. The protocol is gRPC in protocols/cache/cache.proto.
The VMCache server will create some VMs and cache them by factory cache.
It will convert the VM to gRPC format and transport it when gets
requestion from clients.
Factory grpccache is the VMCache client. It will request gRPC format
VM and convert it back to a VM. If VMCache function is enabled,
kata-runtime will request VM from factory grpccache when it creates
a new sandbox.
VMCache has two options.
vm_cache_number specifies the number of caches of VMCache:
unspecified or == 0 --> VMCache is disabled
> 0 --> will be set to the specified number
vm_cache_endpoint specifies the address of the Unix socket.
This commit just includes the core and the client of VMCache.
Currently, VM cache still cannot work with VM templating and vsock.
And just support qemu.
Fixes: #52
Signed-off-by: Hui Zhu <teawater@hyper.sh>
We are creating Store directories but never removing them.
Calling into a VM factory created vm Stop() will now clean the VM Store
artifacts up.
Signed-off-by: Samuel Ortiz <sameo@linux.intel.com>
We convert the whole virtcontainers code to use the store package
instead of the resource_storage one. The resource_storage removal will
happen in a separate change for a more logical split.
This change is fairly big but mostly does not change the code logic.
What really changes is when we create a store for a container or a
sandbox. We now need to explictly do so instead of just assigning a
filesystem{} instance. Other than that, the logic is kept intact.
Fixes: #1099
Signed-off-by: Samuel Ortiz <sameo@linux.intel.com>
reflect.DeepEqual would return false when comparing nil map/slice with
empty map/slice. We would want to return success in such case, since it
is possible for upper layers to send these kind of configs.
Fixes: #844
Signed-off-by: Peng Tao <bergwolf@gmail.com>
Then we can remove the arbitrary sleep waiting for migration
completion when creating a tempalte vm.
Fixes: #728
Signed-off-by: Peng Tao <bergwolf@gmail.com>
The PR moves ahead the start of proxy process for vm factory so that
it waits for both vm and proxy to be up at the same time. This saves
about 300ms for new container creation in my local test machine.
Fixes: #683
Signed-off-by: Peng Tao <bergwolf@gmail.com>
Now that we only use hypervisor config to set them, they
are not overridden by other configs. So drop the default prefix.
Signed-off-by: Peng Tao <bergwolf@gmail.com>
Add additional `context.Context` parameters and `struct` fields to allow
trace spans to be created by the `virtcontainers` internal functions,
objects and sub-packages.
Note that not every function is traced; we can add more traces as
desired.
Fixes#566.
Signed-off-by: James O. D. Hunt <james.o.hunt@intel.com>
Add vm factory support per design in the VM Factory plugin section.
The vm factory controls how a new vm is created:
1. direct: vm is created directly
2. template: vm is created via vm template. A template vm is pre-created
and saved. Later vm is just a clone of the template vm so that they
readonly share a portion of initial memory (including kernel, initramfs
and the kata agent). CPU and memory are hot plugged when necessary.
3. cache: vm is created via vm caches. A set of cached vm are pre-created
and maintained alive. New vms are created by just picking a cached vm.
CPU and memory are hot plugged when necessary.
Fixes: #303
Signed-off-by: Peng Tao <bergwolf@gmail.com>
