This is similar to ae64ab6b82 from #2849 which
did the same for runtime.mkdir.
This makes it possible to specify both host (absolute) or container (relative)
paths.
Signed-off-by: Ian Campbell <ijc@docker.com>
Pulls in:
f816553 Merge pull request #195 from ijc/prepend-rootfs-mounts
0a6d69d Prepend the rootfs mounts to the user specified runtime.mounts.
Signed-off-by: Ian Campbell <ijc@docker.com>
Will return the list of all hosts that are part of a vCenter DataCenter
and choose one at random to use.
Signed-off-by: Dan Finneran <daniel.finneran@gmail.com>
This simply starts a web server serving the specified directory.
It's useful for PXE booting.
Signed-off-by: Rolf Neugebauer <rolf.neugebauer@docker.com>
This currently just copies the kernel/initrd to a destination
and writes the iPXE script there as well. The scheme is flexible
enough to support scp or other means for pushing in the future.
The kernel/initrd are conditionally decompressed (default for arm64).
Signed-off-by: Rolf Neugebauer <rolf.neugebauer@docker.com>
Using the userdat to pass the iPXE script precludes us from passing
actual userdata to the instance. Instead pass a URL to the iPXE script.
The script is expected to be a <base-url>/<name>-packet.ipxe.
When starting a server also serve the iPXE script from memory (not
the file system). The http server needed extending to differentiate
between the serving files and the iPXE script and ServeMux is used
to differentiate based on path.
A subsequent commit will provide an option to write the iPXE script
to disk.
Signed-off-by: Rolf Neugebauer <rolf.neugebauer@docker.com>
prefix packet.net specific functions with 'packet' and make
non-packet specific functions, such as validateHTTPURL()
more generic so that they can easily be moved to 'utils.go'
should they be needed elsewhere.
Signed-off-by: Rolf Neugebauer <rolf.neugebauer@docker.com>
This looks like there are a couple of minor fixes to the
recent KPTI changes but nothing major...
Signed-off-by: Rolf Neugebauer <rolf.neugebauer@docker.com>
This PR correctly plumbs a single context to propagate the containerd
namespace to the necessary commands. Services launched with containerd
after this change will now be in a default namespace of
`services.linuxkit`.
A top-level flag is added to the service command,
`--containerd-namespace` which can be used to change, if needed.
Signed-off-by: Stephen J Day <stephen.day@docker.com>
The scheme we currently have seems relatively usable and
this project has not been maintained for a while.
Signed-off-by: Rolf Neugebauer <rolf.neugebauer@docker.com>
The scripts used to scrape the various repositories to build
and push package for all foreign kernels. They were designed
to be run periodically and provide "official" foreign kernel
packages. Needless to say we did not run them periodically
and the linuxkit packages became out-dated quickly.
Now, we just provide users who are interested in using foreign
kernels the means to build their own package from specific
vendor kernels.
Each script uses slightly different command line arguments
as the location and naming of the kernel packages differ
wildly. The help message provide a working example which
has been tested with a minimal LinuxKit YAML file.
Signed-off-by: Rolf Neugebauer <rolf.neugebauer@docker.com>
Support for this has stalled in the swarmkit project due to lack of maintainer
time to review and support and the existing code no longer works with the
version of containerd used in linuxkit.
Signed-off-by: Ian Campbell <ijc@docker.com>
