Current versions of Docker complain:
Sending build context to Docker daemon 5.12kB
Step 1/4 : FROM linuxkit/guestfs@sha256:62e810869e83dabe04af7ac8fcd81d9f177d986a3424dcb9c4eb5439993c962f
...
Successfully tagged linuxkit/mkimage-gcp:736880776838c8f41c081276d9c017ae6adbf1bb-amd64
Tagging linuxkit/guestfs@sha256:62e810869e83dabe04af7ac8fcd81d9f177d986a3424dcb9c4eb5439993c962f as linuxkit/guestfs:aecc40bf84ce4f4238e06934c46d7cc570a39eed@sha256:62e810869e83dabe04af7ac8fcd81d9f177d986a3424dcb9c4eb5439993c962f
refusing to create a tag with a digest reference
../../pkg/package.mk:82: recipe for target 'tag-y' failed
Just use the digest for now. Perhaps someone will have time to fixup this build
properly at some point.
Make the same change to test/pkg/ltp.
Signed-off-by: Ian Campbell <ijc@docker.com>
Using explicit targets is cleaner and easier to follow.
Also add `forcepush` to pkg and test/pkg and make `push` explicit.
The `dirs` phony was non-existent in all three cases. Likewise `$(DIR)` wasn't
used in that way in pkg or test/pkg and is nolonger used in that way in tools.
Finally, `clean` was bogus in all cases (tools cleaned a thing which never
exists and pkg and test/pkg were a syntax error).
Signed-off-by: Ian Campbell <ijc@docker.com>
golint on pkg/init now complains:
golint...
./init.go:199:2: redundant if ...; err != nil check, just return error instead.
Resulting in a change which doesn't seem like an improvement to me.
Signed-off-by: Ian Campbell <ijc@docker.com>
As discussed before, as we use this in three places, cloning in
base makes more sense.
Update base image.
Signed-off-by: Justin Cormack <justin.cormack@docker.com>
- xfsprogs is required for mkfs.xfs for some tests.
- mount persistent disk on /var/lib instead of /var.
- include host /dev (for loop devices) in mounts.
- /tmp need not be exec.
Signed-off-by: Ian Campbell <ijc@docker.com>
`${CONTAINERD_COMMIT}` is only defined in the `FROM linuxkit/alpine` section
and not the in `FROM scratch` section.
This means we were always checking out the current master at the time of build
rather than the expected commit because `git checkout $CONTAINERD_COMMIT` was
expanding to `git checkout `.
Signed-off-by: Ian Campbell <ijc@docker.com>
The following packages did not work on aarch64:
- linuxkit/test-docker-bench
- linuxkit/test-ltp
- linuxkit/test-ns
- linuxkit/test-virtsock
Signed-off-by: Rolf Neugebauer <rolf.neugebauer@docker.com>
Mounting a directory inside a read only container requires that to be
created in advance, but `runc` worked around that if the rootfs was not
originally read only.
You cannot even bind mount a file that does not exist into a
read only container.
The containerd test is given a disk, as running on an overlay does
not work; however it is also disabled as one of the parts of the test
is failing, needs investigation.
Signed-off-by: Justin Cormack <justin.cormack@docker.com>
The newest tests actually run containerd and therefore have some additional
requirements:
- containerd + shim + runc binaries are needed. We bind these in from the host.
The test code should, by design, be from matching containerd source, assuming
we remember to update test/pkg/container/Dockerfile when we bump
CONTAINERD_COMMIT. 5217b9973b added a reminder
to do so.
- the tests need networking (to pull images). So add dhcp to onboot and bind
/etc/resolv.conf into the test container.
- running containers requires a writeable cgroup mount.
- containerd wants /etc/localtime, so install the UTC one (as we do in
pkg/containerd).
The test image already has `net: host` and `capabilities: all`.
Signed-off-by: Ian Campbell <ijc@docker.com>
The logrus import path has changed, so adjust. Also there is a minor API change
to the containerd.IOCreation() function spec, it now takes a string id which we
can ignore.
Signed-off-by: Ian Campbell <ijc@docker.com>
with cwd of test this was done with:
for i in pkg/* ; do make --no-print-directory -C $i show-tag; done | ( IFS=: ; while read image hash ; do ../scripts/update-component-sha.sh --image $image $hash ; done )
Note that `linuxkit/test-virtsock` (built by `test/pkg/virtsock`) does not
appear to be referenced anywhere.
Signed-off-by: Ian Campbell <ijc@docker.com>
Makefile and Dockerfile are implicit from pacakge.mk.
Need to list the other files consumed by the Dockerfile though.
template.yml is only for manual testing and so is not a dependency of the
standard build.
Signed-off-by: Ian Campbell <ijc@docker.com>
This makes the package actually build reproducibly, with the downside that it
requires changing the hash. Perhaps this should move to tools/alpine.
Signed-off-by: Ian Campbell <ijc@docker.com>
The previous version just created a network name space which does
not allow us to also test additional namespaces, e.g. for unix
domain sockets.
This commit uses runc to create a fully namespaced container to
run a test in. It creates a container, configures the network
interfaces in the new network namespace before starting the
container.
A OCI config.json template is used and then customised for a
given test based on command line arguments.
Finally, instead of iperf, we use the socket stress test from
https://github.com/linuxkit/virtsock as it provides finer-grained
control over the traffic patterns (e.g. long lived vs lots of
short lived connections).
Signed-off-by: Rolf Neugebauer <rolf.neugebauer@docker.com>
This is what CONFIG_DEBUG_SET_MODULE_RONX was renamed to; since we want
that one, presumably we want this one too.
Signed-off-by: Tycho Andersen <tycho@docker.com>
The previous commit moved some filesystems to kernel modules.
modprobe them before checking if they are enabled.
Also update the YAML files for the tests.
Signed-off-by: Rolf Neugebauer <rolf.neugebauer@docker.com>
Make sure we do not remove filesystems we expect to have.
Fix the failure cases for the kernel tests which were not working properly
due to shell code.
Fix some 4.11 kernel changes in config that show up once tests are fixed.
Signed-off-by: Justin Cormack <justin.cormack@docker.com>
0eb21735ae accidentally broke some package
builds by switching linuxkit/alpine to linuxkit/containerd. Let's revert
the ones that shouldn't be there.
Closes#1991
Signed-off-by: Tycho Andersen <tycho@docker.com>
Note that this is not the latest (which was 95efd45db073 at time of writing)
but the next commit 6428b4bad0c2 merges "Port ctr to use client package" breaks
the use of `ctr run --runtime-config` (by removing that option).
This contains https://github.com/containerd/containerd/pull/954 which was
causing some services to fail to start.
All previous uses of 15541037b9 are updated to
5749f2e9e6.
Signed-off-by: Ian Campbell <ian.campbell@docker.com>
The remaining packages will be updated with subsequent commits
to also include the config label.
Signed-off-by: Rolf Neugebauer <rolf.neugebauer@docker.com>
The new version has improvements to the Windows code as well
as initial support for the VMware/virtio based Hyper-V socket
interface.
Signed-off-by: Rolf Neugebauer <rolf.neugebauer@docker.com>
