Dockerfile: move into separate group when running under cgroupsv2

Per the comments in

 https://github.com/containers/podman/issues/14884

there is basically no way to run podman nested in the container in a
cgroups v2 environment (e.g. Ubuntu Jammy) with the processes in the
same context the container starts in.

One option is to run systemd in the container, which puts things in
separate slices, etc.  This is unappealing.

This takes what I think is the simplest approach which is to check if
we're under cgroups v2 and move everything into a new group before
nodepool-builder starts.

The referenced change tests this by running the containerfile elements
on Jammy.

Neded-By: https://review.opendev.org/c/openstack/diskimage-builder/+/849274

Change-Id: Ie663d01d77e17f560a92887cba1e2c86b421b24d
This commit is contained in:
Ian Wienand 2022-07-11 16:16:06 +10:00
parent bd6f610113
commit 6c59182428

View File

@ -138,6 +138,13 @@ RUN \
apt-get clean \ apt-get clean \
&& rm -rf /var/lib/apt/lists/* && rm -rf /var/lib/apt/lists/*
# NOTE(ianw) 2022-08-02 : move this into its own cgroup on cgroupsv2
# hosts for nested podman calls to work; see comments in
# https://github.com/containers/podman/issues/14884
CMD _DAEMON_FLAG=${DEBUG:+-d} && \ CMD _DAEMON_FLAG=${DEBUG:+-d} && \
_DAEMON_FLAG=${_DAEMON_FLAG:--f} && \ _DAEMON_FLAG=${_DAEMON_FLAG:--f} && \
if [ -e /sys/fs/cgroup/cgroup.controllers ]; then \
sudo mkdir /sys/fs/cgroup/nodepool && \
for p in `cat /sys/fs/cgroup/cgroup.procs`; do echo $p | sudo tee /sys/fs/cgroup/nodepool/cgroup.procs || true; done \
fi; \
/usr/local/bin/nodepool-builder ${_DAEMON_FLAG} /usr/local/bin/nodepool-builder ${_DAEMON_FLAG}