oscontainer: Use skopeo, not podman to extract #1671
Conversation
src/oscontainer.py
Outdated
| subprocess.check_call(['find', '-name', '*.tar', '-exec', 'tar', 'xUf', '{}', ';'], cwd=tmpdir) | ||
| subprocess.check_call(['find', '!', '-perm', '-u+w', '-exec', 'chmod', 'u+w', '{}', ';'], cwd=tmpdir) |
There was a problem hiding this comment.
Hmm I think I follow, though some comments for these two would be helpful.
There was a problem hiding this comment.
I'm actually surprised there isn't a command to just extract all the layers for you into a custom directory. That would also ensure it unpacks the layers in the right order. Might be a good RFE for skopeo.
There was a problem hiding this comment.
Or probably podman image extract - it seems like oc isn't using containers/storage (which is the source of the problems here when run nested). Probably this code could get lowered into podman and then oc could vendor it or so.
|
Using I'd proffer a lgtm, but its Friday. @cgwalters merge when ready. |
|
/approve |
I kept running into failures trying to run this in an unprivileged container; basically the containers stack seems to want to try to `chown()` files based on the UIDs, but we don't care about that at all for this. What we really want is the functionality of `oc image extract` but that doesn't exist yet in the podman stack and today cosa doesn't include `oc` (that's a big discussion).
cgwalters
force-pushed
the
oscontainer-extract
branch
from
bd49446 to
da57979
Compare
|
/lgtm |
|
[APPROVALNOTIFIER] This PR is APPROVED This pull-request has been approved by: cgwalters, darkmuggle, jlebon The full list of commands accepted by this bot can be found here. The pull request process is described here
Needs approval from an approver in each of these files:
Approvers can indicate their approval by writing |
I kept running into failures trying to run this in an
unprivileged container; basically the containers stack
seems to want to try to
chown()files based on the UIDs,but we don't care about that at all for this.
What we really want is the functionality of
oc image extractbut that doesn't exist yet in the podman stack and today
cosa doesn't include
oc(that's a big discussion).