Nell Boulle

Hmph, nspawn mounting a tmpfs that's masking it I guess On Oct 24, 2015 5:26 PM, "Chance Zibolski" [email protected] wrote: > Example: > > root@core-01:/home/core/testcore-01 test # echo "test" >...

@krobertson what do you do if an image contains a (pre-populated) /tmp?

Hm, does someone want to put up a PR for that? On Wed, Oct 28, 2015 at 8:29 AM Ken Robertson [email protected] wrote: > Until recently, we were mounting a...

@alban That is what I had been thinking, see https://github.com/appc/acbuild/blob/master/README.md#related-work Do you have any concerns about this?

Sure. With my experiences trying to get actool vendored properly with go1.5, I'm starting to think we should instead vendor acbuild/lib and use that programmatically instead... On Tue, Oct 13,...

I don't really see why that should be necessary. What happened when you didn't have it provided?

ok, I bet @dgonyeo can investigate :-) On Thu, May 19, 2016 at 6:06 PM, Faraz Fazli [email protected] wrote: > @jonboulle https://github.com/jonboulle The container wasn't able to > write to...

Waiting on acpush.

The best way to start from an existing root filesystem is to point begin at a directory: https://github.com/containers/build/blob/master/Documentation/subcommands/begin.md#starting-with-a-local-pre-existing-rootfs You could also probably achieve something similar with [`copy-to-dir`](https://github.com/containers/build/blob/master/Documentation/subcommands/copy-to-dir.md) which works differently...

@dgonyeo how about a container based on CoreOS or something to pick up a more modern systemd-nspawn? That check was removed a while ago: https://github.com/systemd/systemd/commit/4f923a1984476de3441922ee5bf7102ebdd250ef