Aleksa Sarai

Grr, it seems I didn't send this comment earlier when I looked at this... ``` mount --bind -o remount,diratime,strictatime "$DIR" ``` resets the mount flags on the mount. `strictatime` is...

@rata I suggested doing a `umount` above: > Looking at the failures, it seems that there's something weird going on with atime-related mount flags on your system. We could rework...

Ah, I got confused by the fact that the `util-linux` package doesn't contain `/usr/bin/mount` and then looked for the source of the `mount` package from sources.debian.org. To be honest, I've...

Seems like a Docker/containerd issue? runc outputs the error to stderr, it's up to Docker to format it as necessary. I suspect the issue is that our logs go to...

Reading through my (8-year-old :cold_sweat:) comments, I think that there were three issues at play here: 1. A value of `0` being identical to `1` is a quirk of the...

> To be honest I don't think we can have a hard guarantee for memory usage, peak memory usage is not deterministic for one (and, as evidenced by the CentOS...

@rata That is very similar to how runc currently works. We fork a child (`runc init`) and we give it a parsed form of the container configuration. We have to...

The runc binary size doesn't affect the cgroup limit. I can run a 2.5mb limit container (with `RUNC_DMZ=legacy` no less) on my machine. If the runc binary size had any...

`0` in `runc.spec` should be a limit of 0, which the kernel treats like a limit of 1 in most cases. Unset in `config.json` should mean leave. I looked at...

#3931 and #3876 are both ready to merge IMHO. I would be _fine_ with a -rc1 release with just those if we really do need to rush it that badly....