Laura Brehm

I (and lots of others) use the same in Neovim (https://github.com/redhat-developer/yaml-language-server). This would definitely be a big QoL improvement, I'd be happy to review PRs adding these.

Closing since this was addressed by https://github.com/moby/moby/pull/45362

@dcarbone since your issue started more recently and doesn't seem related, please open a new issue with all the requested information + information on how to reproduce it so we...

Oooh, thanks, I can replicate this (though I to use a lower `d`). I think regardless of exit event order guarantees, we can probably try to harden things on the...

I've been a bit busy over the last few days, but in the meantime I've confirmed via bisect that https://github.com/containerd/containerd/commit/5cd6210ad055ec3a0fc263fdde382191c52c40ce is at fault (pre-this commit, we just had big locks...

Yeah, I've been looking and as thoroughly confused as you. However (regardless of _the mystery of who is killing the execs_), since we do consistently receive the exec's exit event...

I pushed a little non-super-polished example of a fix w/ the "wait to emit init exit if there are pending execs" approach here https://github.com/laurazard/containerd/commit/66ac08d89e68990016a3217ab66932d0062a777e?diff=split&w=1 for discussion. @corhere @dmcgowan @thaJeztah thoughts?

Sure! I admit I can't see what a cleaner fix would look like – but maybe my assumptions are wrong. What I'm thinking is: - we want the exec to...

Actually, after thinking about it for a bit I guess this (current) issue was introduced when attempting to fix another issue (lock contention due to locks present for preserving event...

Nice!! That's great :) I'll polish up that fix and prepare a PR.