Skip to content
This repository was archived by the owner on Dec 10, 2019. It is now read-only.

Latest commit

 

History

History
50 lines (32 loc) · 2.35 KB

2014-08-07--17-30.md

File metadata and controls

50 lines (32 loc) · 2.35 KB

US TZ irc meeting.

#docker-dev meetings are held twice a week in the #docker-dev irc group on Freenode. This irc channel is logged all the time, so you can always refer to it.

Held at 10:30am Thursday, San Francisco time.

Please remember that there are now 2 meetings, one in an Asia-Pacific timezone, and this one in the SF timezone.

To scale this sysem we need to gradually step away from mandatory synchronous meetings as a primary tool I would like to experiment with monthly

The goal is to give everyone, especially maintainers, some breathing space to do things with more signal/noise ratio, like reviewing design props and PRs, short onthefly irc sync, and of course writing code :)

Also we should clearly designate the meeting as "non-binding". It's an opportunity to discuss But not where important decisions are taken - that's always in the repo

cpuguy83 is really concerned that we got rid of "rm -f" too hastily, and that we have a short window here before 1.2.0 to change this

so I'd just ask that maintainers take a look and weigh in whether they think "rm -f" or "rm --stop" / "rm --kill" are a better UI / UX ;)

-f is gone today; deprecated there needs to be some context here docker rm -f is super slow if you have a lot of containers because it is trying to stop and not kill

anyway, after some discussion on the issue we agreed we would introduce -s for stop and -k for kill and deprecate -f, in order to be more explicit and the PR got merged it's not in a release yet, but it's in master

also, --timeout is like -t in docker stop and in restart

i think the wisest thing to do right now is to revert the -s and -k change and not introduce --timeout

  • tianon: I'm +1 to reverting the change too, and I'm also -1 on --timeout

  • erikh: we currently send STOP, wait 10 or 15 seconds, then send KILL

  • lk4d4: we can try to send sigterm and sigkill without timeout

  • erikh: ^ is what I was after.