Summary of IRC Meeting in #aurora at Mon Jun 1 18:05:57 2015: Attendees: thalin, dnorris, lexinator, jfarrell, marti_, Yasumoto, kts, mkhutornenko, serb, derfloh, benley
- Preface - Debian packaging IRC log follows: ## Preface ## [Mon Jun 1 18:06:14 2015] <kts>: good morning everyone and welcome to the weekly aurora community meeting [Mon Jun 1 18:06:34 2015] <kts>: let's get started with a roll call and a call for topics [Mon Jun 1 18:06:39 2015] <derfloh>: here [Mon Jun 1 18:06:42 2015] <marti_>: here [Mon Jun 1 18:06:50 2015] <mkhutornenko>: here [Mon Jun 1 18:06:56 2015] <serb>: here [Mon Jun 1 18:07:06 2015] <lexinator>: here [Mon Jun 1 18:08:16 2015] <kts>: does anyone have a topic they would like to discuss? [Mon Jun 1 18:08:25 2015] <Yasumoto>: howdy howdy [Mon Jun 1 18:08:50 2015] <marti_>: could I share something =? [Mon Jun 1 18:09:15 2015] <kts>: marti_: go ahead [Mon Jun 1 18:09:48 2015] <marti_>: ok :) I share you my gsoc project https://github.com/twitter/cli-guide.js [Mon Jun 1 18:09:56 2015] <marti_>: demo: http://marti1125.webfactional.com/cli-guide/ [Mon Jun 1 18:11:22 2015] <thalin>: here [Mon Jun 1 18:11:22 2015] <marti_>: in editor you can write nano and see an example of editor... ctlr + x for scape... [Mon Jun 1 18:12:29 2015] <Yasumoto>: marti_: awesome! [Mon Jun 1 18:12:55 2015] <marti_>: :D [Mon Jun 1 18:13:58 2015] <kts>: that looks neat, more to come? [Mon Jun 1 18:15:02 2015] <kts>: thank you for sharing, marti_ [Mon Jun 1 18:15:31 2015] <kts>: does anyone else have a topic they'd like to discuss? [Mon Jun 1 18:16:28 2015] <serb>: Question regarding this bug I've filled: https://issues.apache.org/jira/browse/AURORA-1335 [Mon Jun 1 18:16:50 2015] <serb>: Is it intended that the kill is send almost immediately after the term of a process? [Mon Jun 1 18:16:56 2015] <derfloh>: vvhurirgfurcvbvfdggbiikjgjlrcckgeeertjuuubtv [Mon Jun 1 18:17:51 2015] <jfarrell>: afternoon everyone [Mon Jun 1 18:18:26 2015] <kts>: dlester: I think that's a legitimate bug - my understanding of the behavior was that we gradually escalate signals in 10sec intervals [Mon Jun 1 18:18:42 2015] <kts>: er 10secs might not be the timeout, but some constant amount of time [Mon Jun 1 18:19:25 2015] <kts>: er serb [Mon Jun 1 18:19:37 2015] <serb>: ok thanks, maybe I have time to dig into it [Mon Jun 1 18:20:27 2015] <kts>: regarding the inherent security problems with the http signaler, would you feel okay about it if we did an X-style thing and had the executor set a random environment variable and pass that in a header? [Mon Jun 1 18:22:22 2015] <serb>: Might work as well. However this sounds a little bit like re-inventing something already readily available for any linux process via SIGTERM and SIGKILL [Mon Jun 1 18:23:47 2015] <kts>: the main advantage imo is that you have more control over how/where the signal handler executes (this is more of a library thing) [Mon Jun 1 18:24:11 2015] <kts>: but yes, please feel free to investigate a patch, and ask here if you need any help [Mon Jun 1 18:24:37 2015] <kts>: moving on to the next topic ## Debian packaging ## [Mon Jun 1 18:24:46 2015] <kts>: jfarrell: you have the floor [Mon Jun 1 18:25:20 2015] <benley>: I'm here too now, but I don't actually have an update on deb packaging. [Mon Jun 1 18:25:41 2015] <jfarrell>: iâd like to get thoughs on cutting a 0.8.1 or 0.9.0 when this patch hits [Mon Jun 1 18:27:09 2015] <jfarrell>: will have a script for build-support on release to push this so it can be used against the apache deb package repo [Mon Jun 1 18:27:10 2015] <kts>: doing 0.9.0 is fine with me, worst case we'll have deprecations that we don't complete [Mon Jun 1 18:27:35 2015] <kts>: are we prevented from releasing debs for 0.8.0? [Mon Jun 1 18:27:46 2015] <jfarrell>: will be pushing thrift there as well so we can switch to consuming it from there instead of from my home directory [Mon Jun 1 18:27:56 2015] <benley>: yeah, I was wondering the same. Isn't packaging kindof a meta release thing? [Mon Jun 1 18:28:14 2015] <dnorris>: it should be part of the release process, no? [Mon Jun 1 18:28:17 2015] <jfarrell>: we dont have the code in trunk currently that makes those packages, so would have to commit that and that would cause the point release [Mon Jun 1 18:28:49 2015] <benley>: Since the plan is to have the packaging metadata bits directly in trunk, I suppose that does make it intrinsically a part of the release process. [Mon Jun 1 18:28:54 2015] <serb>: Yeah, that's true. But I believe the bug somewhat remains and HTTP kill escalation can be a goodie on top [Mon Jun 1 18:29:12 2015] <jfarrell>: so if its 0.8.1 or 0.9.0 up for debate, but wanted to cut it and get it in and available. if we want to try and do a monthly release on this [Mon Jun 1 18:29:52 2015] <jfarrell>: java 1.8, deb, and handful of other tickets over the next say two weeks [Mon Jun 1 18:30:06 2015] <jfarrell>: happy to fill in as RM again for this [Mon Jun 1 18:30:33 2015] <kts>: +1 [Mon Jun 1 18:31:05 2015] <kts>: if we require jdk8 it should be 0.9.0 I think [Mon Jun 1 18:31:24 2015] <jfarrell>: lets plan on rc in the next week/2 weeks then, should give us enough time to get the deb patch finished and in as well as any others we want to group [Mon Jun 1 18:31:27 2015] <jfarrell>: ok 0.9.0 it is [Mon Jun 1 18:31:32 2015] <benley>: sounds good [Mon Jun 1 18:31:56 2015] <benley>: I'll definitely be able to allocate time in the next 2 weeks to getting the deb patch merged [Mon Jun 1 18:32:14 2015] <dnorris>: sounds great, Iâve built debs ussing benleyâs patch, still in the testing phase internally though [Mon Jun 1 18:32:39 2015] <dnorris>: happy to help w/ testing or whatever is needed to push that through [Mon Jun 1 18:32:41 2015] <benley>: dnorris: please let me know if you encounter packaging bugs! [Mon Jun 1 18:32:42 2015] <kts>: any other topics? [Mon Jun 1 18:33:10 2015] <jfarrell>: think we are good, thanks everyone [Mon Jun 1 18:33:37 2015] <kts>: thanks everyone [Mon Jun 1 18:33:40 2015] <kts>: ASFBot: meeting stop Meeting ended at Mon Jun 1 18:33:40 2015
