18:01:40 <h01ger> #startmeeting jenkins.debian.org 18:01:40 <MeetBot> Meeting started Wed Mar 22 18:01:40 2017 UTC. The chair is h01ger. Information about MeetBot at http://wiki.debian.org/MeetBot. 18:01:40 <MeetBot> Useful Commands: #action #agreed #help #info #idea #link #topic. 18:04:08 <KGB-0> 03Valerie R Young 05master ce00640 06jenkins.debian.net 10(6 files in 3 dirs) Revert "reproducible debian: show the "show all" button when summary tag supported" * 14https://deb.li/ZPuF 18:04:27 <mapreri> o/ 18:04:44 <h01ger> the agenda for todays meeting is in TODO in j.d.n.git 18:05:09 <h01ger> #topic say hi / agenda 18:05:33 <h01ger> so todays agenda is short: 18:06:21 <h01ger> 1. j.d.n status 2. j.d.o migration next steps 3. jenkins sprint 4. AOB 18:07:40 <h01ger> #topic jenkins.debian.net status 18:08:11 <h01ger> we have a new plugin to monitor jstat for the jenkins process https://jenkins.debian.net/munin/debian.net/jenkins.debian.net/jenkins_jstats.html 18:08:35 <h01ger> after writing this olasd pointed out, that munin-contrib has very similar plugins, which i should probably use insteac 18:08:39 <h01ger> instead 18:09:21 <h01ger> but the real problem is *understanding* those graphs and estimating how/if it can be useful to deal with the executor starvation on jenkins 18:10:03 <mapreri> tbh, I have no clue what those acronyms or whatever those few capital letters on the graph are. 18:10:22 <h01ger> i've also sketched some design notes how to run those 150 reproducible build jobs, which are constantly running, from (systemd-)cron instead of jenkins, wchih would also solve the executor starvation problem… i might get to work on this at the reproduicble hackathon in hamburg 18:10:41 <h01ger> mapreri: man jstats will tell you 18:11:31 <h01ger> man jstat even 18:12:20 <h01ger> actually no 18:12:55 <h01ger> https://docs.oracle.com/javase/6/docs/technotes/tools/share/jstat.html 18:13:05 <mapreri> yes, there are some explanation there 18:13:51 <h01ger> that oracle page has a bit more info than "man jstat" 18:14:01 <mapreri> but apart from "it's nice monitoring things", is there a reason for this? I reckon we're not seeing any memory-related problem, except (maybe, since we have no clue about what's going on?) the executors not starting. 18:14:36 <h01ger> my guess its that it might be garbage collection related, or thread execution or dunno 18:14:37 <mapreri> I'll try to go through the jstat docs 18:15:03 <h01ger> i also wanted to see whether jenkins.war has *enough* memory 18:15:31 <h01ger> #info java help still welcome. if you know java, please read the log & come around 18:15:50 <h01ger> the other topic coming to my mind is diskspace… 18:16:27 <mapreri> (on an aside, this semester I have a java-related course (curse, even), guess it'll come handy) 18:16:33 <h01ger> though i have nothing to say about this except that i should look into this 18:16:45 <h01ger> mapreri: poor you, lucky us! ;-) 18:18:55 <h01ger> anything else about jenkins.d.n? 18:19:10 <mapreri> not from me, everything's been very cool lately 18:20:25 <h01ger> :) 18:21:13 <h01ger> #save 18:21:21 <h01ger> #topic jenkins.d.o migration 18:24:17 <mapreri> I think we are stuck at the "somebody should make update_jdn run on jdo" step? 18:25:55 <h01ger> yup 18:26:02 <mapreri> then again, I wonder if for us it wouldn't be easier to write a different update_jdo, since most of update_jdn wouldn't be useful in any way, not being root there. 18:26:08 <h01ger> yup 18:26:11 <mapreri> juts thinking aloud now 18:26:13 <mapreri> just* 18:27:39 <h01ger> #info someone should start update_jdo ("rewrite" of update_jdn from scratch, with only 5% the functionality) 18:28:10 <h01ger> #info basic/main function: configure jobs with jenkins-job-builder. also deploy scripts + their config 18:29:12 * h01ger has put this higher on his agenda now 18:29:20 <h01ger> we need this start… 18:30:07 <h01ger> #topic 18:30:14 <h01ger> jenkins sprint spring 2017 18:30:20 <h01ger> #topic jenkins sprint spring 2017 18:30:32 <h01ger> a jenkins sprint in spring 2017 wont happen 18:30:49 <h01ger> however, there will be https://wiki.debian.org/ReproducibleBuilds/HamburgHackathon2017 18:31:13 <h01ger> to which i've also invited fil to come around, as he lives just a few subway stations away ;) 18:32:05 <mapreri> handy 18:32:10 <h01ger> during the sprint i probably rather work on importing guix's and fdroid's reproducible.json results, but i'll definitly also be open to and supportive of jenkins.d.o hacking 18:32:42 <h01ger> other people interested to work on jenkins.d.(n|o) are also welcome to join! 18:32:52 <mapreri> (as h01ger knows I still don't know if I will attend, as I have clashing events where other people expects me…) 18:34:14 * h01ger nods 18:34:24 <h01ger> #info https://wiki.debian.org/ReproducibleBuilds/HamburgHackathon2017 18:34:32 <h01ger> #info people interested to work on jenkins.d.(n|o) are also welcome to join! 18:34:36 <h01ger> #save 18:34:58 <h01ger> #topic AOB 18:35:06 <h01ger> (any other business) 18:40:30 <h01ger> i guess not :) 18:40:34 <h01ger> #endmeeting