20:00:50 #startmeeting diskimage-builder 20:00:51 Meeting started Thu Nov 30 20:00:50 2017 UTC and is due to finish in 60 minutes. The chair is ianw. Information about MeetBot at http://wiki.debian.org/MeetBot. 20:00:53 Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 20:00:55 The meeting name has been set to 'diskimage_builder' 20:01:05 #link https://wiki.openstack.org/wiki/Meetings/diskimage-builder 20:01:21 #topic Announcements 20:02:56 The gate seems as stable as usual after a short period of issues after we tied up the repos from fallback 20:03:25 I'm still working on full zuulv3-ification of the func tests 20:04:49 basics are there, but i'd really like to split up the log files for each test 20:05:13 #topic Action items from previous meeting 20:06:57 it's been a while, with a two week alternating cadence, summit and no meetings 20:07:18 #topic s390x support for DIB 20:08:28 this started out worring with magic unlicensed perl but in it's current state is simply a bootloader type element that uses the environment variables exposed by dib-block-device 20:09:35 is this what the third party ci is testing as well? 20:10:34 it's suggested that s390 3rd party ci could start reporting on dib 20:12:10 we either have to reject this and say it should live outside, or accept it. per my comments in the review, i think after the changes it's now fine, and with no other comments i think it's time to merge 20:13:18 and done 20:14:11 Revert "Fix #1627402: fix grub2 element for ubuntu xenial" 20:14:27 #link https://review.openstack.org/524019 20:15:21 clearly this is not used very much, but we did have inc0 come in and notice the same thing 20:17:48 grub2 is slightly different to the "bootloader" element as it install the efi elements. i don't know if that's intentional or just happenstance, the logs are unclear on the motivations 20:18:11 efi packages i mean 20:18:17 there is also confusion around the baremetal element related to that 20:20:33 so my thinking is to commit that, and at least restore to the last known "good" state, and if someone comes along wanting to modify it again we'll be in a better position to insist on some morme thorough review 20:20:45 Fix /dev/pts mount options handling 20:20:50 ianw: sounds good to me 20:20:54 #link https://review.openstack.org/#/c/522654/ 20:21:03 this is all a mess 20:21:47 it seems dib is the only tool in this space having to do all this magic which makes me very worried, and this is the 3rd or 4th fix up patch 20:23:57 the whole thing started in cebfcf85f9034522a6521c40e1ae9c6d559f585f "Use -t devpts for /dev/pts mounts" 20:24:18 It was observed that the /dev/pts mount vanishes from the host system. 20:24:43 when running in parallel 20:27:33 i guess we'll go with this for now, but any more problems and i think i'm going to insist we just go back to a --bind mount 20:28:17 because, if it's the same kernel issue that makes the /dev/pts points disappear underneath the bind mounts (which i suspect it is) then i will just say "we don't support parallel building on these kernels" 20:30:12 block-device lvm: fix umount phase 20:30:23 #link https://review.openstack.org/503958 20:30:32 i need to review this, no comments at this point 20:32:23 Update Fedora defaults to 27 20:32:25 https://review.openstack.org/521684 20:32:27 #link https://review.openstack.org/521684 20:32:55 slight WIP -- i need to add the f27 mirrors, and as part of that i'll clear out f25 remnants 20:33:36 however, i had a f27 + infra elements built and working in local cloud so that should be good. it remains a good target for python3.6 unit testing 20:34:09 that will be the last thing i'd like in before a release 20:34:33 ergo i imagine a release early next week 20:38:10 cool 20:39:14 ok, that's about it for the monthly clear out, till next time ... 20:39:17 #endmeeting