[00:39] ok, systemd_timesyncd and chrony done; [00:39] now I'm done, tomorrow fixing up py26 unittests [00:39] *sigh* [01:38] blackboxsw: i would probably not add the new depends into x, a [01:39] it probably *is* strictly required. [01:39] but because isc-dhcp-client is part of ubuntu-minimal [01:40] and you're really expected to create a ubuntu without 'ubuntu-minimal' [01:40] it is not likely that cloud-init is to be installed in such a place. [01:43] so.. i'd just leave it be. additionally, adding depends or recommends like that to a stable can be problematic [01:43] see someone complaining about a similar add [01:43] https://bugs.launchpad.net/ubuntu/+source/initramfs-tools/+bug/1633643 [01:43] Ubuntu bug 1633643 in initramfs-tools (Ubuntu) "unnecessary dependency upon isc-dhcp-client" [Undecided,Invalid] [01:44] and then also the squashfuse bug (bug 1628289) [01:44] bug 1628289 in Snappy "snapd should depend on squashfuse (for use in containers)" [Undecided,In progress] https://launchpad.net/bugs/1628289 [01:53] +1 on dropping isc-dhcp-client depends [01:53] will repost the MPs without that [02:01] smoser: thanks for the reference to https://bugs.launchpad.net/ubuntu/+source/snapd/+bug/1756173 [02:01] Ubuntu bug 1756173 in snapd (Ubuntu Artful) "[SRU] 2.32" [Undecided,New] [02:03] I've added a card https://trello.com/c/OTYKbFEu/720-drop-squashfuse-install-logic-from-ccsnap-when-https-bugslaunchpadnet-ubuntu-source-snapd-bug-1756173-lands [02:29] pushed ubuntu/xenial and ubuntu/artful without isc-dhcp-client explicit dependency === shardy is now known as shardy_lunch [13:53] smoser: ./tools/run-centos is super handy =) [13:55] yeah. we need to make it work for suse too [13:56] and actually.. i've even wnated it to work for ubuntu [13:56] when there was some bug trhat didnt reproduce on my desktoip, but c-i saw. [14:05] interesting === shardy_lunch is now known as shardy === r-daneel_ is now known as r-daneel [17:01] blackboxsw: ok... can you give qa-scripts/new-upstream-snapshot a try ? [17:01] you can see the last commit message there. [17:02] i think i have addressed most of the pain in it. [17:02] checking/updating now [17:02] and thanks! [17:03] smoser: we'll drop new-upstream-snapshot script after the SRU or as part of the SRU? [17:03] in the ubuntu/xenial|artful branches [17:03] i think no hurry to drop it. [17:03] so i'd say after. [17:03] but if you want to drop it we can. [17:06] * blackboxsw checks out the refresh patches logic. that was a surprise [17:07] thats why your xenial didnt work [17:07] because the patch would only apply with fuzz [17:07] as we'd changed ds-identify aroudn the xenial patches [17:09] gotcha, yeah I just didn't expect to see a separate commit message for it, and changelog entry. Both make sense as separate items, I just wanted to understand the quilt magic [17:13] smoser: pushed just ubuntu/xenial branch again [17:13] shall I push artful again? [17:14] new-upstream-snapshot works like a charm [17:15] blackboxsw: i uploaded artful all ready. [17:15] ahh hadn't checked, thx [17:15] i don tthink any reason to re-do it. [17:16] changelog didn't end up changing anything, so good there on artful anyway [17:26] I'd like to test some changes in user-data after applying some changes to user-data. So far, I've mostly used cloud-init w/ qemu and genisoimage for testing, but I think there should be an easier way. I've tested w/ manual changes to cloud-config.txt/user-data.txt/user-data.txt.1 in /var/lib/cloud/instances/$instance-id/, and by reattaching the cdrom through the qemu monitor, yet changes aren't picked up. [17:26] Which file on disk should I modify to be able to rerun cloud-init without having to go through the qemu/genisoimage steps? [17:26] ihre: there are several semaphores that block re-running specific cloud-init modules across reboots [17:27] you can clear out those semaphores by running "sudo cloud-init clean --logs --reboot" [17:27] it'll remove all cloud-init former run logs and reboot the machine so cloud-init can re-run "fresh" with any user data you've provided [17:28] I generally iterate using lxc on ubuntu [17:28] via lxc launch ubuntu-daily:bionic myb1 [17:29] lxc exec myb1 bash # then manipulate user-data in /var/lib/cloud/seed/nocloud-net/user-data [17:30] in lxcs, the no-cloud/seed directory is kept across reboots. I'll check how the qemu image detected re-run to better answer your question [17:31] hrm, I actually have to run on an errand, will get back to that in ~45 mins [17:32] blackboxsw: Thanks, didnt think of using a container to test it, that seems a nice alternative as well. [17:32] Previously, I just removed a/the semaphore(s) by hand and reran cloud-init modules --mode final. [17:33] it's at least the fastest iterative method I've found for dev-test cycles. [17:33] * blackboxsw heads out [17:33] No rush on my end, the current way works fine though, it just feels a little cumbersome at times, anyways, thanks for taking the time to look into it :) [17:39] blackboxsw: your quilt refrsh differs in behavior from mine. [17:40] blackboxsw: smoser: it's sensitive to ~/.quiltrc [17:40] use --quiltrc - [17:40] (just a guess) [17:42] yeah. [17:44] we have to do that in git-ubuntu [17:44] smoser: i'm guessing this is about reproducibility? [17:49] blackboxsw: grab tip of qa-scripts and try again, would you ? [17:50] nacc: well, yeah. [17:50] [17:50] http://paste.ubuntu.com/p/3BwC5n3v4P/ [17:50] smoser: yeah [17:50] we run 'new-upstream-snapshot' to create a new upstream snapshot of cloud-init. [17:50] and that is a run here versus on chad's system. [17:50] yeah, you want your snapshot to be the same as chad [17:51] similar idea for us, we want our same-argument builds to be contentfully the same === r-daneel_ is now known as r-daneel [18:33] smoser: just pushed https://code.launchpad.net/~chad.smith/cloud-init/+git/cloud-init/+ref/ubuntu/xenial [18:36] blackboxsw: woot. [18:36] mine to yours : http://paste.ubuntu.com/p/spBd2CCrKg/ [18:36] new feature there to strip bug numbers on sru [18:36] and make you give an SRU bug [18:53] ahh right I totally forgot to strip bug#'s on artful [18:56] smoser: what should I do w.r.t. devel/arful [18:56] smoser: what should I do w.r.t. devel/artful. should we strip bug#'s [18:58] blackboxsw: oh. shoot. yeah i guess we should. right ? [18:58] we are supposed to. [18:58] we can just re-upload [18:58] yeah, I think so, yeah I'll repropose and do that right. [18:58] forgot that part of the process obviously [18:59] well now the script does the right thing [18:59] you forgot, i forgot too. [19:11] https://code.launchpad.net/~chad.smith/cloud-init/+git/cloud-init/+merge/342329 just --forced a new-upstream-release for artful [19:12] merge conflicts against existing ubuntu/artful. [19:13] for all the bug comment drops [19:13] blackboxsw: how bout this.. [19:16] git checkout ubuntu/artful [19:16] git reset --hard ubuntu/17.2-35-gf576b2a2-0ubuntu1_17.10.2 [19:16] new-upstream-snapshot [19:16] dch --release --distribution=artful [19:16] git push chad.smith HEAD --force [19:16] basically, lets just re-do it. [19:16] yeah I did basically that. [19:16] git checkout -b ubuntu/artful a2ce41af52c619c13da66f84ee2a789b9df8f3a2 [19:16] and re-ran everything [19:16] I'll use your git-fu here and try again [19:17] yours is probably ok now. [19:17] let me see. i thought you were saying it wasnt [19:17] In the UI I see merge conflicts [19:17] in debian/changelog [19:17] but all the other content is good [19:19] yours is good. [19:19] k. i'll just build and push and push over upstream [19:21] excellent. and the moral if this story is: Chad don't push branches late in the evening [19:22] i'm quite happy with those changes to that script. [19:22] i think we have covered most of the pain poitns of it. [19:23] smoser: I love it. though I throw eggs at any shell script. [19:23] * blackboxsw will become better versed in shell someday. [19:26] blackboxsw: fair enough. it is lots of variables and such in there. definitely "shell script" [19:26] smoser: right so although you pushed the upload, nothing showed in artful-proposed yet so we're safe to re-push right? [19:27] I'm looking for the proposed build queue in launchpad again [19:27] https://launchpad.net/ubuntu/xenial/+queue?queue_state=1&queue_text=cloud-init [19:27] or https://launchpad.net/ubuntu/artful/+queue?queue_state=1&queue_text=cloud-init rather [19:28] i just uploaded artful [19:29] ok great so that list tells me you just re-submitted same unapproved upload 30 seconds ago [19:29] yep [19:29] now do we ping in ubuntu-release for attention on artful/xenial? [19:30] we can, yeah. [19:30] will do. [19:43] ok smoser I'm going to get to work on updating out manual cloud test scripts for this SRU https://hackmd.io/jlq3C4qbSgurZ_DZ5GTiuw?both [20:25] smoser: I hadn't tested review-mps yet on non-master branches. I didn't want to automatically botch the merges [20:26] all 3 branches are approved [20:26] I can use review-mps to land them later if you don't land them now === r-daneel_ is now known as r-daneel [20:58] thanks for the tip re: lxc/lxd blackboxsw, that is definitely an easier workflow than using qemu :) [20:59] good to hear ihre, I'll probably write up a blog post about it (and use of other cloud-init CLI commands=) === r-daneel_ is now known as r-daneel [21:03] i'll keep an eye out for it! [21:51] ok smoser landed your ubuntu/(artful|xenial) mps and fixed up qa-scripts:review-mps tool to properly handle --upstream-branch origin/ubuntu/* [22:09] thanks for the cloud-init release email BTW https://lists.launchpad.net/cloud-init/msg00145.html === blackboxsw changed the topic of #cloud-init to: Reviews: http://bit.ly/ci-reviews | Meeting minutes: https://goo.gl/mrHdaj | Next status meeting: Monday 4/2 16:00 UTC | cloud-init 18.2 released (03/28/2018)