[00:22] <blackboxsw> personally I like seeing your separate commits post my last review. there's gotta be a way to do that once someone has already git push --force'd to a remote.
[00:23] <blackboxsw> I guess I could just git diff rhrper-repo/branch-name and see it before I pull
[00:24] <blackboxsw> personally I like seeing your separate commits after my most recent review.*
[01:41] <smoser> rharper: so templates/chrony.conf.debian.tmpl:
[01:41] <smoser> has a non-ascii char in it.
[01:41] <smoser> and as yo udiagnosed, that is getting decoded somewhere as ascii in python27 jinja2
[01:41] <smoser> and it just isnt ascii
[02:00] <smoser> http://paste.ubuntu.com/p/c5jxV6Rg45/
[02:00] <smoser> that recreates the basic failure
[02:04] <smoser> https://stackoverflow.com/questions/22181944/using-utf-8-characters-in-a-jinja2-template
[02:06] <smoser> well, ignore that. that failure wasnt right. but its somewhere around there.
[08:25] <apollo13> blackboxsw, rharper: thanks, any reason to prefer manual_cache_clean over /etc/cloud/cloud-init.disabled or just personal pref?
[12:57] <rcj> rharper: I see that you're working on chrony support (according to blackboxsw in https://cloud-init.github.io/status-2018-03-19.html#status-2018-03-19) Will that land for bionic?
[13:43] <smoser> rcj: yes. will land today
[13:43] <smoser> rharper: we need your network-ipv6 fix today too
[13:44] <smoser> rcj: but the images to my knowledge do not have chrony installed.
[13:45] <smoser> cloud-init will not install chrony if systemd-timesyncd is present, it will just configure that.
[13:57] <rharper> smoser: network-ipv6?
[14:01] <smoser> 10 second timeout
[14:02] <rharper> oh, netplan
[14:02] <rharper> yes, that would be really good to get an upload into bionic at least
[14:02] <rharper> even if it needs an SRU to artful
[14:03] <rharper> I have the PR in
[14:03] <cyphermox> I'll review in a bit and upload that with some other bugfixes
[14:03] <rharper> thx
[14:05] <smoser> rharper: i had thought that that was in cloud-init ... sorry i missed rthat.
[14:06] <rharper> smoser: I had a branch to disable-ra in cloud-init but after discussion with stgraber I realized the underlying issue was related to how systemd-networkd handles RA
[14:06] <rharper> and what the default netplan setting was
[14:06] <rharper> so that's probably why you thought we had some cloud-init work to do
[14:26] <smoser> rharper: and yes... thank you for seeing that. i had in 2 places seen the ~ 13 second network config and just thought "wow, their dhcp server is really slow"
[14:27] <rharper> yeah; it actually come up under the autopackage test failures on bionic in lxd containers because the network wasn't yet on line and apt update/install failed
[14:27] <smoser> well,. that needs fixing :)
[14:28] <smoser> sleep 5 && echo "everything ready!"
[14:28] <rharper> they added some retry logic
[14:29] <rharper> https://lists.ubuntu.com/archives/ubuntu-devel/2018-February/040138.html
[14:36] <smoser> very much a need in Ubuntu for "wait until system is booted" command.
[14:41] <rharper> folks disagree what "booted" means; I think that thread demonstrates that
[14:41] <Cyclohexane> Is there a known issue with any of these packages? https://pastebin.com/kwT1Fs8c I'm following https://docs.aws.amazon.com/AmazonVPC/latest/UserGuide/vpc-migrate-ipv6.html#ipv6-dhcpv6-rhel and it works absolutely fine on a fresh install, then I run yum update and the whole thing goes caput, networking dies and the server becomes inaccessible...
[14:43] <rharper> did you reboot or just after the yum update completes ?
[14:44] <rharper> cloud-init isn't active after boot, so an update to cloud-init won't affect a running instance;
[14:46] <Cyclohexane> i reboot yeah, as the amazon guide says
[14:47] <rharper> cloud-init at that level isn't going to affect the network config; it's just dhcp on eth0
[14:48] <rharper> if you look at the instance console-log, that should show if cloud-init ran, it dumps the network state of the instance
[14:50] <Cyclohexane> yeah cloud-init runs, but and updates per 99-custom-network.cfg but then when cloud-init runs netstat -rn the route table is dead so it loses connectivity
[14:50] <Cyclohexane> it works fine prior to running yum update, can reboot and it comes back up with ipv6 connectivity
[14:51] <rharper> maybe the dhcp update then ?
[14:51] <rharper> I suspect you may need to bisect your update ; upgrade a few at a time to see if keeps networking, or not
[15:09] <smoser> http://paste.ubuntu.com/p/FsDxksfVS3/
[15:24] <smoser> rharper: $ grep -r can.t templates/
=== dpb1_ is now known as dpb1
[16:02] <Cyclohexane> rharper: it seems to be this https://bugs.centos.org/view.php?id=14585
[16:08] <rharper> smoser: ok, pushed an update to the branch to update template and drop the related changes
[16:11] <rharper> Cyclohexane: yeah, not much info there;   you could 1) yum upgrade  2) make a copy of /etc/sysconfig/{network,network-scripts}  3) add the 99-custom-networking.cfg to /etc/cloud/cloud.cfg.d/ and then 4) cloud-init --force --debug init --local  which will re-run the initial cloud-init stage that renders network config;  then compare /etc/sysconfig/{network, network-scripts} contents with your copy and see whats different
[16:28] <Cyclohexane> rharper: https://gist.githubusercontent.com/bytestream/cb6fa966875b902cdb34986047eca1b3/raw/1183714df137a7df9441600f242cab46df52ffa7/gistfile1.txt
[16:34] <rharper> Cyclohexane: if /etc/sysconfig/network-scripts/ifcfg-eth0 is the same before and after, then it's not the config that cloud-init is generating
[16:37] <rharper> Cyclohexane: if possible, you could add a second interface and configure that with a public ip, or a second vm in the same VPC both with dual nics in the same network; you could  then hop through the second interface to see what things look like after reboot
[16:39] <blackboxsw> smoser: we good on https://code.launchpad.net/~chad.smith/cloud-init/+git/cloud-init/+merge/342761?
[16:41] <Cyclohexane> rharper: it's something to do with changes between cloud-init-0.7.9-9.el7.centos.2.x86_64 and cloud-init-0.7.9-9.el7.centos.6.x86_64, I downgraded cloud-init after yum update back to .2 and it's working fine again
[16:49] <dpb1> Cyclohexane: you should file a bug, or add to that one you found already
[17:22] <smoser> blackboxsw: acked.
[17:22] <smoser> tahnks
[17:27] <smoser> rharper: you must have cherry picked my commit from trunk ?
[17:27] <smoser> 0f7745619ab0a61d7dee5bde43e1e970ddf4a9b6 is on your branch but is
[17:28] <smoser> never mind
[17:28] <smoser> noise here
[17:45] <smoser> rharper: responded on your mp
[17:46] <smoser> i like it. thank you
[18:13] <rharper> smoser: checking, thanks
[18:27] <rharper> smoser: fixed  https://code.launchpad.net/~raharper/cloud-init/+git/cloud-init/+merge/339438
[19:30] <smoser> blackboxsw:
[19:37] <rharper> https://code.launchpad.net/~raharper/cloud-init/+git/cloud-init/+merge/343122
[19:49] <smoser> rharper: isnt' there some argparse -> bash completion ?
[19:49] <rharper> smoser: maybe
[19:49] <rharper> I'll look
[19:54] <blackboxsw> ooooh raharper... nice I did bash completion for landscape, will check
[19:54] <rharper> blackboxsw: yeah, dpb1 mentioned that
[20:09] <blackboxsw> rharper: yeah per smoser's comment python3-argcomplete might do some of this work for us, and doesn't add any more dependencies besides that package
[20:09] <rharper> well, I was interested in static generate of the file
[20:10] <rharper> like build-time run argparse to shell code and pack that up
[20:10] <blackboxsw> time-saver :)
[20:10] <rharper> rather than having python runtime during tab-tab
[20:10] <smoser> yeah. you'd think something like that would exist
[20:13] <dpb1> ntp merged?
[20:16] <rharper> dpb1: y
[20:16] <smoser> https://code.launchpad.net/~smoser/cloud-init/+git/cloud-init/+merge/343127
[20:17] <dpb1> rharper: woop!!!
[20:17] <dpb1> upload?
[20:17] <smoser> rharper: the files you got for templates
[20:17] <smoser> they were exactly from debian/ubuntu ?
[20:18] <smoser> with the "can?t"
[20:18] <rharper> yes
[20:18] <rharper> look at bionic's chrony.conf
[20:19] <rharper> lemme fire up sid
[20:20] <rharper> yeah
[20:27] <smoser> blackboxsw: around?
[20:27] <blackboxsw> smoser: yep
[20:28] <blackboxsw> hangout?
[20:28] <smoser> yeh
[20:28] <blackboxsw> release related
[20:29] <smoser> https://code.launchpad.net/~smoser/cloud-init/+git/cloud-init/+merge/343120
[20:31] <smoser> https://code.launchpad.net/~smoser/cloud-init/+git/cloud-init/+merge/343127
[21:02] <smoser> man. that exception_cb is a mess.
[21:02] <smoser> who wrote this stuff
[21:02] <smoser> readurl and wait_for_url are much different in what they pass
[21:32] <smoser> blackboxsw: i'll be back in in ~ 3 hours i guess.
[21:33] <blackboxsw> sounds good. just pushed robjo's changes  into my nettools branch, will have a devel branch up for you  before you get back. if you find the net-tools changes good I'll check in later to see if we want to land that
[22:09] <blackboxsw> pushed https://code.launchpad.net/~chad.smith/cloud-init/+git/cloud-init/+merge/343136 for bionic release to include ntp/chrony fixes
[23:58] <smoser> blackboxsw: merged yours and uploading