[13:35] what's the minimal cloud-init config to just specify the ssh public key for the default ubuntu user? [13:36] system_info -> default_user -> ssh-import-id: foo? [13:42] ok, found a way, just needed to specify user ubuntu again and override what I wanted [14:04] ahasenack: you can just do: [14:06] ssh_authorized_keys: [14:06] - ssh-rsa AAAA...H5qV7NZ mykey@host [14:07] smoser: under "users->name: ubuntu", right? [14:07] well you can put it there. [14:07] but you dont need to. [14:07] top level applies to default user [14:07] ah, good to know [14:07] does that work for "shell: /bin/bash" too? [14:07] or that one has to be under a user [14:08] no. that has to be in user [14:08] ok [14:08] * smoser didn't know that ahasenack was a elite shell user. [14:08] zsh? or maybe he just loves the old school ksh [14:08] what's yours, zsh? [14:08] ksh! [14:08] i jsut use bash. [14:09] wow. ksh. [14:09] I use bash too, that's what I'm setting it to [14:09] i dont think i've ever seen someone not previously employed by ibm that used ksh. === shardy is now known as shardy_afk === shardy_afk is now known as shardy [18:56] rharper: on my desktop i do not have /run/log/journal/ [18:56] what creates /run/log/journal/*/system.journal ? [19:12] journald [19:12] did you persist your systemd journal ? [19:13] smoser: you probably did, so it's /var/log/journal [19:13] I did that on diglett which does not have a /run/log/journal, but /var/log/journal [19:14] rharper: ok. so it will either go to /run/log/journal// or /var/log/journal? [19:14] yeah [19:14] /var/log/journal/ [19:15] if you mkdir /var/log/journal and run a command that tells journald to use a persistent directory instead of /run [20:01] @smoser, are we good to land my MP? [20:06] dojordan: yes. and i'd like to get that uplaoded to bionic sooner than later. [20:23] awesome! I assume you do the red button pressing? [20:25] dojordan: yeah. [20:25] i'll get uploaded today i hope. [20:25] meged and uploaded. [20:26] great, thanks! [21:12] blackboxsw: https://code.launchpad.net/~penick/cloud-init/+git/cloud-init/+merge/335286 [21:12] i can't think of a reason to not pull that [21:12] reviewing [21:13] oops missed that one [21:13] I recall reading it, just never committing to the review [21:13] ohh right no CLA, yeah I didn't have a concern with it. will officially accept :) [21:15] approved [21:15] want me to land it? [21:15] blackboxsw: sure [21:15] ^ == more testing of review-mps script [21:15] powersj: can i easily run a jenkins ec2 on https://code.launchpad.net/~smoser/cloud-init/+git/cloud-init/+merge/336366 [21:16] works in my local gtest [21:16] I only have jenkins jobs for ec2 on master [21:17] smoser: we have 4 approved branches now in cloud-init review queue. Can I land all of them? [21:17] yeah. [21:17] powersj: you think its enough for me to just run that branch ec2 once on a few tsets ? [21:17] and then merge ? [21:17] blackboxsw: please do land all those. yeah. [21:18] smoser: that is fine by me [21:30] blackboxsw: qa-scripts/launch-ec2 [21:30] quick review? [21:30] http://paste.ubuntu.com/26446608/ [21:30] creates help like: http://paste.ubuntu.com/26446617/ [21:31] +1 smoser yeah instead of invalid help text/ thx for the additional logging info on zone (was lacking that on my last deploy/test attempt) [21:32] land away. and land anything you like in there as you see it. I'll do the same. if we break something we can fix it up as we use it [21:33] since there are ~95000 different types on ec2, i didn't put them all there. [21:34] and didn't use choices (as is done in the softlayer one) [21:53] smoser: did the following branch close all four of those related gce bugs https://code.launchpad.net/~illfelder/cloud-init/+git/cloud-init/+merge/334777? [21:53] I'm trying to verify now [21:53] and oops commit message didn't contain the bug references [21:55] blackboxsw: that commit message does need work. [21:55] updating [21:55] :-( [21:55] i can update now [21:55] grr yeah [21:55] need a force push [21:56] saw that too late [21:56] just reset --hard HEAD [21:56] and force push now [21:56] HEAD^ [21:57] force pushed it out [21:58] will repush once commit message is fixed. [22:00] blackboxsw: please read. [22:08] lgtm [22:18] blackboxsw: i'll come back in in 3 or 4 hours [22:18] and will upload to ubuntu whatever you have in trunk at that point [22:18] feel free to grab https://code.launchpad.net/~smoser/cloud-init/+git/cloud-init/+merge/336497 [22:19] i'll just upload whateve ryou have in trunk [22:19] https://code.launchpad.net/~smoser/cloud-init/+git/cloud-init/+merge/336366 [22:19] that one too [22:19] and /me out [22:19] later [22:26] smoser: will do, I'm fixing dojordan's commit message to reflect retries etc instead of infinite loops etc. [22:28] (thanks) [22:48] dojordan: does the updated commit message look ok https://code.launchpad.net/~dojordan/cloud-init/+git/cloud-init/+merge/334341 ? [22:48] just let me know if you have any changes (or not) and I'll merge away [22:52] dojordan: saw a join/logout drop IRC message [22:52] dojordan: does the updated commit message look ok https://code.launchpad.net/~dojordan/cloud-init/+git/cloud-init/+merge/334341 ? [22:52] 3:49 PM just let me know if you have any changes (or not) and I'll merge away [22:52] New messages since you tabbed out [22:57] was actually about to reply [22:57] technically it is an infinite polling loop assuming we are getting 404s from IMDS, but we will only retry DHCP a max of 5 times [22:58] technically a VM can be in this polling state for a long time (day or two) [22:58] also @blackboxsw, what time zone are you in? [23:00] dojordan_: Mountain, still have an hour before I go poof. [23:02] ah, that explains how 3:49 PM is in the future [23:07] thx for the correction, I had forgotten about the exception_cb behavior [23:07] ok will revert that part but wanted to keep the reference to EphermeralDHCPv4 so we can look back easily on commit history when pushing this functionality to other clouds [23:08] s/clouds/datasources [23:09] dojordan_: updated one last time. https://code.launchpad.net/~dojordan/cloud-init/+git/cloud-init/+merge/334341 [23:10] lgtm [23:12] ok waiting on one last azure instance boot on my side..... will merge in 10 mins [23:26] dojordan_: on bionic with your changeset. my azure machine doesn't come back up on reboot. I can run "sudo cloud-init clean --logs --reboot" with your deb installed. This reproduces a fresh-install test. It seems my instance doesn't come back up..... I'm trying to see what gives now as I thought all IMDS polling was based on present of a /var/lib/cloud/data/poll_imds file. [23:26] hrm, I know this was working on earlier commits within your branch. [23:27] it shouldn't even create the file in the first place since the ovf_env.xml on the ISO doesn't contain that flag [23:27] and from my understanding you wouldn't even be providing the flag to turn that file on anyway.... [23:27] ^ right [23:27] exactly [23:29] did you create the marker file? [23:29] before rebooting? [23:29] also do you have serial logs? [23:31] I hadn't in this case, it was a fresh install. but grabbing logs now [23:37] hrm I hadn't enabled boot diagnostics on this instance which I think captures serial log right? trying another deployment (without this branch) to make sure my "cloud-init clean" on tip of master doesn't actually break Azure anyway [23:38] it actually might [23:39] because once we remove the ISO (after we report ready), how will the VM get the ovf_env [23:39] also, if you can get me the deployment ID and subscription id I can grab the logs for you [23:40] ok just validated the released version of cloud-init in bionic can cloud-init clean --logs --reboot on azure and come back freshly installed. [23:41] hrm [23:41] and tip of master works too 17.2-18-gd14fa1a2-1~bddeb [23:42] so 17.2-13-g6299e8d0-0ubuntu1 (published version) is good too. [23:42] will grab deployment id now [23:42] before I upgrade to your branch [23:42] "id": "/subscriptions/12aad61c-6de4-4e53-a6c6-5aff52a83777/resourceGroups/srugrp10/providers/Microsoft.Compute/virtualMachines/my-b3", [23:43] great [23:43] I'll start diggiung [23:43] I'll also add your ssh account to my instance if you like [23:43] i thought it didn't come back up [23:43] ssh ubuntu@40.70.56.77 [23:44] dojordan_: I deployed a fresh vm [23:44] ah [23:44] to test tip and published versions too [23:44] so I'm about to install a deb made from your branch on it [23:48] fell free to connect to my shared term with 'byobu' dojordan [23:48] feel free* [23:50] ok new cloud-init installed from your branch, no marker file present [23:50] last cloud-init run worked fine [23:52] @blackboxsw, not sure if i can type in your shared term but sounds good [23:53] ok so we're kicked for a fresh install now [23:53] s/install/cloud-init run/ [23:53] what should happen is subsequent ssh connections to the same IP will present us with an ssh key error (key has changed on new cloud-init run) [23:54] this happens within about 30 seconds on my last two attempts (with tip and published version of cloud-init in bionic) [23:55] i assume my ssh key wont be there, so my connection refused makes sense [23:55] yours? [23:55] yours? [23:55] correct. your's shouldn't exist again.... but I'm getting no response yet [23:55] from ubuntu@40.70.56.77 [23:55] though azure portal says it's up [23:55] * blackboxsw checks the cli [23:56] ssh: connect to host 40.70.56.77 port 22: Connection timed out [23:57] https://paste.ubuntu.com/26447610/ (logs from bionic3) [23:58] meh: az vm boot-diagnostics get-boot-log --ids /subscriptions/12aad61c-6de4-4e53-a6c6-5aff52a83777/resourceGroups/SRUGRP10/providers/Microsoft.Compute/virtualMachines/my-b3 [23:58] Please enable boot diagnostics.