| === apollo13_ is now known as apollo13 | |
| [12:28] <otubo> Odd_Bloke, quick question on PR 428: Would you like a unittest for kernel_version() function? We already have tests for swap file. | |
| [12:31] <otubo> Odd_Bloke, oh I see you specified you want both :-) Nevermind. | |
| [12:35] <Odd_Bloke> :) | |
| [12:40] <otubo> Odd_Bloke, but I'm not sure I can test the kernel_version function without issuing os.uname() again. I would have to hard-code a kernel version assuming I know the actual kernel version in where the test will run. | |
| [12:43] <Odd_Bloke> otubo: I noticed this when reviewing your generator PR: https://github.com/canonical/cloud-init/pull/452 | |
| [12:45] <otubo> Odd_Bloke, Oh, didn't see you already had this PR! Apologies :-) | |
| [12:46] <Odd_Bloke> otubo: Oh, no, I noticed that one of the other systemd files was incorrect when I was grepping around to confirm that your change matched the rest of the templates. Your fix already landed, this is a different file. :) | |
| [12:47] <Odd_Bloke> otubo: Regarding testing kernel_version(), I was thinking that you'd mock out os.uname() and test that if you return a few different valid kernel versions (ideally formatted how we would find across a few different distros) it produces the expected tuple. | |
| [12:48] <otubo> Odd_Bloke, oh nice one, indeed. | |
| [12:54] <Odd_Bloke> Thanks for the review! | |
| [12:59] <Odd_Bloke> Oh, huh, I thought I had a committer +1 on https://github.com/canonical/cloud-init/pull/391 but I don't; could someone give me a +1 so I can hit the green button? | |
| [13:10] <meena> Odd_Bloke: i could but its not worth much | |
| [13:17] <Odd_Bloke> :) | |
| [13:25] <Odd_Bloke> Similarly, just looking for a committer +1 here: https://github.com/canonical/cloud-init/pull/452 | |
| [13:27] <otubo> Odd_Bloke, I don't think I'm a committer, am I? | |
| [13:29] <Odd_Bloke> Nope, that's Chad, Ryan, Scott and myself. | |
| [13:37] <powersj> ahem | |
| [13:37] <powersj> Odd_Bloke, +1 on 452 | |
| [13:40] <Odd_Bloke> powersj: Thanks! | |
| [13:43] <powersj> +1 on 391 | |
| [13:44] <Odd_Bloke> Thanks! | |
| [14:51] <meena> otubo: Odd_Bloke tried to give me reviewer permissions before, and it didn't pan out | |
| [14:52] <meena> my ✅ is still grey | |
| [14:52] <Odd_Bloke> meena: I gave you enough permissions that we can _request_ reviews from you, we still want to have a committer's +1 even on BSD-specific changes. | |
| [15:07] <meena> aaaaaaahh | |
| [15:59] <blackboxsw> sorry lucasmoura just pushed that missing release -> UBUNTU_RELEASE commit to the sru test refactor https://github.com/cloud-init/ubuntu-sru/pull/128 and merged your RandomPassword validatrion | |
| [15:59] <blackboxsw> validation rather | |
| [17:12] <falcojr> whats the point of these "ssh "${SSHOPTS[@]}" ubuntu@$VM_IP -- grep Trace /var/log/cloud-init.log" lines in the manual verification? | |
| [17:13] <falcojr> there's no trace in the logs so the script is failing | |
| [17:13] <blackboxsw> falcojr: generally with set -ex in the script we should have made that line whats the point of these "ssh "${SSHOPTS[@]}" ubuntu@$VM_IP -- ! grep Trace /var/log/cloud-init.log" | |
| [17:13] <blackboxsw> sorry if I switched the script to set -ex, I should have adapted those lines to expect no matches | |
| [17:13] <blackboxsw> or fail quickly if they hit it | |
| [17:14] <blackboxsw> note the bang "!" above | |
| [17:14] <falcojr> gotcha. I saw some with a bang and some without, so was wondering | |
| [17:14] * blackboxsw looks over the refactor cloud tests and fixes that throughout | |
| [17:15] <blackboxsw> gce and azure scripts it seems | |
| [17:15] <rharper> blackboxsw: btw, this swap: auto for swapfile is a PITA, all of my lxd deployments use ZFS and ZFS does not support swapfiles ... so having to go VM route; lxd vms IIRC don't have cloud-images for Xenial (right? ) so that'll have to be a one-off to get a xenial cloud-image up; should ahve a PR up for that today | |
| [17:16] <blackboxsw> gce/softlayer I mean , though I ffixed softlayer in the PR https://github.com/cloud-init/ubuntu-sru/pull/130 | |
| [17:16] <Odd_Bloke> I think we've figured out that you can launch xenial VMs but the lxd agent won't work without the HWE kernel so you can't use any of the `lxc` commands to interact with it. | |
| [17:17] <Odd_Bloke> Someone else please confirm though, we've been learning a lot over the past couple of days so I might have got mixed up. :p | |
| [17:17] <blackboxsw> rharper: sorry, right so falcojr was using ssh into the xenial lxc container instead of lxc exec specifically for xenial https://github.com/cloud-init/ubuntu-sru/pull/129 | |
| [17:19] <blackboxsw> strange thing rharper is that paride was still able to lxc launch ubuntu-daily:xenial --vm --profile <vm_profile> yesterday I thought (and I know I was able to do the same last week, but can no longer do that today) | |
| [17:19] <rharper> well, I think I'm going to add an ext4 loop device as a storage pool to lxd | |
| [17:19] <blackboxsw> rharper: trusty via lxc launch --vm is an official "not gonna do it" as far as images. Xenial has some support, and bionic+ is good to go | |
| [17:19] <rharper> then file-based swaps will work and I can just use the normal lxc path | |
| [17:19] <falcojr> @rharper I tested some other swap srus | |
| [17:19] <falcojr> I can take this one too if that's easier | |
| [17:20] <rharper> I've got the test written; just a matter of running it on all of the releases | |
| [17:20] <falcojr> kk | |
| [17:25] <blackboxsw> falcojr: pushed the "! grep Trace" commit into sru/refactor-manual-clouds | |
| [17:31] <blackboxsw> o/ sparkiegeek | |
| [17:31] <sparkiegeek> \o blackboxsw | |
| [17:34] <sparkiegeek> I'm trying to understand how MAAS needs to send user-data to cloud-init for the machines it boots. AFAICT MAAS is currently base64 encoding a script, and sending it as application/octet-stream but I might be misunderstanding | |
| [17:40] * blackboxsw is peeking through maas code again to see how cloud-config userdata is specified directly to cloud-init. Generally I believe providing any cloud-init userdata types should be acceptable and interpreted correctly by cloud-init https://cloudinit.readthedocs.io/en/latest/topics/format.html | |
| [17:42] <blackboxsw> In maas case, I know that cloud-init gets some user-data/ or at least vendors data through a curtin config passthrough that maas provides (in order to make sure curtin and cloud-init knows how to talk back to the maas proper log/event endpoints). But I don't recall exactly how maas passes cloud-init user-data specifically (whether through curtin config or the maas datasource). | |
| [17:43] <blackboxsw> ahh user-data for maas comes from the maas seed url. to the instance. | |
| [17:47] <blackboxsw> so generally, if that user-data exposed to the instance is a supported cloud-init user-data format (link above). Cloud-init should be able to properly interpret it. I see in the MAAS datasource, that cloud-init will attempt to decode the content of user-data url if present. It does expect user-data to be binary encoded content. | |
| [17:47] <blackboxsw> https://github.com/canonical/cloud-init/blob/master/cloudinit/sources/DataSourceMAAS.py#L22-L27 | |
| [17:48] <blackboxsw> sparkiegeek: not sure if that is what you were looking for as far as context | |
| [17:50] <sparkiegeek> blackboxsw: thanks, that's plenty to chew over | |
| [17:58] <blackboxsw> -/me merged ubuntSRU verification branch https://github.com/cloud-init/ubuntu-sru/pull/128 will adapt/iterate on the softlayer results. | |
| [17:58] <blackboxsw> falcojr: I think I handled your initial review comments (and probably expect there may be a couple of extra things you run across in your GCE test run) | |
| [18:06] <rharper> sparkiegeek: you may also be interested in the curtin handling of maas provided cloud-config, https://git.launchpad.net/curtin/tree/curtin/commands/curthooks.py#n1344 | |
| [19:29] <rharper> fun fact: eoan, bionic, and xenial never had the bug around swap: {size: auto}; as it was an regression introduced with the fix for fallocate, on Jan 23; 6603706eec1c39d9d591c8ffa0ef7171b74d84d6 ; it only regress focal; blackboxsw so, for SRU verification is it OK to only verify focal ? | |
| [19:36] <blackboxsw> good pt. rharper yes focal only | |
| [19:37] <blackboxsw> given we never introduced the bug in the first place.... thank goooodness for not SRUing for 6 months ;) | |
| [19:37] <blackboxsw> ohh wait carry that cost forward :/ | |
| [19:38] <rharper> hehe | |
| [19:39] <rharper> I spent too long trying to figure out why I couldn't recreate on eoan, thinking it was lxd vm related; turns out not a bug in 19.4-33 | |
| [20:18] <rharper> blackboxsw: https://github.com/cloud-init/ubuntu-sru/pull/132/files | |
| [20:19] <blackboxsw> on it rharper | |
| [20:19] <rharper> you also mentioned, https://git.launchpad.net/cloud-init/commit/?id=723e2bc1 (nfs style paths) ; is that still open/needed ? | |
| [20:19] <blackboxsw> rharper: yep in the hidden trello board, still has your gravatar on it :) | |
| [20:20] <blackboxsw> falcojr: you didn't grab the above commit for verification did you? ^ | |
| [20:21] <falcojr> blackboxsw rharper , no I have not grabbed that one | |
| [20:31] <rharper> blackboxsw: falcojr: ok, I'll take that one next | |
| [20:37] <blackboxsw> lucasmoura: a followup change request on https://github.com/cloud-init/ubuntu-sru/pull/122#pullrequestreview-436136291 | |
| [20:38] <blackboxsw> falcojr: you are on gce next for SRU verification right? and you don't plan on oracle? | |
| [20:38] <blackboxsw> if not oracle , then I'll grab it (and re-find my credentials) | |
| [20:39] <lucasmoura> blackboxsw, ack | |
| [20:39] <falcojr> blackboxsw yes, I'm doing GCE (should have it done soon)...I could jump or Oracle too | |
| [20:40] <blackboxsw> falcojr: I'll wrap up SRU reviews and ping you to when I'm done to see where we are both at. | |
| [20:41] <falcojr> sounds good | |
| [20:52] <meena> Odd_Bloke: how do i take on one of those cloudinit/net refactor bugs? | |
| [21:04] <meena> never mind, i'm too tired. lol | |
| [21:35] <blackboxsw> heh meena, just assign yourself when intererested | |
| [21:38] <blackboxsw> meena: Odd_Bloke was planning on filing bugs for each aspect of the net refactor and we can have devs assign themselves to each bug and mark it in progress | |
| [21:38] <blackboxsw> as they carve out the items | |
| [22:26] <blackboxsw> I'm on https://github.com/cloud-init/ubuntu-sru/pull/133 falcojr for gce verification. minor request for comment on https://github.com/cloud-init/ubuntu-sru/pull/129 to resolve and we can land that | |
| [22:34] <falcojr> Sounds good | |
| [22:40] <blackboxsw> lucasmoura: https://github.com/cloud-init/ubuntu-sru/pull/131#discussion_r444497107 I'm not sure what you are asking. should I add the content of sru-vars.template to the verification output for softlayer ? | |
| [22:40] <blackboxsw> or something else | |
| [22:44] <lucasmoura> blackboxsw, no no. It is just that in the other cloud providers tests we have the shellscript that generated the output on the report as well, like this for example https://github.com/cloud-init/ubuntu-sru/blob/master/manual/ec2-sru-20.2.45.txt#L1 | |
| [22:45] <blackboxsw> ahh ahh, I see. yes I will put that in there. an oversight thanks | |
| [22:45] <blackboxsw> I thought you were specifically asking about including sru-vars.template content alone | |
| [22:45] <blackboxsw> didn't realize I left it out | |
| [22:45] <blackboxsw> all of it | |