blackboxsw | meena: will have to get to it tomorrow. we're wrapping up SRU testing (and should be done tomorrow) | 00:24 |
---|---|---|
blackboxsw | softlayer SRU redone https://github.com/cloud-init/ubuntu-sru/pull/71 repushed. should be done tomorrow w/ puppet testng | 00:25 |
otubo | meena: I could take a look at your PRs. Not sure if 93, 53 or 42. Or all of them. | 12:00 |
meena | otubo: cool thanks | 13:43 |
meena | i tried to write a patch to @goneri's code, but somehow sleep seemed like a much more attractive option | 13:45 |
otubo | meena: all those PRs? Or just one of them? | 13:46 |
meena | Goneri: your code is missing at least one function, _supported_write_network_config | 13:47 |
Goneri | meena, Aloha! Thanks for the information. I plan to work on it next week-end. | 13:48 |
meena | otubo: let's start with #93. but generally, whatever you can independently verify would be great to ➕1 | 13:53 |
meena | Goneri: i might get to it earlier | 13:54 |
Goneri | meena, Yeah! You're a machine :-) | 13:54 |
meena | nah, i just have a very nice evening ritual off sitting in front of the TV with my partner, after the baby is in bed, and we both stare at our laptops | 13:56 |
meena | otubo: #42, by some miracle, is merged. but #69 is not, and isn't freebsd specific | 13:58 |
otubo | meena: awesome, I think I still have some time today. So I'll take a look :) | 13:58 |
Odd_Bloke | meena: I've been using Python since 1.5.x, I have all sorts of trivia stored in my head. :p | 14:47 |
Odd_Bloke | meena: I'm not sure, but I wouldn't be surprised if the mock object doesn't know its fully qualified name. It's just a (special) Python object with a name attribute, really. But maybe there's some trivia for me to learn here. :p | 14:50 |
smoser | someone needs to file an lxc issue for pstart fails. | 15:01 |
smoser | lxc-pstart really isn't doing anything hacky. basically it just | 15:02 |
smoser | - creates a profile with its own network | 15:02 |
smoser | - modifies container to be on that profile | 15:02 |
smoser | - inserts an 'init' and sets 'lxc.initcmd=<that>' | 15:02 |
smoser | - starts container and runs stuff | 15:03 |
smoser | - stops container | 15:03 |
smoser | - removes profile | 15:03 |
Odd_Bloke | meena: One minor change requested on #93. :) | 15:03 |
smoser | and lxc-proposed-nsapshot just does that, and then uses lxc publish. | 15:04 |
Odd_Bloke | smoser: Is there an easy way we can dump the lxd commands that lxc-pstart is performing? A bug report would probably have less friction if we don't need upstream to use our tooling to reproduce. | 15:06 |
smoser | it does have pretty good debug output | 15:07 |
smoser | Odd_Bloke: https://hastebin.com/levavikayi.cs | 15:09 |
smoser | oh | 15:10 |
smoser | there it is | 15:10 |
smoser | --dump-commands | 15:10 |
smoser | Odd_Bloke: https://hastebin.com/agolazazin | 15:14 |
meena | Odd_Bloke: will have to sit until 19:00, or else you propose a change, or, even push directly to my branch | 15:26 |
MrGeneral | Hello guys: | 15:28 |
MrGeneral | 2019-12-12 14:55:06,854 - __init__.py[DEBUG]: Searching for local data source in: ['DataSourceNoCloud', 'DataSourceConfigDrive', 'DataSourceOpenNebula', 'DataSourceDigitalOcean', 'DataSourceAzure', 'DataS> | 15:28 |
MrGeneral | 2019-12-12 14:55:06,855 - handlers.py[DEBUG]: start: init-local/search-NoCloud: searching for local data from DataSourceNoCloud | 15:28 |
MrGeneral | 2019-12-12 14:55:06,855 - __init__.py[DEBUG]: Seeing if we can get any data from <class 'cloudinit.sources.DataSourceNoCloud.DataSourceNoCloud'> | 15:28 |
MrGeneral | 2019-12-12 14:55:06,855 - __init__.py[DEBUG]: Update datasource metadata and network config due to events: New instance first boot | 15:28 |
MrGeneral | Can't get cloud-init in proxmox to be read from ide2 | 15:28 |
MrGeneral | Any hint? Sorry for the spam btw. | 15:28 |
Odd_Bloke | meena: No rush. :) | 15:52 |
Odd_Bloke | MrGeneral: What is "ide2"? | 15:53 |
MrGeneral | Odd_Bloke , Cloud-init drive in proxmox :\ | 16:00 |
MrGeneral | is there any way to specify the /dev/srID for cloud-init? | 16:00 |
Odd_Bloke | MrGeneral: Before we get into specific solutions, let's make sure we understand the problem. Can you open a bug at https://bugs.launchpad.net/cloud-init/+filebug and attach the tarball produced by `cloud-init collect-logs` to it, please? | 16:01 |
Odd_Bloke | That'll get us the info we need to properly help you out. :) | 16:02 |
MrGeneral | I'll try thanks Odd_Bloke | 16:02 |
MrGeneral | yep :) | 16:02 |
MrGeneral | trying to sort the issue tho | 16:02 |
MrGeneral | supposedly, the data source is a cdrom generated by proxmox mounted in the server | 16:02 |
MrGeneral | though as I install an ISO and load the cloud-init drive it sets by default from /dev/sr2 | 16:02 |
MrGeneral | actually | 16:02 |
MrGeneral | sr1 | 16:02 |
MrGeneral | instead of sr0 | 16:02 |
MrGeneral | Ouch was finally able to get it working @Odd_Bloke, just didn't setup the network! | 16:13 |
MrGeneral | let's find out why. | 16:13 |
MrGeneral | resize, hostname, root login enabled, etc. did work. | 16:14 |
MrGeneral | k, I know what to do I think. | 16:14 |
MrGeneral | @Odd_Bloke may have more info. So, I install the iso which already comes with cloud-int, then I create a template in proxmox. everything is confgiured other than networking, any idea? | 16:55 |
ahosmanMSFT | blackboxsw: any updates on the bug merge? | 17:01 |
Odd_Bloke | MrGeneral: Not really; gathering the info I requested earlier is really the next step. | 17:07 |
MrGeneral | Got it all working, Odd_Bloke, thanks, was my fault tho. :P | 17:07 |
MrGeneral | Windows with Cloud-init WILL be a pita tho! | 17:08 |
Odd_Bloke | Glad to hear it's working! :) | 17:08 |
MrGeneral | Thanks sir :) | 17:08 |
Odd_Bloke | blackboxsw: Am I remembering correctly that the Softlayer SRU PR is ready for re-review? | 17:09 |
blackboxsw | Odd_Bloke: yes it is https://github.com/cloud-init/ubuntu-sru/pull/71 I repushed late yesterday | 17:12 |
blackboxsw | increased a timeout to an unreasonable amount | 17:12 |
Odd_Bloke | Cool, will look in a minute. | 17:14 |
meena | Odd_Bloke: im too old/young to be in a rush | 17:15 |
blackboxsw | ahosmanMSFT: I left you comments on the PR | 17:19 |
blackboxsw | ahosmanMSFT: https://github.com/canonical/cloud-init/pull/84#discussion_r356369812 | 17:20 |
blackboxsw | last item and should be good | 17:20 |
ahosmanMSFT | blackboxsw Just fixed and pushed | 17:25 |
blackboxsw | thx | 17:30 |
rharper | blackboxsw: Odd_Bloke: https://github.com/cloud-init/ubuntu-sru/pull/75 | 19:03 |
blackboxsw | reviewing | 19:09 |
blackboxsw | rharper: https://github.com/cloud-init/ubuntu-sru/pull/66 | 19:10 |
rharper | reviewing | 19:10 |
meena | Odd_Bloke: pushed. (and rebased again) | 19:40 |
Odd_Bloke | meena: Merged, thanks! | 20:05 |
meena | Goneri: any chance you can tell me what function _supported_write_config is supposed to be? | 20:05 |
meena | Odd_Bloke: what about my other seventeen patches & bugs tho? | 20:07 |
Odd_Bloke | Oh wow look what's that over there behind you | 20:07 |
* Odd_Bloke runs | 20:07 | |
meena | Odd_Bloke: a santa hat. | 20:08 |
meena | incidentally, we're also watching Rare Exports rn. | 20:08 |
blackboxsw | rharper: btw that is a painful script to have to have written for intermittent Openstack deployment failures :/ | 20:11 |
rharper | =( | 20:12 |
rharper | took better part of a day; but it's written now =) | 20:12 |
meena | how do i ask cloud-id to give me debug (trace!) output from why it thinks it's on a None cloud? | 20:14 |
powersj | ha I have a merge to update our FAQ with that very command | 20:14 |
powersj | sudo DEBUG_LEVEL=2 DI_LOG=stderr /usr/lib/cloud-init/ds-identify --force | 20:15 |
powersj | we should clean that up fwiw | 20:15 |
powersj | and give cloud-id an alias to run that | 20:15 |
blackboxsw | nice powersj (your FAQ radar was up on that ;) ) | 20:15 |
meena | powersj: i've hacked up ds-identity to do the right thing… but cloud-id / cloud-init don't | 20:17 |
Odd_Bloke | meena: As background, cloud-id just reads data that cloud-init wrote when it was running. Josh has pointed you at the command that does the actual determination. | 20:17 |
Odd_Bloke | Hmm, interesting. | 20:17 |
Odd_Bloke | meena: Can you describe the problem in more detail? | 20:17 |
meena | actually, let'stry to run that command, and see wht happens, cuz i just looked at my /run/cloud-init/cloud.cfg and i modified it, so it might be wrong: https://gist.github.com/56a994dc4ccba1a81d7024c24bcd2aa4 | 20:19 |
meena | https://gist.github.com/27ac6b7ad02e1fa9de9145b1f4458f4d ← output | 20:24 |
meena | cloud-id is stil saying none | 20:25 |
rharper | cloud-init pulls from the instance-data.json in /run/cloud-init ; so unless you've re-run cloud-init after making ds-identify changes, you won't get a new instance-data.json file | 20:28 |
rharper | meena: so, 1) modify ds-identify 2) forcefully run it; 3) cloud-init clean --logs 4) cloud-init init --local; cloud-init init | 20:29 |
rharper | that'll be enough to have cloud-init go through the datasource and write out instance-data.json | 20:29 |
meena | rharper: ds-identify is giving the right response: datasource_list: [ Hetzner, None ] | 20:30 |
rharper | look at instance-data.json | 20:30 |
meena | i think the problem is still that i'm on Goneri\s branch and the network stuff is write-only… | 20:31 |
rharper | right, if you can;t reach the hetzner network IMDS, then you won't get the DataSourceHetzner to indicate that it was found; and no datasource was processed | 20:32 |
meena | now let's fix that before i continue to nonsensicle debugging… | 20:34 |
meena | rharper: https://twitter.com/emilyst/status/1204823144956977152 apropos of almost nothing. (except, perhaps, https://github.com/canonical/cloud-init/pull/53 ;) | 20:43 |
Odd_Bloke | blackboxsw: https://trello.com/c/6Pr4OxuI/48-sru-verification-10-azure-do-not-lock-user-on-instance-id-change doesn't have anyone's face on it but is marked complete, and I don't see a verification script in the repo. | 20:52 |
Odd_Bloke | Was that intentional? | 20:52 |
blackboxsw | Odd_Bloke: yes was intentional as we pinging the microsoft bug author on that one since we can't reproduce that failure mode easily | 20:53 |
blackboxsw | so I marked complete, but no owner for us Odd_Bloke | 20:54 |
Odd_Bloke | blackboxsw: But don't we need the SRU template stuff for our exception? | 20:54 |
blackboxsw | Odd_Bloke: nope. The SRU template stuff is for us to continue to go above and beyond our existing exception in cases where we think we might have a regression. Our exception is really just. does cloud-init CI and integration test suite continue to pass, and can manual upgrade validation on ## clouds validate we haven't regressed (plus maas and cdoqa) | 20:57 |
blackboxsw | Odd_Bloke: any other one off validation we do (with SRU templates) is just performed as we see fit (because we don't officially have to do that anymore due to our exception) | 20:58 |
blackboxsw | Odd_Bloke: rharper we were planning on dropping some of this one-off validation additional work when we've cleared a couple of SRUs without detecting potential regressions during SRU validation process. | 20:58 |
blackboxsw | earlier it had been fairly frequent where we have detected some potential regression during our one-off SRU validation. | 20:59 |
Odd_Bloke | "The test case should be developed as a part of each resolved bug or new feature." <-- this makes it sound like it's expected that we are validating bug fixes and features | 20:59 |
blackboxsw | Odd_Bloke: I had been those test cases as only those applicable to the given manual cloud test (as we require unit tests for any new feature/bug fix that is added in general) | 21:01 |
blackboxsw | and in the previous 6 or so SRUs, we don't attach any manual SRU one-off validation test output to the SRU bug | 21:01 |
blackboxsw | only the cloud-specific test | 21:02 |
blackboxsw | and maas/cdoqa | 21:02 |
Odd_Bloke | OK, fair enough. | 21:04 |
Odd_Bloke | I hadn't realised (or had forgotten :p) we were doing more than we needed, that may have modified my opinion on which issues needed additional verification. | 21:05 |
blackboxsw | we are just *that* amaaaazing | 21:05 |
blackboxsw | yeah, which is also why our reviews on this one-offs (non-cloud) don't really need to be too strict. But, good to show that the feature is exercised without regression to common install behavoir. | 21:06 |
blackboxsw | rharper: https://github.com/cloud-init/ubuntu-sru/pull/75/files#r357378780 one question otherwise +1 on openstack | 21:26 |
meena | def _supported_write_network_config() is part of distros/__init__.py | 21:28 |
meena | is there a way to set a breakpoint in the distro base class, so that when i run what rharper suggested, I'll know which methods i need to fix / reimplement / restore? | 21:31 |
Odd_Bloke | meena: Is what he suggested not addressing the issue? | 21:32 |
meena | Odd_Bloke: as mentioned previously, @goneri's branch is currently write only, in terms of networking, he was quite zealous in removing some quite rusty code… instead of trying to grease it | 21:34 |
meena | but that rusty code was still capable of answering questions like, are we connected? | 21:35 |
meena | is this interface up? | 21:35 |
* meena taps mic: is this thing on? | 21:36 | |
Odd_Bloke | meena: Sorry, is this related to the cloud-id issue, or a separate thing? | 21:36 |
meena | yesno | 21:36 |
Odd_Bloke | blackboxsw: rharper: https://bugs.launchpad.net/cloud-init/+bug/1849731 <-- do you have any suggestions on how to construct a test for this? | 21:37 |
ubot5 | Ubuntu bug 1849731 in cloud-init "Introduction of subplatform around 18.5 broke ConfigDrive" [Undecided,Fix committed] | 21:37 |
meena | it's python debugging: how do i find out when my parent class is being called, so i know which methods i need to implement | 21:37 |
Odd_Bloke | meena: I'm afraid I still don't really understand what you're trying to do. | 21:39 |
Odd_Bloke | But maybe trying to do One Big Branch which will address a bunch of issues is what is tripping (you|Goneri) up? | 21:40 |
meena | https://stackoverflow.com/questions/6954116/ruby-s-method-missing-in-python this. i basically need to log here | 21:40 |
Odd_Bloke | Yeah, that would do what you described. | 21:41 |
meena | thanks for rubbery duckling | 21:42 |
meena | even though you doubted me 😒 | 21:42 |
Odd_Bloke | meena: My concern with using the __getattr__ method is that it presupposes that you will be running the code in a way which will actually call all the methods that are missing. | 21:46 |
Odd_Bloke | If this is about working out which of the removed methods should be restored, I think you're better inspecting the codebase to work that out. | 21:47 |
meena | but there's so much code 🙀 | 21:48 |
meena | but, yeah, might be more feasible | 21:49 |
Odd_Bloke | meena: FreeBSD has a grep implementation though, right? ;) | 21:49 |
meena | i use ripgrep, and emacs' jump to definition | 21:50 |
meena | well, on my Ubuntu (elementary) laptop. on the servers its mostly vi, cuz i keep forgetting I've installed nvim | 21:53 |
Odd_Bloke | meena: https://github.com/canonical/cloud-init/pull/61/files#r357391834 <-- I broke down how I looked at one of these, as I figured my working might be useful reference | 21:57 |
Odd_Bloke | meena: I think my most important realisation was that all of our Distro subclasses are also called Distro, so the `Distro.get_devicelist` calls are referencing cloudinit.distros.freebsd.Distro, _not_ the super-class. | 21:58 |
Odd_Bloke | Which is _extremely_ confusing when reading that code. | 21:58 |
blackboxsw | Odd_Bloke: for testing config drive you could create an lxc with a /config-drive directory (via lxc-proposed-snapshot on bionic) | 22:02 |
meena | Odd_Bloke: i do have that much figured out already, but, yeah, it's incredibly confusing, and, unnecessarily so | 22:02 |
blackboxsw | Odd_Bloke: this may be helpful https://github.com/cloud-init/ubuntu-sru/blob/master/bugs/lp-1673637.txt | 22:04 |
blackboxsw | it's a test case seeding config-drive content on lxc | 22:05 |
blackboxsw | and the validation would be cloud-init query v1.subplatform | 22:06 |
powersj | blackboxsw, so are we all set on the SRU minus some internal testing? | 22:08 |
blackboxsw | powersj: yep, I'm adding a nit for puppet validation of csr. I think Odd_Bloke is grabbing the configdrive | 22:08 |
blackboxsw | I'll attach the cloud logs now | 22:08 |
blackboxsw | so SRU will be ready minus internal testing | 22:08 |
powersj | ok so monday, darn | 22:09 |
blackboxsw | just attached all SRU logs to the bug https://bugs.launchpad.net/ubuntu/+source/cloud-init/+bug/1854872 | 23:06 |
ubot5 | Ubuntu bug 1854872 in cloud-init (Ubuntu Eoan) "sru cloud-init (19.2.36 to 19.3.41) Xenial, Bionic, Disco, Eoan" [Undecided,Fix committed] | 23:06 |
blackboxsw | ahosmanMSFT: and I just published tip of cloud-init to Ubuntu Focal 20.04: cloud-init v. 19.3-74-g129b1c4e-0ubuntu1 | 23:09 |
blackboxsw | coincidentally this will also drop a package dependency on ifupdown2 from focal as well | 23:10 |
tribaal | awww rharper spent some time on my unfinished branch... sorry man! | 23:34 |
tribaal | I should not have opened it (even though it's a "draft" PR | 23:35 |
Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!