=== dionysus70 is now known as dionysus69 [09:47] Odd_Bloke: login for posting is okay, but login for retrieval of the raw text?? [09:47] hello, i am "alive" again === cpaelzer_ is now known as cpaelzer [11:54] docs patch, as requested by Odd_Bloke: https://github.com/canonical/cloud-init/pull/658 [11:55] falcojr: ready for review: https://github.com/canonical/cloud-init/pull/655 [11:55] smoser: ready for re-review: https://github.com/canonical/cloud-init/pull/621 [11:56] i haven't checked, but this might break https://github.com/canonical/cloud-init/pull/653 again 🙈 === dionysus70 is now known as dionysus69 [13:54] meena: more. thanks. [13:54] comments. [13:55] Odd_Bloke: any reason why https://github.com/canonical/cloud-init/pull/653 didn't get merged ? [14:11] smoser, btw, i have no idea how dmidecode in ds-identify works, given that it's missing translation for ⅔ of the keys [14:17] meena: well, we only ever read chassis_asset_tag sys_vendor product_name product_uuid product_serial product_serial [14:17] yeah, and all those are covered in the dmi_decode func [14:17] right ? [14:19] oh, OK [14:19] i didn't realise i was doing too much work [14:22] smoser: re: #653, it doesn't have an Approve on it. [14:40] Odd_Bloke: yes it does [14:40] i can hit the "quash and merge" button [14:41] oh, is a ACK not required on ubuntu branches ? [14:42] so... i can ack it for you. but i woudl like to suggest that running "new-upstream-snapshot" is preferable. [15:11] 10/28 18:57:23 Hawk_X> BTW I didn't have time to spin the instance, hopefully I will try it tomorrow and will share the log output over screenshot of the console. Have a great night/day to you all! [15:11] 11/6 08:59:35 Hawk_X> Good morning/afternoon/evening to you all! [15:11] 11/6 09:00:35 Hawk_X> Gentlemen/Ladies. I'd like your wisdom on another question. Is it possible to libvirt to make a "Trunk" bridge of many other bridges. [15:11] 11/6 09:02:10 Hawk_X> For example: I have bridge A, B, C and D. And I want to create bridge X to receive traffic of bridges A, B, C, D. The use case is that I want to have access to these bridges over a single one to ease the connectivity. [15:11] 11/6 09:03:40 Hawk_X> Two scenarios: Bridges A, B, C, and D are host-only. [15:11] 11/6 09:04:39 Hawk_X> Two scenarios: Bridges A, B, C, and D each one are tagged VLANs to the LAN. [15:11] 11/6 09:06:21 Hawk_X> That is, there is a physical port that receives a 802.1q trunk with all the VLANs tagged. [15:11] 11/6 09:06:58 Hawk_X> Mmm no that I writing it I believe I have to clear out my question. Let me go back to the drawing board.... [15:11] 11/6 09:10:13 Hawk_X> but the host-only bridges question I believe is valid, I want to have access to all of them. Can I tag them (A, B, C and D) so I can receive them as a "Trunk" on bridge X? [15:11] 11/6 09:10:31 Hawk_X> Thanks to all for your assistance! :-) [16:06] rharper, I have updated a UT for configDrive in the https://github.com/canonical/cloud-init/pull/647 let me know if any comments on this. will be working on OVF and openstack update soon [16:10] smoser: Yeah, we only have our branch protection on master, because we don't want to squash-merge the ubuntu/* branches (because we want the merge commit and the separated commits for the changelog updates etc.); whoever uploads should push the tree they uploaded from directly to those branches. [16:40] smoser: It's not clear to me that `new-upstream-snapshot --update-patches-only` is really what we want for the daily branch: it's intended for minimal fixes to unbreak daily builds, and we don't need a whole new upstream snapshot to do that. IMO, it's easier to understand what actually broke the daily build (and therefore what will break when we do a `new-upstream-snapshot` in the ubuntu/* branch) if we [16:40] only have the specific fixes in that branch. [16:55] it is exactly what --update-patches-only was intended for. [16:55] update-patches-only only updates the patches. its exactly what you did. [16:56] just with some automation in place and consistency. [17:02] i didnt realize there was a "daily" branch. i guess i'd forgotten that. but i dont really think it changes anything. [17:07] But we don't need or want a new upstream snapshot in the daily branch: that changes it substantially in my view. [17:10] (We only ever use the ubuntu//daily branch to merge into "ubuntu/ merged into master" in the daily recipes.) [17:12] (I'm not suggesting we should ever not use the tooling for the actual packaging branch, to be clear.) [17:13] https://paste.ubuntu.com/p/7cn363pHYs/ [17:13] thats the diff it produces. lines 79 down. [17:14] had you ran it more recently, it would have only modified the one file that was changed. so you'd only see the changelog and lines 115 to 140. [17:14] basically the same as your PR in https://github.com/canonical/cloud-init/pull/653/files [17:15] the other changes/churn that youre' seeing is actual churn. it updates the quilt files. [17:15] OK, maybe I ran it incorrectly locally, because I saw it pulling in the whole of master. [17:15] That churn is still noise. [17:16] on a daily branch ? that is there for machiens ? [17:16] and .... maybe i didn't mention that i fixed https://github.com/canonical/uss-tableflip/pull/65 [17:17] you can make it not update the debian/changelog if uyou want, that'd be easy enough. [17:17] Oh, I didn't set SKIP_BRANCH_NAME_CHECK=1 in my environment, so it performed a full merge? [17:18] it did do the merge before https://github.com/canonical/uss-tableflip/pull/65 [17:18] it left it merged. [17:18] but SKIP_BRANCH_NAME_CHECK is just skipping that my detached head was not named "ubuntu/" [17:18] because... i dont care. [17:19] shoot. [17:19] bah. antoher pr coming [17:23] https://github.com/canonical/uss-tableflip/pull/66 [17:29] My bad. Wrong IRC channel :-( Have a good one! [17:48] smoser / Odd_Bloke : applied your review suggestion [17:52] meena: i think you missed one. or i wasn't being clear. [17:52] smoser: if you mean eblah and sblah then, i didn't miss that, i just find that name less clear [17:52] https://github.com/canonical/cloud-init/pull/621/files#diff-153c7aabaff1f846f28ae50b2c562549689a68cf0f94128dd1da31bb450d8f83R93 [17:53] - except (IOError, OSError) as e: [17:53] + except subp.ProcessExecutionError as e: [17:54] and then [17:54] https://github.com/canonical/cloud-init/pull/621/files#diff-15e919a60d374d575c257871c4cd44989c4b923c248f7bb75c352d0e8d3a759eR1965 [17:54] i'd move _is_contaienr_freebsd above _upstart too (upstart is also 2014 era linux) [17:54] i thought chromeos uses it :P [17:55] oh. i didnt even realize that was still the case. [17:55] and wrt short names... that is golang speaking. [17:56] in a 30 line function like that, with 3 variables, its just not that important that they be descriptive. [17:57] it's not about being descriptive, it's about being able to tell them apart [17:57] ename and sname have the same shape and size [17:58] my dyslexia is going to be tripped 3 times by that [17:58] then just drop '_field' [17:59] sys_field and kenv_field -> sys (or maybe lsys or something) , kenv. [17:59] but who cares. [17:59] the next person who refactors that can care :P [17:59] so line 93 there in dmi.py is the only thing i really care about. [18:00] oooh! i thought i had fixed that long ago! [18:01] fixed [18:05] (btw, i have run the ds-identify code on my local machine and in a vm on Hetzner) [18:05] (ds-identify happily recognizing stuff…!) [18:05] meena: push ? [18:05] i hit reload i still see [18:07] smoser: https://github.com/canonical/cloud-init/pull/621/commits/cc5fb806ce69af6dcc6b8d137107820fce093df6 [18:07] i did push [18:09] what the heck is going on [18:09] i load https://github.com/canonical/cloud-init/pull/621/files [18:09] and it shows me line 93 has 'except (IOError)... i see you did push that. [18:09] even in a new browser [18:09] what am i doing wrong. [18:10] oh. you didnt change it there. you cahnged it somewhere else. [18:10] oh, i'm on line 110 [18:10] (which i'm fine with too) [18:10] but you should mention that one in your commit message. [18:10] i changed it where i copied it from lolol [18:10] adn then just read you rcommit message and make sure it is up to date. [18:10] push -f [18:10] and if you would... squash and rebase. [18:11] i will at the end when we're done with the review [18:11] (i really dont like having to paste my own commit message from yours in the description... i'd much rather you do it) [18:11] fix that OSerror on line 93 and i'm happy with that [18:12] i just did [18:13] 19:11 <@smoser> (i really dont like having to paste my own commit message from yours in the description... i'd much rather you do it) ← haa haa. [18:13] yeah, that does seem tedious [18:13] well, the issue last time was that you had a "signed off by" and other things in your commit messages. [18:14] but did *not* have those in the description [18:14] aah [18:14] i dont mind copy and paste if thats all i have to do. [18:22] smoser: squashed [18:23] smoser: btw, it's not signed-off-by, it's Co-Authored-by [18:23] in this case , co-authored with you [18:25] yeah. so maybe it was someone else... its just a pain. [18:25] in launchpad we had a lander [18:25] anad the lander linted the commit message. [18:25] and did the right thing generally. [18:28] github is so underdeveloped as development tool for bigger teams [18:45] meena: patches are welcome https://resynth1943.net/articles/github-source-code-leak/ [18:45] :) [18:53] hello guys, can someone give me a direction on phone_home? [18:53] thanks [19:12] direction? [19:56] itlinux: what kind of direction? [19:56] how do I get the token for example to get auth [19:56] thanks meena: [19:58] here is an example I was looking but not sure how to get that data [19:58] phone_home_cli: "curl -i -X POST -H 'X-Auth-Token: gAAAAABc5UscwS1py5XfC3yPcyN8KcgD7hYtEZ2-xHw95o4YIh0j5IDjAu9qId3JgMOp9hnHwP42mYA7oPPP0yl-OQXvCaCS3OezKlO7MsS-ZCTJzuS3sSysIMHTA78fGsXbMgCQZCi5G-evLG9xUNrYp5d3blhMnpHR0dlHPz6VMacNkPhyrQI' -H 'Content-Type: application/json' -H 'Accept: application/json' http://192.168.0.121:8004/v1/d3779c949b57403bb7f703016e91a425/stacks/demo_waf/3dd6ce45-bb8c-400d-a48c-87ac9e46e60e/resou [19:58] rces/wait_handle/signal" [20:04] here is another example phone_home_url: "https://webhook.site/5f8cd8a7-b051-4648-9296-8f6afad34c93" [20:04] thanks guys [20:13] meena: Heads-up, we both missed one other `net` in your PR so I fixed that, (rewrapped some lines), rebased onto master (to save an "Update branch") and force-pushed to your branch. [20:14] If that breaks you because I broke your git history, please feel free to berate me (and then I can help fix it :p). [20:23] Odd_Bloke: eh, i don't mind broken git history [20:23] itlinux: that second example is probably the instance-id, which often uses a UUID [20:25] meena: how do I get that? is it self gen how would I call it in the script. thanks [20:26] itlinux: for instance, here's what i have have on my Hetzner server: smbios.system.uuid="c9832204-776c-4697-8349-8772138bd3f3" [20:27] here's from my local server: smbios.system.uuid="0ee069ca-c08d-47f7-bd66-351cd5de8616" [20:28] so, hetzner now uses that as instance-id: https://github.com/canonical/cloud-init/pull/630 [20:29] but maybe we have the variable somewhere accessible thru jinja or user-data? [20:29] Does anyone know? [20:31] itlinux: Those are two examples of how you could use phone_home, not examples of how you must use it. If you can explain a little more about what you're trying to achieve with `phone_home`, we'll be able to be more helpful. [20:34] I want to make sure it completes the process if not rerun the script [20:34] @Odd_bloke: [20:36] itlinux: `phone_home` can report failure/success, but it doesn't give you a mechanism to retry anything. (If by "rerun the script" you mean "rerun the script which will launch fresh instances", then that does still make sense.) [20:36] yes correct [20:37] I shorted the version but you got what I wanted to achieve [20:37] itlinux: It's not 100% clear to me what you're asking for help with: do you have a `phone_home` configuration which isn't working as you expect? Or you need a higher-level overview of how you might use `phone_home` for your purposes? [21:10] I have not used the feature so I am trying to adopt it. @Odd_Blake: [21:18] one more question is there a way to pass an ssh to a specific user? [21:28] itlinux: yes, https://cloudinit.readthedocs.io/en/latest/topics/modules.html#users-and-groups has a ssh_authorized_keys under user [21:30] itlinux: So phone_home allows you to configure a URL which cloud-init will report information to; the docs are https://cloudinit.readthedocs.io/en/latest/topics/modules.html#phone-home. So to use it, you would need a persistent HTTP server which cloud-init could POST to. Note, however, that it doesn't report the success/failure of cloud-init, and it may not be executed at all if cloud-init fails to [21:30] complete for whatever reason. [21:30] So it may not be appropriate for the retrying mechanism you're talking about. [21:41] @Odd_Bloke: I see that but no example on how to specify the user do you happen to have an example [21:42] got it... [21:42] need to do less tasks :) at the same time [21:43] :)