[09:17] goood morning peeps [09:46] Gooooood day :) [10:50] niemeyer: ｇｏｏｏｏｏｏｏｏｏｏｏｏｏｏｏｏｏｏｏｏｏｏｏｏｏｏｏｏｏｏｏｏｏｄｍｏｒｎｉｎｇ！ [10:56] https://github.com/juju/juju/pull/11323 - "pod spec" becomes "k8s spec" [10:57] nice [10:57] I think "pod spec" was a confusing term because it wasn't just about pods [10:57] Muy buenos días a todos! [10:57] Facu: buen día! no tennis today, right? [10:58] hola Chipaca! nop, I'm quarantined [10:58] ok [10:58] Dmitrii-Sh: Facu: can we have a meet in ~15 minutes? [10:58] Chipaca, sure [10:59] ack [10:59] i've got my manager hat on so beware :-) [11:13] Chipaca, niemeyer, I started this https://docs.google.com/document/d/12dha4GkmjWusAN6cMMZsQKKK-qfpVrvPFw6PRX0wUNA/edit# (so we can collect requirements or discussion items for when we actually talk/decide this) [11:14] Chipaca, I'm ready now! Room 57? ( https://meet.google.com/veq-yfqm-kdk ) [11:16] Facu: sure [11:16] Dmitrii-Sh: ^^ [11:16] Chipaca, we're there [12:21] jam: 👋 [12:21] Chipaca: Is that the hand-washing emote? [12:22] jam: no clearly that's 👐 [12:22] or maybe 👏 [12:23] jam: i was pinging you because i'm (experimentally) changing how our standups work [12:23] niemeyer: ^ you too, probably, at least for now :-) [12:23] jam: there'll be a google doc where you write what you worked on (and maybe what you're hoping to get to); this needs to be written before the standup itself [12:24] then at the standup you just call out things that bring up questions, or blockers, or just highlights; the team reads the standup doc after the meeting to catch up with the rest of the detail [12:24] jam: this serves three purposes [12:25] jam: 1. focus and shorten the standups themselves [12:25] jam: 2. feed into a weekly 'whattup with chamrcraft' email [12:25] jam: 3. make it less hypercritical to have the standups such that everybody can make them every day [12:26] jam: hopefully this makes sense to you [12:26] Chipaca: and a google doc vs a trello/kanban/discourse/something else? [12:26] jam: a trello exists but it's less detailed than this, yes [12:26] Chipaca: I'm fine with it, just have gotten used to other project managementy tools. [12:27] our trello: https://trello.com/b/SGgb8v0h/charmcraft [12:27] Chipaca: the google doc that I see linked above is https://docs.google.com/document/d/12dha4GkmjWusAN6cMMZsQKKK-qfpVrvPFw6PRX0wUNA/edit which points to "Charm Tool Spec" is there a different doc that I missed? [12:27] still nascent, i should add [12:27] the standup doc is a different one [12:27] Facu: Dmitrii-Sh: jam: niemeyer: https://docs.google.com/document/d/1sJQ5-OtK83x2KdePZfp1AC_3FQJrvzXaWFEc_n-4vxY/edit# [12:28] Facu: Dmitrii-Sh: if you could read what i just told jam to double-check it (1) matches what we discussed, and (2) i didn't forget anything please :-) [12:28] jam: the trello will have doing/.../done lanes but does not have them yet fwiw [12:29] Chipaca, looks fine [12:30] jam: i'm still thinking about how best to keep track of some of the things, as quite a bit of at least outreach / library work is not a flow of work [12:31] jam: i also brought forward the standup half an hour fwiw (so we won't have Facu on Mondays once quarantine is up, but otherwise should work for all) [12:32] Chipaca: everyone is away on Friday? What party day is this? :) [12:33] jam: the feast of St. Arnold of Soissons (patron saint of beer) [12:33] I *may* work on Friday, if PyCamp is cancelled here [12:33] jam: i asked for holiday because i was going to run in lisbon :'( [12:34] Chipaca, run in lisbon? [12:35] Facu: lisboa meia maratona [12:35] Wikipedia says the Feast is August 14, are you just trying to get a day of for extra beer? :) [12:35] oh [12:35] Chipaca: you were running from Friday through Monday? That is true dedication [12:36] well, some tourism was also planned [12:37] jam: wrt the feast, maybe we're observing it on the orthodox calendar [12:37] Chipaca: I thought you were just running to Lisbon and back. [12:38] https://www.google.com/maps/dir/London,+UK/Lisbon,+Portugal/@44.7763435,-13.6239022,5z/data=!3m1!4b1!4m14!4m13!1m5!1m1!1s0x47d8a00baf21de75:0x52963a5addd52a99!2m2!1d-0.1277583!2d51.5073509!1m5!1m1!1s0xd19331a61e4f33b:0x400ebbde49036d0!2m2!1d-9.1393366!2d38.7222524!3e2 [12:38] says its about 375hrs [12:38] but you're running, not walking, right? [12:57] Chipaca: I also don't have edit rights to the doc where you want me writing what I've been up to :) [12:57] fixing... [12:57] jam: please reload [13:02] Chipaca: fixed. thanks. Is the intent to have 10min standups/ break other discussions out of it? [13:04] jam: that route cheats because it puts you on a ferry, that's non-canon [13:04] jam: i am not too religious about it, in that i'm ok with a little discussion happening in the standup, but as soon as it gets rambly yes it should break out [13:07] jam: some research is still needed on the subject of pontoon running shoes [13:07] and at what point are you just standing on a boat [13:07] Chipaca: when you hook up outboard motors on the back ? [13:09] jam: maybe it should be a venetian bicycle and call it a biathlon [13:09] or, swimming, but i suck at swimming [13:09] * Chipaca also sucks at running [13:09] ignore that last comment, it was from my knee [13:18] Chipaca, can not edit the doc [13:18] Facu: reload? [13:18] Facu: or use your canonical acct [13:19] Chipaca, reloaded, thanks [13:19] Facu: also, https://artsandculture.google.com/asset/among-the-sierra-nevada-california/IQE1CY9y_Rfy5A [13:20] Chipaca, nice! the author removed the AT-AT! :p [13:23] Facu: it was shrouded by the revisionists [13:23] jeje [13:31] jam: standup? [13:31] yep, just getting out of my last one [13:59] 10 events, sitting in a queue [14:29] we're getting better! that was only double the lenght it should've bene :-| [14:29] been* [14:29] Sorry... /o\ [14:31] niemeyer: we'll get there :-) [14:37] niemeyer, where do you propose to start piling examples? (my only suggestion is that this shouldn't be too public, so far, until we have some APIs more stable) [14:38] Facu: Best thing I can think of right now is a repository under canonical/ with "examples" or "samples" as part of the name, and one example per subdir [14:38] Facu: WE might then take PRs on that [14:38] Facu: That's just one idea though [14:39] niemeyer, ack [14:52] niemeyer: Facu: how about an 'examples' branch in operator itself? [14:55] Dmitrii-Sh: WRT gnuoy and the ceph-iscsi charm, what he needs is a check that he's using the framework in a sensible way (and not a line-by-line python code review) [14:56] ok [14:57] niemeyer: how're you for meetings? can i make your life worse in that sense [15:11] Chipaca: On a call but can do shortly [15:21] * Facu -> lunch [16:14] Facu: does that right any bell? [16:14] charm push ./ cs:~dmitriis/nats [16:14] ERROR cannot post archive: cannot put archive blob: failed to PUT object 3701adfe8af7ba58-3c2ab4da9ad292fc from container charmstore-blobs [16:14] caused by: failed executing the request http://10.24.0.23:8080/v1/AUTH_acf64c68bc7349c98dec46b7e0e72d9f/charmstore-blobs/3701adfe8af7ba58-3c2ab4da9ad292fc [16:14] caused by: Put http://10.24.0.23:8080/v1/AUTH_acf64c68bc7349c98dec46b7e0e72d9f/charmstore-blobs/3701adfe8af7ba58-3c2ab4da9ad292fc: read tcp 10.25.10.42:51284->10.24.0.23:8080: read: connection reset by peer [16:15] It succeeded after I retried [16:15] Dmitrii-Sh, let me see [16:16] Dmitrii-Sh, no clue anyway... that's the old charmstore? [16:16] yes [16:21] ah, don't know much about its infrastructure [16:21] sorry [16:22] Chipaca: Apparently, there is no way to install and disable a snap. So the race condition I was after was because of the following sequence: [16:22] 1) snap got installed; [16:22] 2) systemd service start attempts by snapd followed that before a config was rendered by a charm; [16:22] 3) the charm then rendered the config but a `systemctl service start ` made by it then failed once. It then succeeded on a retry. [16:23] Facu: ok, np [16:24] Dmitrii-Sh: i'm not sure i understand what you mean [16:24] Dmitrii-Sh: snaps can ship services that are disabled until they are configured [16:24] Dmitrii-Sh: snap services can be disabled after installing the snap, but there is no way to say 'install this but disable the services' [16:25] maybe that's what you mean [16:25] Dmitrii-Sh: what is the snap? [16:25] Chipaca: yes, maybe this is a bug of a particular snap - services shouldn't be started before a config is supplied from an external entity (the charm in this case). [16:26] and, yes, there's no 'install --disable' [16:26] Dmitrii-Sh: as a blanket assertion that is false [16:26] "services shouldn't be started before a config is supplied" is false, i mean, in general [16:26] Chipaca: it's the nats snap created by the Anbox team [16:26] it is perfectly possible to ship a service that works without a config [16:26] if it does need a config, then its install hook should disable the services until that config is provided [16:27] easy to do [16:27] Chipaca: yes, in this particular case it didn't make sense to start it before config is available. [16:27] and the fix was to do snap stop nats --disable on install [16:27] and then start --enable after the config was renderd [16:27] rendered* [16:30] Dmitrii-Sh: the real fix is for the snap's install hook to do 'snapctl stop --disable' itself [16:37] ok, I'll suggest that to Simon [16:43] Chipaca, do we want or not to have copyright header in otherwise empty files? (`__init__.py`) [16:43] Facu: nope [16:43] ack [16:43] Facu: that's why the previous checker explicitly ignored empty files [16:43] -not -empty [16:43] nice [16:44] `find` is so complex [16:44] tbf the -type f should be the first predicate, but ¯\_(ツ)_/¯ [16:55] Chipaca: SOrry, left one to join another.. I'm available now.. how're things there? [17:43] Got another person asking me about how to easily retrieve pod fqdns and expose them to a different app without relying on ingress-address == cluster-ip https://bugs.launchpad.net/juju/+bug/1867168 [17:43] seems like this will be a common theme for stateful apps that use headless services in K8s https://kubernetes.io/docs/concepts/services-networking/service/#headless-services [18:06] Moved all tests to unit tests, removed Makefile, improved README for dependencies and how to test: https://github.com/canonical/operator/pull/184 [18:07] niemeyer: and I'd stepped out to get some sunlight [18:09] niemeyer: i need to get dinner started, but it's not a 100%-of-my-concentration thing once i've kicked it off so we can probably chat while i do it, if that works for you -- ping me on signal or tg if it works for you too [18:09] * Chipaca will probably be able to work fine with 9 fingers after all [18:11] Chipaca: Sounds dangerous :) [18:11] We can do it tomorrow [19:13] if I do juju status, I see for postgresql/0: Workload=active Agent=idle [19:14] for my-super-charm the Agent is also idle, but the Workload is in "unknown" state... is this ok? there's something charm side that should change that? [20:36] * Facu eods [20:36] Facu: 👋 [20:36] Facu: have a good one