=== kadams54 is now known as kadams54-away === kadams54 is now known as kadams54-away [07:36] mornin' all [07:42] morning rogpeppe, everybody [07:43] mhilton: hiya [07:44] rogpeppe: If you find time it'd be great to get a review of https://github.com/juju/charmstore/pull/219 [07:44] mhilton: looking [07:49] mhilton: reviewed [07:49] rogpeppe: many thanks [12:10] rick_h_: gui charm branch is going well, almost done, going afk, will propose after lunch [12:25] frankban: <3 ty much [12:54] uiteam taking over guimaas to prep for demos for the UOS session today [13:22] * frankban afk === kadams54 is now known as kadams54-away === kadams54-away is now known as kadams54 === kadams54 is now known as kadams54-away === kadams54-away is now known as kadams54 [13:57] uiteam I need two reviews + QA for https://codereview.appspot.com/174170043 (GUI charm/python). Anyone, thanks! [13:57] frankban: will do but will probably be after standup [13:57] rick_h_: thanks np, I'll ping Jeff too [13:57] frankban: so we'll shepard it through probably after your EOD to land [13:57] frankban: ty for helping with that! [13:59] rick_h_: doe sthe next charm release include a GUI one? anyway, I am available for releasing the charm later or on my monday morning, and now good luck with UDS [13:59] frankban: yes it will. There's one card left before a new gui release with added services [13:59] I'll be trying to pull it all together and want to make sure whatever release we do to prodstack on monday includes it [13:59] cool [13:59] frankban: ty, I'll let you know how far we get today [14:16] * frankban afk again, trying to finish lunch [14:28] uiteam I've got a demo of gui running on the develop branch. If you plan on borking trunk please don't :) (I'm pretty sure I'm safe) [14:43] well it appears that my internet is shot today :/ [14:45] hatch: kadams54 found a bug in prepping for the demo with hiding units on machine view [14:45] hatch: kadams54 will leave this up on guimaas for debugging and such after the session [14:45] Hit me [14:45] kadams54: hatch so one of the services is in error, with a relation in error [14:45] when I hit hide, it doesn't hide [14:46] if I hide a second service, both the new one and the first one that failed to hide at first now hides [14:46] Ah, so something with the error status is borking the hide action. [14:46] probably [14:47] Well, I guess that's better than hitting hide on the second service and having it do something completely random :-) [14:47] yea [14:50] hatch: morning, https://codereview.appspot.com/174170043 [14:53] rick_h_: bleh ok :/ [14:53] frankban: awesome will take a look [14:53] today it appears I'm on dialup lol [14:53] lol [14:53] you can watch the animated gif of our UOS session [14:54] uiteam anyone want to be in the actual hangout for the chat? the etherpad for notes is at http://pad.ubuntu.com/uos-1411-whats-new-and-upcoming-in-the-work-of-juju-ui-engineering and the irc room is #ubuntu-uds-devops-1 [14:56] rick_h_: what's the hangout URL? [14:56] kadams54: no idea yet, working on it [14:57] rick_h_: lol yeah I'll watch the stream in B&W via rabbit ears on my old crt [14:58] frankban: so this get() method that's on the config - is that provided by the json serializer? [14:59] hey gui team - i'm helping you get setup [14:59] has anyone made a hangouts on air link for the broadcast yet? [14:59] lazyPower: no, I thought a sessoin chair or someone was going to [15:00] hatch: it's a default method on all python dicts [15:00] lazyPower: do I need to create just a hangout on my canonical account and get you a link or how does this need to work to be on air? [15:00] doing it now [15:00] lazyPower: ty [15:00] https://plus.google.com/hangouts/_/hoaevent/AP36tYek4Fu7CwFd9SQQ9f4CrPfzUOyHQ_E-pq1YS2zjwGaUPtyYBw?authuser=0&hl=en [15:00] hatch: python -c 'help({}.get)' [15:01] ohh intersting [15:01] thanks [15:01] much nicer than all my conditionals [15:05] hatch: FWIW, to take a look at all dict methods: python -c 'help({})' and that works for all python objects [15:06] dammit, hangouts won't let me stay on. [15:07] rick_h_: rather than hopping in and out i'll just stay off. sorry. [15:08] the video isn't loading for me, might just be my connection though [15:08] http://summit.ubuntu.com/uos-1411/meeting/22387/whats-new-and-upcoming-in-the-work-of-juju-ui-engineering/ [15:15] kadams54: are you on the AS bug? [15:16] Yeah, I will after the demo. [15:18] alright cool [15:23] frankban: why lambda vs function? [15:24] read some docs, dont' really see when it's useful [15:25] hatch: simple function, just one line, mostly a personal preference, sometimes I find the slightly more expressive... [15:25] hatch: no hard rules there [15:25] got it [15:25] I see you added a port validator [15:25] that's pretty nice [15:32] frankban: code looks good just pulling the branch down [15:32] but at 80kbps it's taking a while :) [15:33] :-/ thanks for looking at it [15:33] no thanks for picking up my slack :) === tvansteenburgh1 is now known as tvansteenburgh [15:46] frankban: +1's [15:47] hatch: cool, landing it [15:53] uiteam call in 7 kanban please [15:58] hatch: kadams54 ok, talk is over so you guys can have the guimaas for debugging if that helps [15:58] some issues with the demo on a live env it looks like [15:58] uiteam gui has the guimaas for a bit for new feature debugging [15:58] rick_h_: Thanks - I'll dig into the bugs after standup. [16:58] rick_h_: What's the publicly-exposed URL for the guimaas? [17:05] kadams54: check out https://docs.google.com/a/canonical.com/document/d/1v3Bs7JU_mYrwRFbXWcQzNRje_EYz0nqqz9yaR4Nadro/edit [17:06] kadams54: http://maas.jujugui.org/MAAS is the maas side, not sure which item you're looking for [17:07] rick_h_: thanks [18:03] rick_h_, hatch: I think I found the bug that came up during MAAS [18:03] er, during demo. [18:03] oh? [18:03] spillit [18:04] Yeah, a change I made in fixing the last bug caused this one. The machine change event wasn't being fired for fades, so the token in MV wasn't being re-rendered. [18:04] I found that while working on decreasing change event noise [18:05] So it's actually already fixed. [18:05] I think. [18:05] :-) [18:07] lol [18:07] awesome [18:07] you added a test for it, right? :) [18:08] yes [18:09] but the bug that was in the demo was the service token not being hidden [18:09] nothing to do with the mv [18:15] ah, true [18:15] dangit [18:54] kadams54: that was in develop, is this fix in your current branch? === kadams54 is now known as kadams54-away === kadams54-away is now known as kadams54 [19:14] rick_h_: Yes, though as hatch pointed out, the problems you saw were in service view, not machine view. [19:14] So I think we still have bugs [19:34] rick_h_: I'm having problems SSHing into MAAS. I'm assuming you took the SSH keys off Launchpad? [19:35] kadams54: yes, just ssh-import-id [19:36] I have three keys on launchpad… not familiar with ssh-import-id… did it grab all three, or just one? [19:38] kadams54: what's your lp username? [19:38] kadams54 [19:38] https://launchpad.net/~kadams54/+sshkeys [19:38] kadams54: ok all three imported try again [19:40] Hmm, still no go. What username should I be using? kadams54? [19:40] ubuntu@ [19:43] http://pastie.org/private/f5gr8xhesfyw7qiosjiew [19:45] rick_h_: Wow, OK, I think I may have badly misunderstood something due to my sshuttle newbie-ness. I was expecting it to use the SSH keys to login, so when I'd get the password prompt above I assumed something was wrong. [19:46] But then I tried the password specified in the Google Docs and that seems to have worked? [19:46] At least, sshuttle reports I'm connected [19:48] I always read that as ss-shuttle :) [19:53] rick_h_: so this api layer - we rely on a separate module to make the requests so really it's just a thin wrapper around that modules public method [19:54] hatch: right, that's my thought. [19:54] so my question is, how far does this rabbit hole go? Do we want to create a new 'network' layer? [19:54] if we want to offer this api to general users we'll need to offer a stack [19:54] hatch: sec, let's chat in 3min [19:54] sure just ping whenever [19:56] hatch: meet you in the standup [19:57] joining [19:58] rick_h_: just having an issue joining [19:58] couple mins [19:59] kadams54: you're in? [20:04] rick_h_: Maybe? Just realized that sshuttle was dying after I started it up; had to take it out of daemon mode to see the message. [20:04] Apparently I need to reboot in order to use it :-\ [20:06] kadams54: reboot your machine or the maas controller? [20:06] Reboot my machine. It's a workaround for a bug in MacOS X 10.7 [20:07] I also have an AWS env up and running that I'm currently using the debug a potential fix [20:07] So I'm going to hold off on rebooting for a bit. [20:07] k [20:07] I was able to replicate the bug on relations with errors [20:07] hatch: It's a problem in findUnrelatedServices [20:08] ahh [20:19] uiteam: OK, general juju question here. I had juju-gui, wordpress, and mysql deployed in my AWS env. The JS code was throwing an uncaught exception when trying to process a wordpress loadbalancer relation that looked like it was meant to be connected to nginx. [20:20] But I had no nginx deployed… so why would the relation even exist? [20:20] kadams54: to my knowledge, it shouldn't. [20:20] is your env still up? [20:21] jcsackett: I'm switching juju-gui-source on it right now, so no, but it will be available again shortly. [20:21] kadams54: that's alright, i want you to look at something via cli anyway, if you can. [20:21] kadams54: juju status wordpress [20:22] check the relations on that--then we'll know if the gui is processing something weirdly, or if your juju env has a weird thing. [20:25] relations: [20:25] db: [20:25] - mysql [20:25] loadbalancer: [20:25] - wordpress [20:26] https://ec2-54-68-167-143.us-west-2.compute.amazonaws.com/services/:flags:/as/ [20:27] kadams54: I think wordpress has a peer relation [20:27] What's a peer relation? I've heard that term before, but never really understood what it was. [20:28] well essentially it's a relation to something else installed in the same charm [20:29] kadams54: they don't have lines anywhere because it's part of the same service [20:30] Ah, OK, that would make sense [20:30] Peer relations are a different matter, in that peer relations are the mechanism by which service units share information internally. It's not unusual for peer service units to maintain convenient caches of distributed information in their own peer settings, and this inevitably involves generating settings keys at runtime. [20:30] ^ kadams54 [20:30] from the docs [20:31] I don't know how to properly use one, I just know what it is haha [20:31] uiteam: OK, https://github.com/juju/juju-gui/pull/657 needs reviews and QA to fix this bug. [20:31] lol [20:31] It's a very small PR, but needs a real env for QA. [20:32] kadams54: can you add a comment why it might not exist in the code :) [20:32] huh, so peer relations are generated automatically? that's peculiar. [20:33] hatch: Done. [20:33] kadams54: so when you real real env, would lxc work? [20:33] Yeah? [20:34] Shoot, now that I think about it, sandbox ought to have the same problem. [20:34] As long as you deploy a service that has a peer relationship. [20:34] but I don't think it did [20:34] I've used wordpress a lot.... [20:35] so is this actually the fix? hehe [20:36] if it doesn't come up in sandbox or lxc, history says it's a timing issue. [20:42] kadams54: I'm spinning up an lxc env to test this out [20:43] NOOOOOOOO! [20:43] Just kidding. [20:43] lol [20:43] maybe last bug? [20:43] * hatch ducks [20:43] lol [20:43] haha [20:43] hatch ducks [20:43] quack quack [20:44] * kadams54 suppresses the urge to punch ducks. [20:50] kadams54: so I am getting an error in develop when even trying to use the gui because of a hook failure in mysql [20:50] did you get the same bug? [20:51] kadams54: hatch what's up? [20:51] so on lxc I deployed juju-gui wordpress and mysql via the cli [20:52] switched to the guy and related wordpress and mysql [20:52] now the gui fails with a hook error in utils [20:52] (the mysql service failed on start) [20:53] I am not sure why mysql failed on start....but I hope that someone else can reproduce this [20:53] because I can't actually use the gui heh [20:54] ok so this is something different than the show/hide thing? [20:54] I can't even get that far [20:54] the gui is unresponsive on load to the canvas [20:54] heh, what did you break? :P [20:54] I'll take a screenshot of the traceback [20:54] hatch: so this is even before using the develop branch [20:55] this is using the local charm version [20:58] this is because agent_state_data is undefined on utils.js:1418 [20:58] ^ kadams54 didn't you do something about agent_state_data resiliency recently? [21:04] Sure seems like I did, didn't I? [21:04] https://github.com/juju/juju-gui/pull/640 [21:04] hatch: ^ [21:05] looks like we need to also add some here [21:05] although I'd like to know why there is no agent_state_data though [21:05] was that juju-core bug ever created? [21:05] Odd [21:06] You shouldn't be getting that error if you have the patch in that PR [21:06] Are you sure your local charm has the latest code, including that fix? [21:07] hatch: And no, the juju-core bug was not created [21:07] At least, not that I know of [21:08] uiteam: Jenkins merge builds seem to be hung. The current build's been going for 1 day and 1 hour. [21:08] http://ci.jujugui.org:8080/job/juju-gui-merge/782/ [21:08] kadams54: no the bug I'm seeing is not related to your branch/fix - it's in the relation line decoration [21:08] code [21:08] kadams54: you can just kill the build [21:09] although you may need to log into the box to then kill the node process [21:09] hatch: you said utils.js:1418, right? [21:09] correct [21:09] not in develop though [21:09] in release [21:09] hatch: Yeah, that's the code I changed. [21:09] ohh so develop has this fixed? [21:10] Yes. [21:10] ohh ok cool heh I'll try swapping the branch [21:10] hatch: I've never gotten on the jenkins box… not sure how? [21:16] kadams54: ok you are correct develop does have it fixed :) [21:16] * hatch shame [21:25] hatch: yay! [21:25] lol [21:42] kadams54: I have juju-gui wordpress-mysql [21:42] if I highlight mysql then fade wordpress juju-gui shows up [21:42] wasn't this what your branch fixed? [21:44] Yes. [21:44] Looking to see if there's a regression. [21:46] Oh, ah ha [21:46] I did fix it [21:46] On the PR that had its merge build hang for a day. [21:46] So the fix still hasn't landed on develop [21:46] Grr. [21:46] hmm ok - is that merge running now? [21:46] yes [21:47] ok cool well once it's done and you update your pr can you ping me for the qa? [21:47] Once it lands successfully, I'll rebase it into the PR. [21:47] Yup [21:47] my env is all set up for it now so it'll be quick [22:04] hatch: PR updated [22:04] pulling down thx [22:04] kadams54: hatch so once things land is the guimaas still setup? [22:04] e.g. can we change the source to master and then back to develop and QA the exact same env? [22:05] rick_h_: Not sure I follow… [22:06] kadams54: so taking the gui in guimaas and setteing the source tree back to stable right now [22:06] Yup [22:06] kadams54: when we think a fix is landed, I'll update it back to develop and see if it works [22:06] Yeah, sounds good. [22:06] Which reminds me, since I rebooted, I ought to be able to sshuttle in [22:08] Hurrah! I'm on the MAAS box! [22:09] +1 [22:10] we should really sort the added services tokens [22:10] they are always in a different order [22:10] :) [22:11] uiteam: need one more (tiny) review on: https://github.com/juju/juju-gui/pull/657 [22:12] Got it [22:17] Makyo: thanks! [22:38] uiteam: i need two reviews, one QA on https://github.com/CanonicalLtd/blues_browser-charm/pull/13 [22:38] lxc env is fine for QA. [23:35] rick_h_: Branch is landed. Update guimaas at your convenience.