=== cppforlife is now known as cppforlife_ === seelaman` is now known as seelaman === 18VABY2RP is now known as lazyPower [01:54] wallyworld: hey, yes I think i got it [01:55] great :-) hope i made sense with my ramblings [01:55] you need sleep, so we'll catch up tomorrow [01:55] you did, I think I stayed as close as possible :p [01:58] no problems, we can see how it turns out and adjust if the idea is too insane [02:01] wallyworld: ow please sleep is a poor replacement for coffee [02:03] wallyworld: ok i have updated now its backwards compatible and consistent [02:04] if you go for a setstatus it will use the agent [02:04] and if you call explicitly you get the right ones [02:04] perrito666: awesome ty, will look after my cuurent meeting and provide feedback, you need to sleep :-) [02:04] now yes I need to sleep [02:04] cheers man === kadams54 is now known as kadams54-away [03:19] wallyworld: updated the diagram, do you want to take a look and see if it's what you had in mind? === axw_ is now known as axw [03:20] sure, ty, looking [03:22] axw: looks good ty [03:22] ta [04:38] axw: if you have a moment, could you please look at a small review for a 1.22 bug http://reviews.vapour.ws/r/748/ [04:39] wallyworld: sure [04:44] wallyworld: reviewed [04:44] need to make lunch... bbs [04:45] tyvm [05:08] axw: so by the looks, Datastore in code is what's in the spec as Storage Provider, and Datastore.Kind is analogous to Storage Provider type? [05:09] also, i did find in the spec --disks for add-machine [05:29] wallyworld: no, Datastore is an early name I came up with for storage instance [05:29] I may just rename it to StorageInstance ... [05:29] ah right ok [05:30] Kind is block|filesystem|object [05:30] :q [05:30] oops [05:30] and a StorageInstance will reference the provider type / pool from which it came [05:31] wallyworld: correct === kadams54 is now known as kadams54-away [06:08] thanks axw, i need to buy some glasses [06:08] wallyworld: nps. to be fair, someone may want the string; I don't think the backend should be in the business of rendering though [06:09] its job is to tell the user about changes in the model [06:09] wallyworld: there's a followup to change the API right? [06:09] yes, but the megawatcher already does some rendering i think [06:10] does there need to be an api change? [06:10] wallyworld: I'm sure you'll need to change something in apiserver/params at least... [06:11] let me look [06:11] wallyworld: sorry, I assumed you were splitting up the branch [06:11] I believe you need to update AllWatcher, which uses MegaWatcher [06:11] i've not ever done much with megawatcher so am not sure of all the artefacts [06:12] wallyworld: re rendering, perhaps you should double check with fwereade what he wants it to do [06:12] axw: looked in params, it seems the api layer just json marshals the struct directly [06:13] huh, ok [06:13] but i can see a params_test that will fail [06:13] so it does [06:13] i'll wait for the failure and fix the params_test [06:14] i think i had it at theback of my mind that i just needed to change the machineinfo directly but couldn't remember why === urulama__ is now known as urulama [06:37] wallyworld: any opinion on the name of the thing I was talking about before? StorageInstance? StorageUnit? (it's currently "Storage" in state, which has too many meanings...) [06:37] um [06:38] the thing being an instantiation of a store declared in a charm [06:39] maybe StorageInstance just to make it not Storage and we can search/replace later if a better name comes up [06:39] sounds good [06:56] o/ [07:02] morning dimitern [07:03] wallyworld: just sent you an email with a UML diagram - please take a quick look when you have a moment to see if it's what you had in mind [07:03] hey axw [07:03] axw: sure, ty [07:04] * dimitern is amazed [07:04] no blockers todat yay! [07:05] :) [07:05] and here I am with nothing to land [07:05] just in time for the 5 branches I had in mind :) [07:05] woot [07:05] aww ;) [07:06] hmm.. after all the live testing my ec2 bill grew a bit - $0.32 now hah [07:07] axw: thanks yeah, that's something we can start with and evolve over time [07:07] wallyworld: cool [07:34] dimitern: thnx for review on list charms [07:35] dimitern: m moving them into a separate client but have taken ur feedback on board :0 [07:35] anastasiamac, cheers :) [07:36] anastasiamac, I'd like to have a look when you propose it please [07:36] dimitern: thnx! m hoping later on tonite - got my hadns full of kids atm [07:37] dimitern: hands that is [07:37] anastasiamac, np [08:06] dimitern: committed updates with new facade [08:06] dimitern: i have copied CharmInfo only in this PR [08:07] dimitern: I'd like to land it b4 copying the rest of charm stuff from old client [08:07] dimitern: like AddLocalCHarm, AddCahrm and ResolveCharm [08:07] dimitern: this way old client's dealings with charm will be "deprecated" :) [08:08] anastasiamac, good plan, but we shouldn't remove them from the old facade [08:09] dimitern: agreed.. Hence "copy" :) [08:10] anastasiamac, right :) [08:11] dimitern: thnx for double-checking! it's amazing to work with ppl that r diligent and care! [08:15] anastasiamac, oh c'mon :) thanks [08:20] dimitern: :D === Spads_ is now known as Spads [08:57] good morning, I'm quite new to Juju and having problems [08:57] am I in the right place to ask for some help? [09:04] Muntaner, hey, welcome! [09:04] Muntaner, You can ask questions either here or in #juju [09:05] I asked in Juju, if it's ok I can copypaste my question here [09:05] Muntaner, This channel is more about Juju development, while #juju is more for users, charmers, etc. [09:05] Muntaner, no need - I'll have a look [09:06] ok, thanks :) [09:17] morning [09:35] morning TheMue [09:35] TheMue, feeling better today? [09:35] dimitern: heya [09:36] dimitern: in IT I would "project not yet done, but it's getting better with each iteration" ;) [09:36] would say [09:37] TheMue, glad to hear it :) [09:39] dimitern: yeah, me too. today I'm working from a convenient place in the living room, having my cat beside me. better than alone in the home office. [09:41] TheMue, :) nice [09:41] dimitern: yep [10:01] * TheMue looks around [10:05] TheMue, voidspace, sorry guys - omw [10:05] dimitern: o/ === rvba` is now known as rvba [12:22] good morning [12:43] perrito666: o/ [13:15] perrito666: feeling better today? [13:18] wwitzel3: yes tx a lot, was a state of mind more than of body and I really didn't want to share my frustration in negative ways [13:19] my grand mother used to say that you should not do hangouts when angry [13:19] perrito666: yea, I can understand that [13:19] perrito666: haha [13:20] since she died before google rose above the search engine I might be missquoting her :p [14:31] TheMue, voidspace, ping [14:32] dimitern: pong [14:32] TheMue, hey - we have approval for the sprint (check your mail) - please submit the form [14:32] dimitern: just seen Sarahs mail :) [14:33] TheMue, :) [14:58] ericsnow, perrito666 ping [14:58] wwitzel3: pong? [14:59] oh i guess I'm a minute early [14:59] says my clock I still have 1min [14:59] perrito666, wwitzel3: I'm going to the upstart-systemd-migration call [14:59] dimitern: my flights are bad for the times we're planning, cannot come early enough or have to leavy too early. I would would prefer coming on Monday during morning and leave on Saturday then. what do you say? [15:02] TheMue, if there are no suitable flights - yeah, but I'd still wait to hear from the travel agent to be sure [15:03] TheMue, she might have better options to offer [15:04] dimitern: sure, but my usual portal is very good, checking lots of flight. I have also to do one hop [15:04] perrito666, wwitzel3: I did wrap up the tests for instance and conn_availzones [15:05] TheMue, which airport you usually take - hamburg or bremen? [15:06] dimitern: bremen, hamburg is badly to reach [15:07] dimitern: but maybe this time it's worth it, will take a look. from here to hamburg it's about 2:30h (with train and sub). [15:12] dimitern: cool, thanks [15:52] is there an api for hook tools or is the only way to use them to invoke them from the shell? [15:52] e.g. relation-set [15:53] I think it would be nice to be able to use them programmatically from within a hook or action [15:55] we only run one jujud per instance, right? [15:55] I believe so [15:55] my understanding is that we *used* to run one jujud per agent (or something like that) [15:56] ericsnow, 1 jujud machine per instance [15:56] you could use a local http service (is that insane?) [15:56] then that could be wrapped for different languages [15:57] or, rpc server as we do for api/apiserver [15:57] but that might be overcomplicated [15:57] sinzui: do you mean 1 jujud *per* machine per instance? [15:58] ericsnow, 1 jujud as a machine agent per instance. 0-* jujud unit agents per machine. [15:59] sinzui: so there may be more than one jujud running? [15:59] ericsnow, add-machine will give you a machine with just 1 jujud running on it. deploy will usually give 2 jujuds running as machine and unit. [15:59] sinzui: ah [15:59] ericsnow, and with subordinates, many more juju unit agens get added [15:59] sinzui: but there is only one upstart job for jujud (or one for each jujud that runs)? [16:00] sinzui: so one jujud per machine agent [16:00] ericsnow, there is an upstate job for the machine, and for each service running [16:00] jujud_ip=os.Getenv("JUJUD_IP") [16:01] perhaps [16:01] sinzui: ah, okay, that's just what I needed to know [16:01] sinzui: thanks :) [16:02] ericsnow, this is what I see on my apache3 machine with has a lot on it [16:02] http://pastebin.ubuntu.com/9762266/ [16:04] * perrito666 notices he forgot to add something to his patch [16:04] Hi voidspace ericsnow perrito666 TheMue dimitern wwitzel3: can anyone help with these two issues that block 1.21 an 1.22? https://launchpad.net/juju-core/+milestone/1.21-rc1 [16:05] sinzui: taking a look [16:07] sinzui: perfect [16:13] dist-upgrades don't happen on running instances, right? (other than as part of the initial start-instance stuff) [16:14] since everything in juju is series driven :) [16:15] ericsnow, correct. Deploy a new service to get the new os, not upgrade an existing one [16:28] sinzui, i have a clue re bug 1403738 [16:28] Bug #1403738: upgrade tests fail on multiple substrates with revision 24c1b80d

[16:30] sinzui, I think the fix for that bug wasn't backported to 1.21 and hence current blocker - bug 1411502 [16:30] Bug #1411502: ERROR upgrade in progress - Juju functionality is limited

[16:30] ah, thank you dimitern [16:32] sinzui, yep - I've just checked the source of 1.21 - no backport in sight === kadams54 is now known as kadams54-away [16:38] dimitern, can you delegate the two tasks you highlighted in mail for the 1.21 bugs [16:38] to some us based folks so we can get them landed today if possible [16:40] alexisb, two tasks? I haven't actually had time to look at the other bug, just the upgrade one [16:41] well the backport tasks [16:41] alexisb, sure === kadams54-away is now known as kadams54 [16:43] ericsnow, wwitzel3, can any of you find some time to backport https://github.com/juju/juju/pull/1343 to 1.21, so we can unblock people affected by bug 1411502 please? [16:43] Bug #1411502: ERROR upgrade in progress - Juju functionality is limited

[16:44] dimitern: if wwitzel3 can't, I'll do it [16:44] dimitern: ping [16:44] ericsnow, thanks! I'll assign it to you - please sync-up with wwitzel3 and reassign if needed [16:45] dimitern: did you think that goamz already supported the DescribeNetworkInterfaces call? [16:45] dimitern: it doesn't look to me like it does [16:45] dimitern: I can implement ec2 filtering without it [16:45] voidspace, it does, but the filtering by instance-id is not implemented by the ec2test server [16:47] dimitern: ah, I can't see it in the goamz source code [16:47] dimitern: just filtering by network name might be *easier* anyway [16:51] voidspace, well, I started doing it for ec2 (the Environ.NetworkInterfaces() implementation) and realized I can't test it without the filtering [16:51] voidspace, and resorted to using Instances instead and the NetworkInterfaces from there [16:51] voidspace, unfortunately there were too many distractions today to finish it :/ [16:51] dimitern: ok [16:52] dimitern: I have a basically done implementation *without* using DescribeNetworkInterfaces [16:52] just filtering on the provided network IDs [16:52] but I have an oddly failing test - possibly due to the test server [16:52] just investigating [16:54] voidspace, cheers - i need to step out now though [16:54] happy weekends y'all ;) [16:55] dimitern: o/ have a good weekend === kadams54 is now known as kadams54-away === kadams54-away is now known as kadams54 === urulama is now known as urulama__ === kadams54 is now known as kadams54-away [17:45] ericsnow: have a moment for a completely non work related question? [17:45] perrito666: sure [17:45] ericsnow: priv === mbarnett` is now known as mbarnett [18:27] EOW [18:27] happy weekend everyone === kadams54 is now known as kadams54-away === kadams54-away is now known as kadams54 [18:52] sinzui: does bug #1411502 actually apply to 1.22? [18:52] Bug #1411502: ERROR upgrade in progress - Juju functionality is limited

[18:52] sinzui: from what dimitern indicated, this should be fixed there already (with 1403738) [18:53] ericsnow, I think the change is in 1.22 I believe 1.22 was master at the time [18:53] ericsnow, If you agree, I will remove the task [18:54] sinzui: k [18:54] sinzui: I have a patch up for review that backports the 1.22 fix [18:55] ericsnow, okay. I will remove the 1.22 since the change comes from 1.22 [19:02] ericsnow, perrito666 do we have an idea on this bug: [19:02] https://bugs.launchpad.net/juju-core/+bug/1410876 [19:02] ?? [19:02] Bug #1410876: Error executing lxc-clone: lxc_container: utils.c: mkdir_p 220 Not a directory - Could not destroy snapshot %s - failed to allocate a pty; Insufficent privileges to control juju-trusty-lxc-template

1.22:Triaged> [19:02] alexisb: I haven't looked at that one [19:02] wwitzel3: ^^^ ? [19:03] alexisb: I haven't either [19:04] sinzui, lp 1410876 is a blocker for 1.21 right? [19:05] alexisb: I am not sure I even know how to reproduce it [19:05] alexisb, I think so since it was found in openstack testing [19:05] well if we need repro steps to make progress we need to make that request in the bug [19:07] sinzui: do you know how to reproduce this or should I ask for more info? [19:08] perrito666, We will need more info. We obviously never saw this in our testing [19:09] sinzui: ah, i assumed since you re-classified it you knew how to make this happen [19:09] perrito666, No the stakeholder makes it important. it was found in oil testing [19:17] aghh I hate whatever shortcut my client has that allows me to close this [19:19] ericsnow: wwitzel3 did you guys knew that monday apparently is a holiday there? [19:19] perrito666: yep [19:21] ericsnow: wwitzel3 then all we said was happening on monday will be happening on Tu, right? :p [19:21] I am going to be sooo alone in here on monday [19:21] perrito666: yep [19:52] back, ericsnow I haven't looked at that one [19:52] perrito666: yeah, Mon = Tue ;) === kadams54 is now known as kadams54-away === kadams54-away is now known as kadams54 === kadams54 is now known as kadams54-away === tvansteenburgh1 is now known as tvansteenburgh === kadams54-away is now known as kadams54