=== marlinc_ is now known as marlinc [01:53] sidnei, its missing some of the waits for units running stuff, but the python-env branch of the new deployer is working against pyjuju [01:55] * sidnei gives it a shot [02:07] sidnei, pushed some fixes for terminate [02:07] er reset [02:07] hazmat: ok, not using that on the pyjuju one yet, still doing nova delete [02:09] hazmat: seems like it worked, except for the waits :) [02:10] sidnei, nice [02:10] hazmat: https://pastebin.canonical.com/91208/ [02:11] time for bed then :-) [02:13] cheers === ismail is now known as Guest64702 === jppiiroinen is now known as jppiiroinen|afk === marcoceppi_ is now known as marcoceppi === wedgwood_away is now known as wedgwood [12:58] marcoceppi: so I did a talk on juju @ a lug last night [12:58] and the drupal6 charm is broken [12:58] should I file a bug to unpromulgate? [12:58] jcastro: I'd at least file a bug [12:59] I'm a little unclear as to the unpromulgate process, does it just happen whenever or does it have to go through orphaned process, etc [12:59] no clue [13:00] Plus, if we unpromulgate it, where to we keep the charm for historical reasons? ~charmers/ will likely cause some oddites with ingestion in the GUI [13:01] well in any case [13:01] this would make a good test case to see what happens [13:01] * marcoceppi nods [13:01] can we kick it back to brandon's namespace? [13:01] jcastro: I can't, no one but brandon has access to that namespace [13:01] BUT [13:02] it looks like he has a copy in his personal branch already so that would "just happen" [13:02] First audit casualty? [13:06] yeah [13:07] audit by failing a demo [13:08] marcoceppi: I am publishing the new calendar now [13:08] for reviews [13:09] hmm. that was known broken.. [13:09] marcoceppi: ttrss looks like it's ready for round 2 [13:09] i ended up pushing a drupal7 one to my ns [13:09] when i was doing a demo. [13:10] hazmat: yeah, I think it's been general knowledge that it's broken. Just not sure what to do about it until the recent audit sessions [13:10] no more mr nice guy [13:10] it doesn't work ---> see ya! [13:10] agreed [13:11] * marcoceppi warms up the unpromulgater [13:11] if anyone has any cycles today for the review queue, that would be <3 [13:13] marcoceppi: hey so so far Andreas has been asking for more charm-writing content in the charm school [13:13] want to concentrate on that this afternoon? [13:13] jcastro: sure [13:16] I sent you instructions for running the ubuntuonair thing [13:16] since it'll be you and mims today [13:17] jcastro: oh yeahh [13:18] thanks [13:33] does juju go version only support constraints like arch, cpu-cores and mem? [13:34] freeflyi1g: yes, for now [13:35] mgz: thx [13:35] feedback welcome about what you want most [13:36] mgz: in python version, maas-name is very useful :) [13:52] hazmat: is there anyway to fix a machine agent showing "down"? I've ssh'd into the unit and restarted it, but the status is the same. [13:53] (besides redeploying) [14:06] jcastro: are you guys going to be at velocity? [14:15] No, we got declined [14:19] wedgwood, could you pastebin the machine agent log [14:19] hazmat: it's looking like a zookeeper problem [14:20] wedgwood, restarting the machine agent should resolve that [14:20] wedgwood, connectivity? [14:20] hazmat: well, the zookeeper keeps dying [14:20] wedgwood, g+? [14:21] hazmat: sure [14:21] i landed some fixes to do better back off on retry, but its only in the ppa as of two days ago. [14:21] * hazmat refills on coffee [14:51] ahasenack: hey do you have any specific things you'd like covered in the charm school? Or just general charm authorship stuff? [14:53] jcastro: hm [14:54] jcastro: "best practices" for interface creation I guess, and simple examples of set and get relations, with emphasis on the fact that relations might not be established yet so it should noop [14:57] jcastro: or maybe for this first one just explain the hooks === cereal_b1rs is now known as cereal_bars [15:28] jamespage, does java set the default maxheap size based on the ram of the system it's running on or is it set when the jre is compiled? [15:37] gnuoy, its never clever - there is a jre default and then whatever the application specificies [15:37] cassandra does some auto-sizing [15:37] others don't [15:37] jamespage, lovely, thanks [15:38] I was poking around the help for juju-core 1.10, what version (if any) can you specify an alternate .juju/environments.yaml (or different .juju "home")? I couldn't find it any of the help topics for juju-core [15:52] marcoceppi: JUJU_HOME iirc [15:52] sidnei: yup, thanks! [16:30] is there a snippet somewhere on how to do a wait loop for apt-get in case an other charm is installing? [16:31] * avoine think every charm should use that [16:31] it's like 2:30 in the morning and you're working on charms? [16:31] here it's noon [16:32] oops… wrong channel :) [16:32] avoine: :) [16:32] avoine: There is talk of using aptdaemon to resolve this problem. Not sure if there are any snippets though [16:33] ok [16:33] I'll check it [16:35] s/is/was/ [16:35] avoine, marcoceppi: so the latest python juju (and I believe juju-core) only allows serial hook execution in the same container [16:35] That might actually not be the right package, but there's supposedly a way to queue packages for installation so you don't collide with apt-get in parent [16:36] thus avoiding a principle and subordinate charm trying todo conflicting operations at the same time [16:36] jamespage: that's super helpful to know [16:37] marcoceppi, hazmat fixed that for us during the HA work on OpenStack as we hit that issue *alot* [16:37] (around January I think) [16:37] excellent. I should write more subordinate charms [16:43] that also made it into juju core (trunk atm afaicr) [16:45] hazmat: any idea when the ppa will be updated with a more recenty juju-core release? [16:46] marcoceppi, when there's a new release [16:46] marcoceppi, there are several ppas associated to juju-core [16:46] I thought core was up to 1.14 or something (where 1.10 is currently in the ppa) [16:46] 1.11 [16:46] is trunk [16:47] mramm2, when's the next core release? [16:47] I think next friday [16:48] would have been this friday but Dave is traveling [16:48] I also think we should move to earlier in the week in case there is something serious that requires a followup release -- don't want to have to do that over the weekend! [17:02] For those waiting for the Ubuntu On Air, we'll be starting in a few mins [17:09] The Ubuntu on air page isn't quite working for us, so you can follow along here: http://www.youtube.com/watch?v=yRcqSjOGweo&feature=youtu.be [17:12] If you have any questions please feel free to ping me with your questions! [17:55] Questions? [17:56] marcoceppi: how about talking a bit about implicit relations [17:57] and scope [17:58] paraglade: queued up! [18:00] marcoceppi: this might be something actually to cover when you get into subordinates [18:00] paraglade: probably which we'll probably spend a whole 'nother session on [18:00] :) [18:04] cool thanks! [18:04] m_3, marcoceppi thanks or the charm school. [18:04] m_3: Thanks for running us through charms! [18:04] arosales: sure [18:04] thanks peeps for tuning in! === defunctzombie_zz is now known as defunctzombie === defunctzombie is now known as defunctzombie_zz [20:44] m_3: Hey -- Could you review this one when you get a chance? https://code.launchpad.net/~davidpbritton/charms/precise/landscape-client/add-landscape-relation/+merge/161497 [20:59] Oh boy, where is charm.log in go-juju deployments? [21:00] /var/lib/juju/units//charm.log isn't quite there anymore :) [21:03] marcoceppi: look in /var/log [21:03] m_3: ah, /var/log/juju/unit-name.log perfect [21:03] that will make archiving logs in juju-core a hell of a lot easier [21:04] dpb1: I'll try today, but most likely Monday [21:04] m_3: monday is fine. [21:05] m_3: just a gentle reminder. :) [21:05] (now that sprint season is done for a while) [21:05] gotcha [21:05] actually at a conference next week [21:05] _then_ the season's over for a bit [21:15] I don't know if it's just me, and my expectations, but I feel like "the cloud" has been noticeably more responsive since using the juju-core version === wedgwood is now known as wedgwood_away [22:17] is this the right place to look for help with the openstack charms? [22:17] it's not -wrong-, anyway :) but it is late in the day.. [22:19] well, it's worth a shot [22:21] i'm trying to help someone i work with get this openstack deployment straightened out. he's used juju charms to deploy mysql/rabbitmq/keystone/glance/swift/etc in HA. but it seems to have left the default database for each of the openstack services set to sqlite [22:22] there's a relation in juju, so i'm not sure what's been done wrong [22:22] and there are too many moving parts for me to just dive in and fix it, having never set it up myself [22:24] using glance as an example, i read through the 'juju debug-log' as i removed and re-added the relation to mysql. glance says something to the effet of 'database settings changed, will try again' but it never does === Makyo is now known as Makyo|out