=== axino` is now known as axino === dpm is now known as dpm-afk === dpm-afk is now known as dpm [08:51] Hi, I have a question about accessing into mongodb shell when using Juju2.0 [08:52] is it appropriate to ask it here? [08:54] sure kina_ [08:54] its just a bit quiet this time of the day [08:54] whats up? [08:55] I have been trying to access by following this https://github.com/juju/juju/wiki/Login-into-MongoDB [08:56] but I get "auth failed" [08:56] you're trying to login to the controller mongo instance? [08:56] magicaltrout: yes that's right [08:56] never tried it [08:56] let me find a running juju instance somewhere and test [08:57] magicaltrout: thanks! [08:58] BTW, with v1.25, I could successfully access into mongodb [08:58] hmm [08:59] juju 1.25 wasn't controlled by mongodb was it? [09:00] it was. but different version i think [09:01] it was like mongodb v2.xx, and juju 2.0 uses mongodb 3.2 [09:04] fair enough [09:04] shows what i know [09:08] yeah i get the same kina_ [09:08] that page is certainly stale [09:09] if babbageclunk or someone is around they might know [09:09] thanks for checking it out! [09:09] or it was edited a few days ago [09:09] maybe i'm on an older beta [09:09] and thats trunk or something [09:09] hmm, not sure about that. [09:10] I'm now using the latest juju 2.0beta13 [09:11] with just installing the mongodb-clients from ubuntu, it will install v2.6, but the mongodb server is v3.2. [09:12] so I also have tried by installing mongodb-shell v3.2, and the output is now a little different but still no access... [09:16] hi kina_ - can you put up a pastebin of what you tried? (Obscure the password first though!) [09:19] I'm just bootstrapping so I can try following the instructions there and see if I can get in to the db. [09:23] babbageclunk: hi [09:23] babbageclunk: thanks, but i just got afk. so will paste it later. maybe two hours later or so... [09:23] kina_: ok cool, I'll be around. === dpm is now known as dpm-lunch [12:29] evilnickveitch: heya, I'm doing a bunch of AU fixes now for the questions [12:29] but I'm going to do /stable/ instead of /2.0/ for the URLs [12:29] jcastro, yeah, i was thinking about that [12:30] I guess it will be good enough until 3.0 comes out :) [12:32] evilnickveitch: if one is really old and doesn't have a lot of views, just downvote it to bury it, editing them will just cause them to get a bump in google [12:33] jcastro, good plan [12:33] if it's lxc related I' [12:33] m just going to vote to close and delete === dpm-lunch is now known as dpm [12:47] hmm, the current docs don't mention how to ask for a charm to get reviewed and promulagated [12:50] jcastro that policy page got rewritten not so long ago didn't it? [12:54] I can't find it at all [12:56] jcastro, no, I think you are right. this is the page it *used* to be part of: https://jujucharms.com/docs/2.0/authors-charm-store [12:57] yeah so basically, we don't have instructions or a review queue right now lol [12:57] :) [13:15] hi, I'm using juju2-beta7 (unchanged xenial repo) and a local lxd controller. It was working for some weeks, now suddenly a "juju status" returns "ERROR refreshing models: connecting with bootstrap config: unable to connect to API" - any hints how to debug are appreciated === mpavone1 is now known as mpavone [14:18] magicaltrout - can you link me @ your maven build plugin? [14:18] my what? [14:20] your maven build that you gave a presentation on last week [14:20] ah [14:20] https://github.com/buggtb/snappy-maven-plugin [14:20] read the readme for the many holes ;) === valeech_ is now known as valeech [15:32] babbageclunk: hi, sorry for the late response, but here is the pastebin [15:32] http://pastebin.ubuntu.com/20879705/ [15:41] kina_: Hi - it looks like those instructions are old, so we need to pass the CA cert in as well. I'm just trying to work out how to do that. [15:42] kina_: The cert is in the /var/lib/juju/agents/machine-0/agent.conf beside the password. [15:55] babbageclunk: i see. thanks! i will try too [16:00] kina_: Actually, you can just pass --sslAllowInvalidCertificates to avoid having to check the cert. [16:01] ah, now i got in [16:01] kina_: That's obviously bad if you're using it in prod, but probably ok for debugging/exploration. [16:02] babbageclunk: yeah, for api usage it is bad indeed. but i just wanted to know the method, just in case [16:08] @cory_fu, @kwmonroe: Here is the upstream PR for the kafka changes: https://github.com/apache/bigtop/pull/134 [17:48] anyone home [17:48] I've been trying for 2 weeks to get juju 2.0 and maas 2.0 working at intel labs [17:49] im about ready to throw up my hands [17:50] endomorphosis: there are a few caveats you must know considering the pre-release beta status to get everything working ... good job making it here! [17:50] we're in a proxied environment and its a pain in the butt [17:50] im not usually the type to ask for help, and I know that its a pre-release beta [17:51] endomorphosis: asking for help is a good thing! [17:52] in any case, I work at the big data solutions at intel, and I wanted to convert our racks to juju / maas / openstack [17:52] endomorphosis: thats great! [17:53] Internally we're working on some packages called "snap" a telemetry package, and "tap" an analyitics platform written in go [17:54] https://github.com/intelsdi-x/snap [17:54] endomorphosis: do you know exactly what is being blocked by your proxy? [17:55] https://github.com/trustedanalytics/platform-wiki-0.7/wiki/Overview%20of%20Trusted%20Analytics%20Platform [17:55] and I wanted to try to integrate snap into juju at some point in the future [17:55] endomorphosis: that will be the easy part [17:56] endomorphosis: how much experience do you have with openstack? [17:56] I have some experience in openstack, we've been using fuel. [17:56] dial tcp getsockopt: connection timed out [17:57] that and MAAS similarly has problems provisioning servers over a proxy (but thats another team) [17:57] I get this error when trying to deploy charms or bootstrap nodes [17:58] endomorphosis: are you setting the proxy params when bootstrapping juju? [17:58] even when it appears like the proxy settings are correct on the host [17:58] set in /etc/environment and /etc/profile [17:59] endomorphosis: have you looked into setting the juju bootstrap specific proxy config params? [17:59] in environments.yaml right? [18:00] endomorphosis: what version of juju are you using? [18:00] 2.0 beta 13 [18:00] ok, thats the issue [18:00] in juju-2.0, there is no more environments.yaml [18:01] endomorphosis: what cloud provider are you bootstrapping to? [18:01] I've tried both LXD and MAAS [18:01] juju bootstrap local lxd --upload-tools --config apt-mirror=http://th.archive.ubuntu.com/ubuntu --config apt-http-proxy=http://10.0.4.1:8000/ [18:01] endomorphosis: ok, great [18:02] endomorphosis: can you add `--debug` [18:02] to that command [18:02] and report back with the output [18:03] endomorphosis: you may also need to set http-proxy depending on your setup [18:07] correct me if I'm wrong but the process is add cloud -> add controller -> add machines -> add model -> deploy service [18:10] endomorphosis: for lxd you don't need to add it, its configuration are included by default [18:10] so, you can just run the command ^^ from stub, and it should stand right up [18:10] endomorphosis: assuming you have lxd configured [18:11] endomorphosis: can you `lxc launch ubuntu:16.04` ? [18:11] right now its at installing cloud-utils [18:12] endomorphosis: then the "bootstrap" process is almost complete [18:12] when I add machines using the proxy, do I also use the same flags as well I'm assuming [18:13] endomorphosis: the flags you provided at bootstrap should be inherited by your models you create on that controller [18:14] endomorphosis: so, once the bootstrap is complete, you should be able to `juju deploy ubuntu` [18:14] cory_fu, kwmonroe: did either of you do anything to the kafka branch in our bigtop repo? It looks like cory_fu merged my bind address changes to it three days ago, but they don't show up in the branch. [18:14] ... I also got some conflicts when doing a pull, which suggests that somebody may have edited the history :-/ [18:15] petevg: I squashed the kafka branch, as our upstream PRs are supposed to be. [18:15] The PR you created this morning needs to be squashed as well [18:15] cory_fu: Ah. You squashed it into the already squashed stuff. [18:16] cory_fu: I know that we squash things, I was just expecting to see my changes, squashed, rather than my changes, squashed together with kjackal's. [18:16] It makes sense now. [18:16] internally there are some people working on this SNAP system and TAP system, to assign VMs / containers to systems based on load and performance of those workloads. [18:17] it is the software defined infrastructure team at intel. [18:18] I was keen to try to glue juju and those together. [18:20] endomorphosis: great idea, I think a few of us have taken stabs at integrating other auto scaling frameworks into juju as well [18:21] endomorphosis: so this is a great place to ask if you have questions regarding that [18:22] quick question, is it possible to check when an specific review of a charm was pushed? [18:22] for instance, cs:trusty/nova-compute-26 [18:23] endomorphosis: I proposed "juju scale" to the team .... I'm not sure if it got much traction though ... possibly you could write the mailing list with your scaling integration project/idea ... this could really help with bringing that "auto scaling" functionality to Juju itself [18:23] endomorphosis: plus, it entirely backs up my idea/proposal :-) [18:24] the project is also written in go, if that helps any [18:24] I'll write the mailing list [18:24] endomorphosis: thats great, thanks! [18:24] I think the bootstrap hung [18:25] endomorphosis: depending on the speed/capacity of your machine, the last few steps can sometimes take a moment [18:27] kwmonroe: Was there anything else before promulgating the bigtop Kafka charm? petevg: What series does it support? [18:27] cory_fu: it doesn't specify a series ... which might be something that we want to fix. [18:28] Wait, it's already promulgated for trusty: https://jujucharms.com/kafka/ [18:29] This would just be to update with the new config option [18:29] Has this been tested on xenial? [18:29] cory_fu: yes. It's just the update of the config. [18:29] And no, I haven't tested on xenial. [18:32] I'm kind of tempted to make "series" an environment variable that the amulet tests respect. So instead of doing amulet.Deployment(series='trusty'), we do something like amulet.Deployment(series=os.envrion.get('charm_test_series', 'trusty')) [18:33] petevg: Not unreasonable. Some charms / bundles include a separate test for each series, just to confirm that, e.g., a change for xenial doesn't break something on trusty [18:33] But that does double the test load [18:34] Bleh. We don't have a zookeeper for xenial, so testing xenial isn't as simple as editing and re-running the tests. [18:35] petevg: The Bigtop ZK should be xenial (and only xenial) [18:37] cory_fu: Right. My checkout of bigtop zookeeper is undergoing surgery right now. Fixing it up and deploying and running tests now ... [18:37] petevg: I don't think xenial is a blocker for the kafka update [18:38] cory_fu: Cool. I will back off on worrying about it, then :-) [18:38] In any case, kafka is as done as I have permissions to make done for now. [18:52] petevg, kwmonroe: Ok, https://jujucharms.com/kafka/trusty/4 is pushed and published [18:56] lazypower: Since both Kevin and I are going to be out this week, can you work with petevg to review and publish the Bigtop Zookeeper charm to resolve https://bugs.launchpad.net/charms/+source/zookeeper/+bug/1603007 [18:56] Bug #1603007: split network environment [18:57] cory_fu - sure [18:57] petevg ping me when you're ready sir [18:57] cory_fu - just for clarity, this is like the extra endpoint bindings support that has landed in the openstack charms yes? [18:58] lazyPower: will do. I won't be right away -- Zookeeper and I are currently arguing about writing things to a config file. [18:58] * lazypower is stepping out to run a quick errand, will brb shortly [18:58] petevg, lazypower: Keep in mind that the new ZK charm should be xenial only. Also, promulgating it from bigdata-charmers will drop https://jujucharms.com/zookeeper/trusty/ and that will need to be re-pushed to cs:~bigdata-charmers/trusty/zookeeper [18:58] petevg - that sounds like typical devops fun :) [18:58] cory_fu - right, same issue we've had with other charms coming into the promoted namespace. ack [18:58] Thanks [18:58] s/promoted/promulgated/ [18:58] lazypower: yep. That it is. [18:59] ok. sure i'll lend a hand. just lmk [18:59] * lazypower poofs for a bit [18:59] petevg: Did we end up making the java relation optional somehow for the Bigtop ZK? [18:59] cory_fu: nope. [19:00] OK [19:00] I think that we discussed it, and you thought that it would set a bad precedent. It was a while ago, though. === cory_fu is now known as cory_fu-vac [19:33] petevg: what's the best zk charm to use for testing some spark-ha stuff? ~bd-dev or ~bd-charmers zookeeper? [19:34] or build something more recent? [19:34] @kwmonroe: probably just ~bd-charmers for now. [19:34] cool, thx petevg [19:35] np [19:45] 2016-07-25 19:43:50 ERROR cmd supercommand.go:458 storing charm for URL "cs:juju-gui-130": cannot retrieve charm "cs:juju-gui-130": cannot get archive: Get https://api.jujucharms.com/charmstore/v5/juju-gui-130/archive?channel=stable: dial tcp 162.213.33.122:443: getsockopt: connection timed out [19:50] does anyone know how to fix this === endo is now known as endomorphosis_ [20:05] does anyone know how to deal with this error message? [20:05] cmd supercommand.go:458 storing charm for URL "cs:juju-gui-130": cannot retrieve charm "cs:juju-gui-130": cannot get archive: Get https://api.jujucharms.com/charmstore/v5/juju-gui-130/archive?channel=stable: dial tcp 162.213.33.122:443: getsockopt: connection timed out [20:06] does anyone know how to deal with this error message? [20:06] cmd supercommand.go:458 storing charm for URL "cs:juju-gui-130": cannot retrieve charm "cs:juju-gui-130": cannot get archive: Get https://api.jujucharms.com/charmstore/v5/juju-gui-130/archive?channel=stable: dial tcp 162.213.33.122:443: getsockopt: connection timed out [20:07] i don't, but you have some routing issue from controller to the www [20:08] i don't, but you have some routing issue from controller to the www endomorphosis_ [20:09] i have proxy setup in the controller, so I dont understand why === natefinch is now known as natefinch-afk === valeech_ is now known as valeech === valeech_ is now known as valeech