[00:00] ah, i see they list public-ipv4 in there as well - an oversight on my part, disregard. [00:03] lazyPower: i couldn't figure it out [00:03] do you hack the url like I did [00:03] or lookup public-ip in the returned value [00:03] it's about as clear and infomrative as the rest of the aws documentation [00:03] technically correct, practically uselless [00:04] hah, i'll +1 that sentiment [00:05] davecheney: none of the subresources are resolving [00:05] i'm getting 404's as well [00:08] davecheney: wait, nailed it [00:08] ip-10-188-28-170.us-west-1.compute.internalubuntu@ip-10-188-28-170:~$ curl http://169.254.169.254/latest/meta-data/public-ipv4 [00:08] 54.176.211.32 [00:09] themonk: ^ [00:09] themonk: can you try on your host please [00:53] lazyPower, davecheney , its working, Thanks :) [00:58] oh [00:59] it's working isn't the answer I was hoping for [00:59] that command wasn't supposed to fix anything [00:59] just give us more information about what had gone wrong [01:00] davecheney: aws vooodooo === CyberJacob|Away is now known as CyberJacob === CyberJacob is now known as CyberJacob|Away === fuzzy is now known as Ponyo === kadams54 is now known as kadams54-away === kadams54 is now known as kadams54-away === wwitzel31 is now known as wwitzel3 === CyberJacob|Away is now known as CyberJacob === liam_ is now known as Guest79610 === CyberJacob is now known as CyberJacob|Away === liam_ is now known as Guest73903 === rogpeppe1 is now known as rogpeppe === liam_ is now known as Guest35660 [12:07] Following instructions at http://www.ubuntu.com/download/cloud/install-ubuntu-openstack. Have installed a Utopic 14.10 server. There is no ppa:cloud-installer/testing for Utopic! Do I a) re-install Trusy server for MAAS or b) use the TRusty ppa? [12:07] with Utopic === kadams54 is now known as kadams54-away === kadams54-away is now known as kadams54 [13:08] jcastro: as promised, the bug is now there === kadams54 is now known as kadams54-away === liam_ is now known as Guest73986 [16:14] Question for the juju community. [16:15] I am working on zookeeper, there does not appear to have a symbolic link for config-changed, but the main zookeeper-common file has a "config-changed" command (so it can handle config-changed calls). How does juju call config-changed if there is no file? [16:17] mbruzek: if there's no symlink, the hook context is not called - as juju looks for "config-changed" in the hooks directory [16:17] its basically, immutable at that point. [16:18] lazyPower the automated test result would disagree with that. [16:18] http://reports.vapour.ws/charm-tests/charm-bundle-test-1415-results/charm/charm-testing-hp/2 [16:18] I am really confused, I would have thought if no config-changed link or file Juju would not call that. [16:19] But I see it failed! [16:19] mbruzek: without a file or symlink, it should just skip it. [16:19] Yes I agree lazyPower [16:19] I agree [16:20] error details: hook failed: "config-changed" [16:21] i see that but i think its a red herring, or a byproduct of something hinky going on in teh charm [16:21] Correct I am looking into it. [16:21] But I am confused at how config-changed can be called at all! [16:26] yeah [16:26] i pulled the charm local, kicked off bundletester and it ran fine [16:26] i think its a byproduct of something going on in CI [16:26] is it consistent? [16:27] lazyPower: precise charm? [16:27] mbruzek: trusty [16:27] let me pull precise [16:27] * mbruzek was unaware there is a trusty, the readme is crap [16:27] mbruzek: charm get cs:trusty/zookeeper works [16:27] mbruzek: http://manage.jujucharms.com/charms/trusty/zookeeper [16:28] 2014/10/31 Charles Butler Removes a zero byte config-changed hook - will need to be revisted to address immutable config [16:30] lazyPower: I am only working on this because it is failing automated tests and marcoceppi commented about that on my bug. === natefinch is now known as natefinch-vote [16:30] mbruzek: well it wasn't a file [16:30] someone did "touch config-changed" [16:30] ack [16:30] ln -s config-changed should resolve that [16:30] i was moving pretty quick [16:31] lazyPower: I am fixing this and will add that back in, bundletester fails on my machine because my lxc containers are out of date, so I think we need apt-get update [16:31] mbruzek: i do that every morning :) [16:31] in the charm [16:32] the bundletester fails the zookeeper on install hook [16:32] oh, i've just relegated to updating my containers manually because thumper hates us [16:47] mbruzek: someone updated zookeper already [16:47] marcoceppi: Yeah but the automated tests fail on my system today. [16:48] it was updated today [16:48] oh? [16:48] around 8:12am [16:48] hrmmm... [16:48] make sure your trusty branch is sync'd with this fix [16:49] Will do thanks, for telling me Marco [16:51] marcoceppi: it is still missing the config-changed symbolic link I think it needs that [16:51] testing my changes now. [16:51] https://code.launchpad.net/~robert-ayres/charms/precise/zookeeper/fix-symlink/+merge/239736 [16:52] marcoceppi: I *just* pulled trusty zookeeper and it does not have this fix [16:52] This MP is for precise [16:52] look at the MP [16:52] exactly [16:52] its a fix for precise [16:52] zookeeper doesn't exist for trusty yet [16:52] you submitted the branch for it [16:52] you need to copy this change over [16:52] should just be able to bzr merge that branch [16:53] I am going to add apt-get update because I think that is what is failing on my lxc with bundletester this morning. [18:02] Hi to all. Could i ask anybody here for help with http://askubuntu.com/questions/545656/different-nova-conf-on-each-compute-node-for-novnc-acccess ? === kadams54 is now known as kadams54-away === kadams54-away is now known as kadams54 === natefinch-vote is now known as natefinch [18:48] aisrael: about tracks and limesurvey, I'm gonna try and see if there are any equivalents available for install. have you already made progress on those but saw yourself blocked? just so I can make an MP against that branch [18:49] jose: Thanks. I haven't done anything else with those two other than file bugs. [18:49] aisrael: cool, wil take a look asap! :) === roadmr is now known as roadmr_afk [19:34] hey guys, I'm wondering if any of you is willing to do a demo of Juju on POWER8 at UOS? [19:38] mbruzek: ?^ [19:38] I am not going to be at UOS [19:39] sounds like randall's gig :) [19:40] it can actually be anything [19:40] rick_h_: i second that *touches his nose* [19:40] rick_h_: I have been doing all the testing on power [19:40] mbruzek: :) rock on [19:40] probably an announcement about "what's new on the GUI", showcasing everything you can do on the gui [19:40] rick_h_: ^ [19:43] tvansteenburgh: Hi there, any update on the charm testing? (I can manually test on ec2, verified one yesterday, but am still keen to see the full test with ec2, hp etc.) [19:47] noodles775: i got a little side-tracked hunting down other testing-related bugs. revisiting your branch is next on my list. can you link me the branch or MP again please? [19:47] mbruzek: UOS is virtual [19:48] Yeah well then I _could_ do a demo [19:48] noodles775: this one right? https://code.launchpad.net/~michael.nelson/charms/trusty/elasticsearch/use-new-charmhelpers/+merge/239935 [19:49] noodles775: or was it the firewall_optional branch? [19:50] tvansteenburgh: the firewall_optional branch would be great (it builds on the previous one to add another feature requested). Thanks. [19:50] noodles775: got it, thanks [19:56] aisrael: not sure if the tracks charm will be available for trusty, the PPA I was using to set up a newer version of Ruby hasn't got some of the packages for trusty [19:56] jose: That's what I was afraid of. [19:56] aisrael: not sure if the tracks charm will be available for trusty, the PPA I was using to set up a newer version of Ruby hasn't got some of the packages for trusty [19:57] will probably have to wait and see if they repackage it [19:58] now, for limesurvey, it appears there's a fix, but lemme double check === mjs0 is now known as menn0 === kadams54 is now known as kadams54-away [21:00] jrwren, you around? [21:02] yes [21:03] jrwren: question about the s/public-address/private-address/ change to mongodb [21:03] tvansteenburgh: yes [21:04] so this will broadcast mongodb's private-address to all related units [21:04] tvansteenburgh: I guess. [21:04] jrwren: love the confidence [21:04] tvansteenburgh: I am not certain what you mean by broadcast, hence my guessing. [21:04] well that wasn't really a question [21:05] it will relation-set hostname=private-address [21:05] yes. [21:05] ok [21:06] i suppose that address would always be reachable from other units in the deployment so it's okay [21:07] just wanted to double-check with you and think about it some more [21:07] tvansteenburgh: and since the other address is never reachable from other units in teh deployment, unless the port is exposed with expose, its better. [21:07] it actually broke a test, so i'm trying to confirm that the test can be changed, and it's not a legit regression [21:07] I've no idea why a public-address would ever be used for relating charms. [21:07] but I only know ec2 and azure. [21:08] it would be necessary for cross-cloud deployments, but i have no idea whether we'll ever do that [21:10] ping balloons [21:10] o/ [21:10] juju supports cross cloud deployment? [21:10] haha, no [21:10] jrwren: that would it be crazy [21:10] you had me very interested for a second :) [21:11] hehe [21:11] sebas5384: it would? why crazy? [21:11] is like multi universes [21:11] i was just giving you a hypothetical reason why you'd need to relation-set public-address instead of private [21:11] but yeah, i think the need exist [21:12] jrwren: thanks, i'll update the test [21:12] tvansteenburgh: thank you. [21:12] tvansteenburgh: for my future reference, was it an ansible test? [21:12] jrwren: no, amulet [21:13] tvansteenburgh: ha! I meant amulet. great, ty. I need to run those next time :) [21:15] hey sebas5384 o/ [21:15] jrwren: it does if you're using the manual provider [21:15] hey lazyPower ! [21:15] sebas5384: hows things my friend? [21:16] lazyPower: it might work,but does it make sense. [21:16] lazyPower: been good, I went to bolivia to attend a DrupalCamp [21:16] does unit-get("private-address") return the same as public-address if the cloud on which it is running doesn't split public/private addresses? [21:16] I showed Ansible and Juju, and how a charm works [21:17] jrwren: depends on what you're doing and how much you care about latency inbetween services [21:17] lazyPower: they love it [21:17] sebas5384: thats great to hear! [21:17] yeah [21:17] have you gents gotten un-blocked with your integration testingi story? i haven't looked @ the repo in a while [21:17] lazyPower: lol... and how the cloud services does network billing :p [21:17] Juju us getting known :) [21:18] sebas5384: i hear ya, i just got invited to give a talk at a pretty significant meetup here in pittsburgh over juju :) [21:19] nice! lazyPower o/ [21:19] sebas5384: really glad to hear you are having a good time spreading the word of juju though. Solid work. [21:19] lazyPower: do you know how are they doing with the vagrant workflow? [21:19] aisrael: Hey you've got some developments on that front. can you relay to sebas5384 what you've been hacking on? [21:20] yeah! and now I'm going to do that at the first Drupal Conference in latin america [21:20] jrwren: private-address returns the instances public address on manual provider [21:20] because it's just a machine(tm) [21:20] is going to be in Colombia Bogotá [21:21] lazyPower: sebas5384: Sure! We've been getting new functionality added to the vagrant images, things like juju-deployer and squid for caching, and setting up a cleaner workflow. [21:21] Once the squid work lands, I'll update the workflow docs. I'm pretty excited about how clean it's looking. [21:22] i cant wait to see what you've been up to [21:22] aisrael: you may be interested in this as well: http://vagrantmanager.com/ [21:23] omg, want [21:23] its free and available [21:25] marcoceppi: how does it obtain a the "public address" with manual provider? What if that manual machine is behind NAT overload? public-address is meaningless at that point, right? I guess its just address at that point since public/private are both meaningless. [21:26] jrwren: right, it's whatever ip/address you gave it when you ran "juju add-machine" [21:26] both public and private address are set to that [21:26] marcoceppi: cool! thanks. [21:27] lazyPower: works as advertised. Thanks! [21:28] aisrael: happy to help :) [21:28] aisrael: i'll ping sebas via email with that followup since he apparently lost connection. Thanks for the snappy reply at any rate. [21:29] lazyPower: Oh, thanks. I didn't notice. Feel free to tag me on the email. [21:31] aisrael: done, linked you two up. He's doing some interesting workload configuration with vagrant and i'm pretty sure he'll be reaching out. [21:33] sorry i'm back [21:34] thanks lazyPower! [21:34] good to know aisrael [21:34] wb sebas5384 - you've got mail if you missed it. [21:34] yeah I sow here hehe === kadams54-away is now known as kadams54 === kadams54 is now known as kadams54-away === roadmr_afk is now known as roadmr === 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