=== CyberJacob is now known as CyberJacob|Away === Kyle is now known as Guest19572 === Guest19572 is now known as Kyle === mwhudson is now known as zz_mwhudson [04:57] Hello. Anyone here? [05:00] ianous: kinda [05:00] but just EODing [05:01] there are others who are around though [05:01] just ask your questions :) [05:01] or say hi [05:01] At this hour...I'm suprised someone's actually here. [05:02] ianous: 6pm ? [05:03] It's 7 am here... [05:03] ianous: welcome to the globe [05:03] *gasps* [05:03] I know, right! [05:03] it's actually nice but I have no idea what timetables people are usually around here. [05:04] You know..peak hours and sorts [05:05] Anyone knows btw about using private openstacks with juju? I"m trying to get behind the logic with generate-image and public puckets and all. [05:05] And I feel kinda lost === CyberJacob|Away is now known as CyberJacob === rogpeppe2 is now known as rogpeppe [11:04] marcoceppi: Have you seen this error in amulet before? As if it's got the 2.x urllib module loaded... http://paste.ubuntu.com/6837138/ [11:08] noodles775: yes, this was an issue I think I patched in trunk [11:09] marcoceppi: cool, I'll run with trunk again. Thanks. [11:10] noodles775: yeah, sorry about that it'll be released with that patch this week [11:34] frankban: you around? [11:35] jamespage, i heard you mentioned that deployer is incompatible with 1.17 which isn't true afaics, what's that about? [11:35] hazmat, "state watcher was stopped" ? [11:36] hazmat, sometime during a deployment, the deployer exits [11:36] jamespage, that's fixed in juju-core trunk [11:36] with that error message [11:36] hazmat, ok - so is juju-deployer compatible with any released version of juju? [11:36] hazmat, http://pastebin.ubuntu.com/6837352/ [11:36] ok [11:36] mramm: yes I am [11:37] jamespage, well that issue can appear with 1.17 .. 1.17.1 has the fix though. [11:37] hazmat, ah - ok - nuding that through now [11:37] hazmat: ok [11:37] and that issue can appear with 1.16.5.. i can try for a work-around. [11:37] hazmat: I that's the thing that was causing the relations not to be added on some of our demos [11:38] and why we have been pushing forward to 1.17.1 this morning === gary_poster|away is now known as gary_poster [14:17] marcoceppi: is amulet breaking with include-base64 a problem you're aware of? We ran into this: https://code.launchpad.net/~doanac/ubuntu-ci-services-itself/amulet-unit-config-fix/+merge/203729 [14:18] ev: I didn't know there was a base64 include option in deployer [14:19] amulet has no idea about this feature, could you file a bug? I can probably get a fix out this week [14:31] marcoceppi: Having trunk amulet on the pythonpath is enough for running tests manually, but not for running via juju test? Do you know what I need there? http://paste.ubuntu.com/6838092/ [14:31] noodles775: yeah, PYTHONPATH isn't pushed in to the testing environment. I'll add that to the whitelist good catch [14:32] marcoceppi: Chris Johnston is on it now [14:35] noodles775: I've got a patch for that landing in charm-tools now, but I'll have to roll a release of charm-tools for the new test [14:35] proably won't land until later tonight [14:35] marcoceppi: Great - thanks. No rush here, I can keep running tests manually :) === gary_poster is now known as gary_poster|away === gary_poster|away is now known as gary_poster [15:02] marcoceppi: bug #1274142 === freeflying is now known as freeflying_away [15:05] cjohnston: ta! [15:11] hi is juju fully installed an a node if i can read : "cloud-init boot finished at Wed,...... +0000. Up 136.3 seconds" ? [15:19] plz help, i have now isntalled juju, do i have to deploy services on my maas-server or on the juju-node? [15:20] if i hit "juju deploy mysql" on the maas-server no answer is prompted from the terminal [15:20] on my node, i can see "cloud1 login:" [15:28] if i enter juju status it prints me nothing.... i get simply no terminal output... [15:28] does someone know what the problem is ? :( === Guest12754 is now known as wedgwood [15:37] tomixxx: what does the output of juju status look like, please put it in paste.ubuntu.com [15:38] there is nothing, only the black background after hitting "enter" [15:38] one of my nodes show me "cloud1 login:" but i dont know what username and passwort is need here... [15:39] do i have to enter "juju status" on the maas-server-terminal or on the node? [15:40] tomixxx: you enter juju status from where you ran juju bootstrap [15:40] marcoeppi: ok, that is the maas-server then [15:40] tomixxx: can you please run `juju version` then `juju status --debug` and add the outputs to paste.ubuntu.com? [15:41] marcoceppi: ok [15:43] http://pastebin.ubuntu.com/6838480 [15:48] tomixxx: did you run sudo bootstrap or just bootstrap? [15:48] sudo juju bootstrap* or just juju bootstrap* [15:48] only "juju bootstrap" [15:48] tomixxx: okay, good [15:49] tomixxx: can you `ssh cloud1.master` [15:49] err [15:50] tomixxx: can you `ssh ubuntu@cloud1.master` [15:50] it says "could not resolve hostname cloud.master: name or service not known" === mhall119_ is now known as mhall119 [15:57] tomixxx: do you have dns running on your maas node? [15:57] tomixxx: paste initctl list | grep maas [15:58] marcoceppi: pastebin.ubuntu.com/6838547 [15:59] marcoeppi: i have set the interface of the cluster-controller to be managed "dhcp and dns" [16:00] btw, i have replaced the host name of the nodes from the suggested one (i guess it was 10.0.0.100-00 or sth like that) to "cloud1", "cloud2" and so one [16:01] but this should not be a problem, i guess [16:01] tomixxx: this sounds like a networking/dns issue with maas. Can you ssh direclty in to the ip address? [16:02] you mean i should try "ssh ubuntu@10.0.0.100.master"? [16:02] no, 10.0.0.100, if that's the IP address it's allocated [16:02] kk [16:02] seems to work, it asks me "Are you sure you want to continue connecting (yes/no)?" [16:03] and in front it says "The authenticity of host '10.0.0.100' cant be established. ECDSA key fingerprint is xx:xx:.. [16:05] should i say "yes" ? [16:10] hey the queue doesn't look so bad today [16:12] hmm [16:12] when i open head in "resolv.conf.d" there is nothing in the file [16:13] then i have a file called "original" with search ... and nameserver ... entries [16:13] maybe, sometime in past, i have modified head for some reason? [16:13] could this cause the problem? [16:35] marcoceppi: are u still here? [16:36] tomixxx: yes, one second [16:36] k [16:50] marcoceppi: i have to go now. But i will be here tomorrow again. if u have some hint for me, you can send me a private message. would appreciate it! [18:27] hi all [18:27] so I deployed a charm from the charm store [18:28] and now I need to upgrade the charm from a local repository [18:28] how can I do that? [18:30] marcoceppi: thoughts? [18:31] roaksoax@pursue:~/test$ juju upgrade-charm --repository . local:hacluster [18:31] error: invalid service name "local:hacluster" [18:31] roaksoax: you can run --switch [18:32] juju upgrade-charm --switch --repository . local:hacluster [18:32] roaksoax: it's very hulk smashish in that it ignores revisions all together [18:36] marcoceppi: ok cool, thanks! [18:36] marcoceppi: roaksoax@pursue:~/test$ juju upgrade-charm --switch --repository . local:haclsuter [18:36] error: unrecognized args: ["local:haclsuter"] [18:37] marcoceppi: got it [18:38] roaksoax@pursue:~/test$ juju upgrade-charm --switch local:hacluster hacluster [18:50] hey lazyPower [18:50] I saw you touching owncloud [18:50] I think the install hook needs an update btw [18:50] last I checked the version was woefully out of date [18:50] yeah [18:51] its not providing a version identifier [18:51] Already in my notes, currently on hold while i review OpenMRS [18:51] ack [18:52] i may tag it, and after the audit branch it and give it some upgrade lovin [19:05] lazyPower, I have this thing where I wish all charms had version options in config [19:07] The picky part of that is if the charm installs from upstream, how do you validate the hash when someone chooses a newer version? [19:07] yeah [19:07] mbruzek and I are actually in a conversation about this as we speak [19:08] http://download.owncloud.com/download/repositories/xUbuntu_12.04/all/ [19:08] in this case you could config the enterprise version [19:09] the .org stuff has an opensuse build system repo, so there's kind of like 3 options [19:11] Hi jcastro I want more details on how you would implement a version option. [19:12] So the 1.9.7 version of OpenMRS is current, I got a review comment that I need to cryptographically verify the file I download [19:12] Any newer version we would not be able to verify. Is that OK? [19:13] yeah that's a good question [19:13] why not? [19:13] * jcastro sees what marco did in phpmyadmin [19:14] https://github.com/charms/phpmyadmin/blob/master/bin/parse_upstream [19:14] heh [19:14] his comment at the top of the file is priceless [19:15] marco++ [19:16] * mbruzek is checking if there is a hash to verify against [19:16] oh jeeze; preg_replace(.../e) [19:17] sarnold, was that to me? [19:17] nope, for marco's parser [19:17] line 10 [19:17] sarnold, the entire thing is glorious [19:17] mbruzek: it is my wailing, gnashing of teeth, and rending asunder my clothing [19:18] please tell me we don't ship that [19:18] lazyPower, wait until you start to find that there are plenty of upstreams that don't sign their releases at all, depressing. [19:18] sarnold, http://sideshowsito.com/hulk_smash_loki.gif [19:18] lazyPower: yes please :) [19:19] i feel like that line is about as useful as goto, in specific contexts, very acceptable, but comes with such a stigma that most developers choose to stray away from it. [19:19] and as much as it may be hated, its useful in this context [19:20] https://bugs.launchpad.net/charm-tools/+bug/1274255 [19:20] lazyPower: when the context appears to be "load random content from the internet and then proceed to execute it", it's hard for me to be too sympathetic [19:20] anyone have issues with this? [19:21] jcastro: doing that will block ingestion [19:21] so I was thinking maybe a good idea would be for config to pass along a URL to an upstream source? [19:21] jcastro: so any charm without a cateogry will not get updated until touched [19:21] rick_h_, what do you mean? [19:21] oh, as in, you use charm-tools as part of the ingestion process and that would break stuff [19:21] jcastro: if charm proof makes that an error, no charm without a category can get updated in the store. They'll go into an error state [19:22] rick_h_, would doing this post-audit make more sense then? [19:22] jcastro: yes, in order for a charm to get past ingestion it needs to pass proof without errors [19:22] jcastro: +1 [19:22] jcastro: it's a warning and should be part of getting 'recommended' but not required as an error imo [19:22] policy vs proof basically [19:23] rick_h_, I've updated the bug, feel free to add on [19:23] jcastro: cool [19:25] sarnold, ah, i was looking into attack vectors utilizing preg_replace, i see where you're coming from now [19:41] hazmat: does juju-deployer not support ;revno= in bzr branches? Doesn't seem to work === zz_mwhudson is now known as mwhudson [20:18] hey sinzui [20:18] are there backport-for-12.04 plans for the 1.18.x series? [20:20] jcastro, We make backports, but I don't know what the SRU policy will be. [20:21] jcastro, there might be blockers such as switching juju to juju-monogo and gccgo [20:21] ok I will ask around [20:23] ev re deployer revision support.. syntax is lp:bzr@revno [20:23] ev, also works for git repos === gary_poster is now known as gary_poster|away === mwhudson is now known as zz_mwhudson [21:44] hi === thumper is now known as thumper-afk [21:52] hazmat: star! Thanks! [21:56] fwereade: hi [21:56] perrito666, heyhey [21:56] you caught me online :p === zz_mwhudson is now known as mwhudson === Kyle is now known as Guest45437 === Guest45437 is now known as Kyle === freeflying_away is now known as freeflying === CyberJacob is now known as CyberJacob|Away