=== thumper is now known as thumper-sushi-ru === thumper-sushi-ru is now known as thumper-sushirun === thumper-sushirun is now known as thumper === Spads_ is now known as Spads === 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 === kadams54 is now known as kadams54-away === fuzzy is now known as Ponyo === CyberJacob|Away is now known as CyberJacob === CyberJacob is now known as CyberJacob|Away [09:17] I have banged out Python3 support for charm-helpers, https://code.launchpad.net/~stub/charm-helpers/py3/+merge/242460 [09:17] I've love a review ASAP since branches like this will grow conflicts with every landing === kadams54 is now known as kadams54-away [11:51] ahoi, what platforms are supported by juju? [13:55] stub: thanks for the py3 charmhelpers branch! gonna review that now [13:56] tvansteenburgh: ta. [13:57] lazypower_: hi how are you? [13:59] lazypower_: I've re-install everything using the guide published of Marco Ceppi (http://marcoceppi.com/2014/06/deploying-openstack-with-just-two-machines/). But the result is the always the same when I try to open the horizon I see just a white page!!! [14:04] someone can help me, please? [14:13] and from host machine a can ping that using either IP address or FQDN. === jcsackett_ is now known as jcsackett [14:22] hm, how does one use t2.micro instance-type on amazon? [14:24] ah. https://bugs.launchpad.net/juju-core/+bug/1336473 [14:24] Bug #1336473: Support new t2 instance types on AWS

[14:24] bummer [14:49] anyone can help me??? [14:58] darknet_: I'm about to follow my directions again to see if I can reproduce the issue [15:01] marcoceppi_: hi Marco, thanks a lot for your reply. I also send y a private message about that issue. [15:02] marcoceppi_: I've re-install everything and I've the same issue...a white page [15:06] greetings darknet_ [15:06] have you confirmed no further relationships are running in your environment? ran juju debug-log and saw no chatter amongst units reporting they are still doing anything? [15:32] can i get a quick pair of eyes on https://code.launchpad.net/~bigdata-dev/charms/trusty/hdp-hadoop/trunk/+merge/242414 ? I need to get this landed for CTS today [15:33] lazyPower: looking [15:33] thanks dpb1 [15:34] lazyPower: why are you removing usr/bin/env? [15:34] lazyPower: what kind of testing have you done? [15:34] i didnt catch that marcoceppi - but a quick push will fix that [15:35] dpb1: deployed on canonistack, and hpcloud leveraging the tests in the charm suite - HDFS DFSADMIN report yields a good cluster, and terragen/sort smoke tests [15:35] lazyPower: why do you have amir's branch in the bundle? wouldn't you want amulet to use the CWD charm? [15:37] marcoceppi: that was inhereted from amirs branch - when i ran the tests i pointed it at ~bigdata-dev [15:38] lazyPower: just drop all that in general [15:38] when i push this from bigdata-dev i'll update to use the cs resource - however in bundles whats the proper format for calling CWD? [15:38] and put charm: hdp-hadoop [15:38] marcoceppi: when i ran that it deployed cs:trusty/hdp-hadoop [15:38] amulet will see that the charm its testing is that and use the directory the test is launched from [15:38] how did you test? [15:38] with JUJU_REPOSITORY set, it didnt deploy that, it picked the CS resource over local [15:38] bundletester and executing the script directly [15:39] eg: tests/10_deploy [15:39] idontbelieveyou.gif one second [15:40] so, I don't know if bundletester sets JUJU_TEST_CHARM [15:40] but if you set that to hdp-hadoop [15:40] it will work [15:41] ok, so you want me to nuke the branch bits and export that env var and re-test? [15:41] wait [15:42] tvansteenburgh ^? [15:42] ? [15:43] * tvansteenburgh reads scrollback [15:45] if the the charm name matches the local dirname amulet will deploy the local charm [15:46] thats not consistent with what i saw lastnight but i'm running another test on it now [15:46] i know thats the case when the service is declared in the test itself - but the bundle was consistently picking cs:trusty/hdp-hadoop [15:47] is the bundle file being loaded by amulet [15:47] ? [15:48] or just picked up implicitly by bundletester [15:48] its being loaded by amulet [15:48] self.d.load(yaml.safe_load(bun)) [15:49] ok [15:49] i see there's an open file descriptor in the test :S might as well fix that while i'm in here [15:52] marcoceppi_: re-open the URL without /horizon I don't see neither the apache's page [15:53] darknet_: are you getting a no response or an empty reponse? [15:59] tvansteenburgh: is it because the service naem doesn't match the charm name? [15:59] as in the service is being aliased in the bundle file [16:00] marcoceppi_: I tried to ping the node from host (my PC) using both IP add. and FQDN and it works correctly.. [16:00] marcoceppi: i haven't seen the bundle file [16:01] is anything actually broken? i was waiting to hear the results of lazyPower's test run [16:01] tvansteenburgh: spinning up vagrant to run again - took me a minute to clean up and re-import the box. I'm in the middle of another test on canonistack so my env is a bit occupied [16:02] d.load() just calls d.add() for every service in the bundle, so it determine the charm-under-test in the same way [16:02] *should* [16:05] i'll know as soon as the vm is booted - i'll ping ya with the results tvansteenburgh [16:05] roger [16:09] lazyPower: I'm progressively submitting diff comments [16:13] awesome. feedback much appreciated [16:32] lazyPower: ok, I'm done, marked needs fixing [16:33] caught one bug, most of the others are just really egregious style errors [16:33] dpb1: solid - i'll get these banged out and pushed within the hour. Cheers for the review [16:40] np [17:01] marcoceppi tvansteenburgh whit: standup === psivaa-holiday is now known as psivaa [17:44] jose, hi [17:44] jose, can we PM now? [17:53] lazyPower, hi [17:53] Hello themonk [18:01] lazyPower, sorry for late replay [18:01] no worries themonk [18:02] tvansteenburgh: appears nothings broken [18:02] \o/ [18:02] tvansteenburgh: i did however export the env var implicitly [18:02] that may or may not have been the magic sauce [18:03] well yeah that'll work too [18:03] although should work w/o it [18:06] yeah i had a variable in there i didnt think to isolate [18:06] next time around. but i'd say its safe to assume it was PEBKAC === kadams54 is now known as kadams54-away [18:13] lazyPower, i have a script.py which is not idempotent, i am passing properties values to config-change using juju set, and because setup.py is not reconfigurable i am setting a lock so that it(setup.py) can not run second time, and mention it in readme, my question is will it be a blocker in charm review? [18:14] themonk: its quite dependent on the context of the immutable configuration - in some scenarios it is neccessary to prevent dataloss for example. [18:14] Can you give some context as to why the setting is immutable? [18:22] lazyPower, the context is it creates certs and insert it in jks and creates DS data which is immutable and many other things which i dont know all :) [18:24] lazyPower, may be it not complex to make it mutable, but right now its immutable. [18:24] themonk: that seems like a candidate for consideration - its going to be up to the reviewer that looks at the code [18:25] if there is a sane and viable way to make it mutable - we would perfer that be the case [18:25] themonk: i suggest you push it to a namespace branch and get a proper code review - otherwise its a bit too subjective to make a proper response. [18:26] lazyPower, hmm, agreed :) doing it, thanks [18:27] anytime themonk - good luck on the review [18:27] lazyPower, :) [18:32] You can now configure 'apt-mirror' in environments.yaml to specify the [18:32] mirror used by all the machines provisioned in the environment. [18:32] apt-mirror: http://my.archive.ubuntu.com === kadams54-away is now known as kadams54 [18:32] \o/ [18:32] my fave feature so far [18:38] dpb1: fixes are up - running tests now [19:43] jose, lazyPower, marcoceppi: let's fire it up in like 5 more minutes? [19:43] sounds good to me [19:43] Sounds like a plan === CyberJacob|Away is now known as CyberJacob [19:50] jose, ok let's do this! [19:50] yep, it's starting [19:51] jose, post the youtube link too for people who just want to listen [19:51] got it [20:01] jcastro: jose marcoceppi mbruzek1 lazyPower working on an initial readme for hte bugs for jujucharms.com, please feel free to suggest updates. https://github.com/CanonicalLtd/jujucharms.com === roadmr is now known as roadmr_afk [20:03] rick_h_: thanks! will take a look once we're done with the Charm School [20:03] jose: all good ty much [20:06] wwitzel3, around? [20:11] awe man, I'm missing charm school? [20:13] yeah [20:14] jose, youtube link pls [20:14] ubuntuonair.com [20:14] sec checking analytics [20:16] or youtube.com/watch?v=NjERFuBs2S8 [20:20] https://github.com/juju/plugins/ [20:22] jcastro: yepo [20:23] https://plus.google.com/hangouts/_/g4roaheffd55jqc2tofvieb42ya?hl=en [20:23] want to join our charm school to explain your new retry hotness? === roadmr_afk is now known as roadmr [20:30] https://github.com/nojhan/liquidprompt [20:32] cory_fu: wow [20:33] sarnold: you following along at home on the charm school? [20:34] lazyPower: no :/ [20:36] sarnold: cory_fu wrote a pretty intense juju plugin [20:36] juju dhx [20:36] Oh, I thought you were impressed by my dhx plugin. :) [20:37] wwitzel3: Please check out the middle / end of the Charm School video we just finished. I'd like to get your feedback on my dhx plugin [20:37] oh :) well, maybe I would be, let me go looking for it :) [20:39] yup, looks pretty cool :) === kadams54 is now known as kadams54-away [20:44] wwitzel3: https://www.youtube.com/watch?v=NjERFuBs2S8#t=1059 [20:44] cory_fu: thank you [20:45] jcastro: sorry, I didn't respond earlier, I'm neck deep in trying to get something reviewed and landed before I go on vacation for 2-weeks [20:50] wwitzel3: Oh, then the plugin review isn't urgent enough to block your vacation. :) Just was curious on your feedback vis what of those features might make sense in core down the road. === kadams54-away is now known as kadams54 [21:14] Someone can help me? I've reported a problem with Openstack that was deployed by Juju.. All relations between charms are ok (green) but if I try to open horizon, I receive an error page like that " This webpage is not available". I tried also to ping the vm node from my host using either its IP address or FQDN and it works!!. I followed this guide http://marcoceppi.com/2014/06/deploying-openstack-with-just-two-machines/. B.R. [21:14] darknet [21:19] I have an open bug on this, and I'm looking for ideas for work arounds. https://bugs.launchpad.net/juju-core/+bug/1392786 [21:19] Bug #1392786: charm has no way to report error state

[21:20] what to do when a charm has an error, if I'm actively avoiding error states? [21:21] e.g. right now my charm takes a source PPA as config option. If you type that wrong then the isntall hook errors. I can catch the exception and let the hook complete without error. Question is, should I? [21:24] anyone can help me? [21:33] darknet_: Is the openstack-dashboard service exposed? (Does it say "exposed: true" in `juju status openstack-dashboard`?) [21:34] cory_fu_: yes all services are exposed [21:35] darknet_: And port 80 is open (listed under ports)? [21:36] cory_fu_: let me check [21:37] jrwren: There is a discussion underway for a feature to allow charms to report information, including more detailed statuses and messages, back to the user. The semantics haven't been finalized yet, though, so it will be a while before it's available. I'm definitely hoping it comes sooner rather than later. [21:40] jrwren: For the time being, how you handle it very much comes down to your specific needs. I would lean toward failing into error state if the user enters an invalid value, since there isn't really a way to inform them that their selection is being ignored [21:40] But, depending on your specific scenario, it may be acceptable to fall back to a default value [21:43] cory_fu_: i don't know what happened but not i see the openstack's dashboard....i don't have any idea why [21:44] E.g., something like an invalid PPA should probably be an error, because otherwise they'd be running the incorrect version / code without realizing it, but something like setting the domain to be used when generating URLs could be defaulted to a valid domain if you detect something completely wrong [21:44] cory_fu: thanks. I'm going to go ahead and fail. [21:45] darknet_: It's possible the machine failed to provision. Try looking through the full output of `juju status` to see if there are any errors that you missed === kadams54 is now known as kadams54-away [21:45] jrwren: Yeah, it's definitely not ideal, but we don't have a better option (yet!) [21:45] lol, jrwren qotw [21:46] lol [21:46] tvansteenburgh: tis my life motto. [21:46] awesome :) [21:48] cory_fu: this is a paste of juju status http://paste.ubuntu.com/9157090/ [21:49] cory_fu_: I don't see any error!!! [21:49] darknet_: is apache2 running on openstack-dashboard? [21:50] yes [21:51] darknet_: And listening properly on 80 & 443 (netstate -tnl)? If so, I guess check the Apache logs, unless marcoceppi has a better suggestion [21:52] * cory_fu tags out, because he just reached the end of his depth. [22:05] cory_fu_: after to run the command "juju set keystone admin-password="test01" and try the login openstack gave me "Something went wrong! An unexpected error has occurred. Try refreshing the page. If that doesn't help, contact your local administrator." [22:31] darknet_: Sorry, I stepped away. Honestly, I haven't done anything with that openstack deployment. You'd really need marcoceppi to help you, I'm afraid [22:31] cory_fu:_ anyway thanks a lot for your support i'll try to ask that to marco [22:32] Good luck. === CyberJacob is now known as CyberJacob|Away [23:07] cory_fu_: this is strange I've reboot all node and now I can open horizon, same error "This webpage is not available"....why?why? === CyberJacob|Away is now known as CyberJacob === CyberJacob is now known as CyberJacob|Away