=== Spads_ is now known as Spads [10:46] Hey marcoceppi, how are you? All cool? I want to bug you about adding me to charmers group. I was told you hold the keys to paradise! [10:48] http://goo.gl/zZXHZO I think of it more like this [10:48] magicaltrout: oups! [12:39] I am connecting to mariadb from remote machine, I am getting error: ERROR 2005 (HY000): Unknown MySQL server host === dpm_ is now known as dpm [12:40] any configuration changes I need to do on Mariadb for connecting from remote machine [12:40] how you trying to connect Raj__ ? [12:41] don't forget the juju unit names don't exist in real world DNS [12:43] mysql -h db_host -u root -p mypassword [12:43] yeah but is db_host an ip, a juju unit, something else? [12:44] it looks like a standard dns issue more than anything juju specific [12:44] db_host is of mariadb [12:44] yeah i appreciate that. [12:44] okay if you do [12:44] ping db_host [12:45] what happens? [12:45] let me try that [12:47] ping is working to db_host [12:47] 64 bytes from 10.0.3.46: icmp_seq=1 ttl=64 time=0.102 ms 64 bytes from 10.0.3.46: icmp_seq=2 ttl=64 time=0.068 ms [12:48] weird [12:48] okay so if you do [12:48] mysql -h 10.0.3.46 -u root -p mypassword [12:48] what happens? [12:48] also if you need to "juju expose" [12:48] although the mysql error looks different to what i'd expect for a closed port [12:49] also standard stuff like, are you pinging from the same box you're trying to run the mysql client from? ;) [12:50] yes mysql client is installed [12:51] if I connect to container and execute :mysql -h 10.0.3.46 -u root -p mypassword it connects but same command from script has this error [12:52] in your script are you injecting some variables as the hostname or something? [12:52] something smells off [12:53] yes , variable captured host name [12:53] okay, I'd echo that or something to make 100% sure you're trying to connect to the correct place [12:53] or just hard code the ip to test [12:54] if you replace the var with 10.0.3.46 in the script [12:54] you'd know if your addressing was screwy or something [12:57] ok, I will try that, anything about error: ERROR 2005 (HY000): Unknown MySQL server host [12:58] that "Unknown MySQL server host" error is 99.9% of the time because it can't resolve the hostname you're passing on -h parameter [12:58] I can't think of a time when that wouldn't be the error [12:59] ok, I will try [13:38] has anyone had 2.0 bootstraps fail with "missing CloudCredential not valid"? seems as if that would mean creds are invalid but my creds look fine to me. === CyberJacob is now known as zz_CyberJacob [14:11] https://www.crowdsupply.com/lime-micro/limesdr bought some ubuntu snappy based SDR thing today [14:12] interesting bit of kit! [14:13] they're actually based in the same office building as us, i might go and annoy them [14:16] magicaltrout: hows things? when're you going to space? [14:16] lol [14:17] busy away admcleod, i'm trying to get Saiku 3.9 released, the guys working on Saiku 4, exit from Meteorite, negotiate contracts with NASA, work in various bits of NASA shit that are supposed to be at GA mid next month but way behind schedule...... [14:17] you know [14:17] the usual ;) [14:17] fire fighting like the best [14:18] as long as you're enjoying it [14:18] i'll enjoy it when i renew my golf membership and disappear for a few afternoons :P [14:18] is there sun there? [14:19] pissed it down yesterday [14:19] 32 here [14:19] 21 C and sunshine today [14:19] usual british weather [14:19] i expect itll be dark and stormy on the 23rd [14:19] lol [14:19] *fingers crossed i dont get kicked out of the eu* [14:20] it's gonna be shit if we leave [14:20] I might join you and get a bunch of passports [14:20] just in case ;) [14:20] problem is I don't want to agree with Dave [14:20] the big forheaded t*at [14:20] +e [14:21] be nice if dave swapped to the leave side [14:21] haha [14:21] not sure how much use my other passport is in terms of living in places. great for holidays... [14:22] Boris is amusing, but I'd hate to see him as PM [14:22] crazy git [14:22] he'd just smoke lots of dope [14:22] take some class A's [14:23] if him and Trump win, i'm using my NASA status to put them both on the ISS [14:26] just flag the aliens [14:26] "its time" [14:26] hehe [14:26] im thinking boris vs trump boxing match [14:26] trump would suck [14:30] you would get to see him without his wig though [14:30] that would be an awful sight [14:49] elastic-stack-devs: xenial-kibana -> https://jujucharms.com/u/jamesbeedy/kibana/6. I'm working on a two actions that will create and remove httpd user and password, as well as some tests but it should be g2g [15:02] running juju 2.0 beta 9. I have successfully bootstrap’d a maas environment. I would like run juju in ha. When I issue enable-ha command, it fails and says I need to create new controller machines using juju add-machines. Am I misunderstanding something? I thought the enable-ha command would try to spin up new machines on maas to act as juju controllers. [16:07] did the juju 2 status set command change? I think the answer is no but I want to double check [16:20] hi Kwmonroe, i want to discuss regarding IBM-IM issue [16:25] sure shruthima - i did see your most recent note about deploying ibm-im, but haven't tried juju attach yet.. [16:26] oh k .After checking in all ways i feel there is some issue in code if iam not wrong, because as resource is getting downloaded but it is not going to unzip part in the code [16:26] i'll try that now. unfortunately, that error isn't very helpful :( i suspect there's something wrong in the fixpack handler [16:27] shruthima: i'll also look for ways to be more verbose than simply "install-ibm-im-fixpack' returned non-zero exit status 1" [16:31] kwmonroe: ok thanks kevin === scuttlemonkey is now known as scuttle|afk === frankban is now known as frankban|afk === scuttle|afk is now known as scuttlemonkey [16:46] yo petevg, i had some beef with https://github.com/juju-solutions/bigtop/pull/15 and the status we're showing. lmk if you'd like more deats on what i think good status looks like.. i don't mind whipping up a PR if you're in a hurry to get this charm +1'd. [16:48] kwmonroe: thanks for the feedback. I'm not in a hurry to get stuff +1ed -- I just want to make sure that the charm looks like we want it to look when we submit it upstream :-) [16:52] kwmonroe : il be logging out from chat now,if any solution for IBM-IM issue, Could you please email us .. [16:54] kwmonroe: I update the status messages, and argued with you about the quorum messaging (with the non automated restarts, I think that the check gets to be too complicated -- I'd like to add it back after we add back the automated restarts, though). [16:55] yup shruthima - will do [16:55] thanks you kevin :) [16:56] hi lazypower .. is it possible for me to get apt..installed state in bash if i use options in layer.yaml [16:59] i have written some piece of code http://pastebin.ubuntu.com/17651116/ ... here i am installing mysql-client using options in layer.yaml [17:00] here is it possible for me to get the state apt.mysql-client.installed' ? [17:00] bdx: there's a layer for kibana, iir [17:01] Prabakaran: no, mysql-client, when defined as part of the layer options will always be installed before the layer runs [17:01] apt.mysql-client.installed is set from the apt layer, where you may need to dynamically install deps [17:02] so i can remove this state 'apt.mysql-client.installed' in this code .. because it will be installed before only [17:02] http://pastebin.ubuntu.com/17651116/ [17:04] marcoceppi [17:05] so i can remove this state 'apt.mysql-client.installed' in this code http://pastebin.ubuntu.com/17651116/ .. because it will be installed before only [17:06] correct Prabakaran [17:06] Thanks kevin for the confirmation [17:06] :) [17:10] np Prabakaran -- you've got a bigger problem in that pastebin though [17:10] Prabakaran: line 3 here: http://pastebin.ubuntu.com/17651116/. bash reactive doesn't support the concept of a 'mysql' object in the same way that python does [17:10] so you can't use a mysql object like you're doing on line 16 [17:10] i was typing this question to you.... [17:11] i just got this answer [17:11] :) [17:11] :) so instead, do something like: [17:11] my_host=`relation_call --state 'mysql.available' 'host'` [17:12] and then use $my_host in your mysqlshow -h command [17:12] similarly, my_port=`relation_call --state 'mysql.available' 'port', etc, etc [17:15] cool .. let me just replace those lines [17:18] i dont need to worry about ssh exchange between containers if i use mysql-client apt package right? [17:19] correct Prabakaran - db consumers can talk to the mysql host with the mysql-client utilty [17:19] Prabakaran: fyi, the settings you can get from the relation are defined here: http://bazaar.launchpad.net/~charmers/charms/trusty/mysql/trunk/view/head:/hooks/db-relation-joined#L81 [17:22] so host, database, user, and password will give you relevant info, but it doesn't look like 'port' is an option on the relation. looks like that is always hard coded to 3306. [17:26] now i got all these.. Thanks kevin .... you made me clear on this.. [17:34] marcoceppi: yea, I was just indicating that it supports xenial now :-) [17:36] big-data: aside from tests and the httpd auth bit, how is https://jujucharms.com/u/jamesbeedy/kibana/7 looking to you guys [17:39] I'm going to put some finishing touches on it and submit to the RQ [17:39] just wanted to get some prelim feedback [17:57] bdx: i haven't been paying much attention on the kibana front, so others may have better feedback.. but how close is yours to the ~containers version (https://jujucharms.com/u/containers/kibana/)? i'm assuming it's the xenial support.. [17:57] and if that's it, is the plan to merge those and have 1 maintained kibana? or is there some reason we might need two? [18:02] kwmonroe: mine is the reactive re-write [18:02] + xenial [18:04] oh snap bdx! that's how much i don't know about kibana :/ i didn't realize the ~containers version wasn't reactive. [18:06] well bdx, +1 on making it reactive, -2 for your readme skills. i like that the ~containers readme talks about deployment and dashboards. would you consider copying the relevant bits from their readme into yours? [18:17] kwmonroe: entirely, I haven't given any cycles to the readme yet thats for sure :-0 [18:17] but yea, totaly down for a swap out === scuttlemonkey is now known as scuttle|afk === scuttle|afk is now known as scuttlemonkey [18:18] there are a few new dashboard bits I might need to commandeer 4 sure [18:26] charmers: what is the prefered method of mitigating the AUTHORS file between layers ? [18:27] charmers: I feel like the AUTHORS file should be generated on charm build by pulling the maintainers out of the top layer? [18:27] bdx: we've talked about that [18:28] bdx: one way would be to generate a copyright file using the debian copyright format [18:29] marcoceppi: and just rm the authors file before pushing? [18:29] bdx: well, instead of authors, it'd be a copyright file [18:29] totally [18:29] but basically, yeah [18:31] it'd look something like this: https://github.com/juju/charm-tools/issues/136 [18:31] bdx: ^^ [18:31] feel free to pitch in on that issue [18:32] it's a bit stalled atm [18:36] I see ... hmmm, seeing as each AUTHOR is represented in their respective layer, can charm build exclude the AUTHORS file, and just have a flag to pull down and include the desired copyright to point at the maintainer from metadata.yaml? [18:37] possibly the AUTHORS parsing doesn't need to be as complicated as the actual merging of the layers ...? [18:38] bc it could just boil down to just not merging any AUTHORS and just generating the desired copyright [18:39] I'm probably way out in left field here, but it seems reasonable === zz_CyberJacob is now known as CyberJacob === redir is now known as redir_afk [20:02] tvansteenburgh: Do you have an eta on bundletester compatibility with beta 9 (in light of the push to clear the queue)? [20:03] aisrael: no. there is a chain of things to fix. jujuclient -> deployer -> amulet -> bundletester [20:03] aisrael: i'll probably get it all working just in time for beta10 to break it again [20:04] hehe /o\ [20:04] aisrael: strongly suggest you just use beta8 for now [20:04] aisrael: lets face it... the queue will never be clear! ;) [20:04] be honest with yourselves :P [20:05] tvansteenburgh: lol, beta 8 it is [20:05] magicaltrout: A boy can hope! [20:05] hehe [20:06] I wound up rolling back to beta7, because apt was having a hard time finding beta8. That may just be weak package manager fu on my part, though. [20:07] In any case ... yeah: going to wade into the queue with a slightly outdated beta. I've done worse things :-) [20:08] And I autoremoved recently, so I don't have a backup of beta 8 [20:19] i'm trying to bootstrap juju2 against an openstack cloud and getting errors, probably doing something wrong. Can someone have a look and see if it's something obvious? https://bugs.launchpad.net/juju-core/+bug/1594958 [20:19] Bug #1594958: Bootstrapping on OpenStack fails with juju 2 [20:28] jhobbs: i don't have much exp with openstack clouds, but in my ~/.local/share/juju/credentials.yaml, my default-region is set to $OS_REGION_NAME and my tenant-name is set to $OS_TENANT_NAME. is your OS_REGION_NAME 'RegionKVM' and is your OS_TENANT_NAME 'admin'? [20:29] and when i say "my X is set to $ENV", i mean it's set to the actual value, not the string "$OS_X_NAME" [20:31] kwmonroe: i used add-credential without using a credentials.yaml - the tenant is correct but it didn't ask for a region name there. The region name is set in my clouds.yaml when I added my cloud, that value is correct, but I have two regions, wonder if that could be related [20:31] default-region isn't an accepted value (at least for openstack) in credentials.yaml - it rejects that [20:34] sorry, only one region in this deployment, so that's not related [20:38] hmph, not sure if i can be any help, but "index file has no data for cloud {RegionKVM } not found" feels like there's some disconnect between the region and the endpoint. you sure it's port 5000 there? [20:41] yeah - it is; i tested it with keystone client. I think it's trying to find a set of images for that cloud in the list of images here http://cloud-images.ubuntu.com/releases/streams/v1/index.sjson [20:42] There are some docs on how to setup a private cloud here, but they haven't been updated for 2.0 https://jujucharms.com/docs/devel/howto-privatecloud === redir_afk is now known as redir [21:10] cory_fu: in bash reactive, is the right way to make an explicit @hook handler? [21:10] @hook 'upgrade-charm' [21:10] function check-fixpack(){ [21:11] Yep [21:11] Though I don't think "check-fixpack" is a valid identifier [21:11] kwmonroe: ^ [21:12] whatchu talkin bout willis? [21:13] the use of a - in a function name... i believe [21:13] i think you're all nuts [21:13] Hrm. Apparently it's fine for function names but not variables [21:14] hey - i didn't even know it was a thing.. apparently ksh/bash/zsh are cool, but ash/dash/csh are not. [21:14] underscores ftw [21:15] jesus f**cking christ... there's some idiots on the news tonight [21:15] its like Im watching Fox or CNN [21:16] magicaltrout: i think you mean "there *are* some idiots on the news tonight" [21:17] that might be true! :P [21:17] admcleod: i might ship all the vote leave's down your way [21:17] i hear the costa del sol is nice for british people looking for an alternative home [21:36] jhobbs: I put some heat on that bug for you [21:37] bdx: cool thanks [22:45] anastasiamac: I updated bug #1594958, sorry, must have posted the wrong pastebin [22:45] Bug #1594958: Bootstrapping on OpenStack fails with juju 2

[22:46] jhobbs: tyvm. looking :) [22:47] jhobbs: do u still have a complete log from this run? if it was at trace-level would be stunning \o/ [22:47] bdx: i got juju 1.25 to work against openstack - I had to specify both metadata-source to bootstrap (as a filesystem path) and image-metadata-url (as an http url) in environments.yaml [22:47] anastasiamac: how do i set it at trace level? [22:47] jhobbs: https://github.com/juju/juju/wiki/Juju-Logging [22:48] jhobbs: could you please also kindly update the bug to say that it works for 1.25? :) otherwise I'll mark it as invalid for 1.25 :D [22:51] anastasiamac: i updated the bug WRT to 1.25; Is there somewhere to find a log for this other than the juju bootstrap itself? [22:52] *the output of juju bootstrap [22:53] jhobbs: if u set the environment variable and re-run bootstrap command as u were (with debug-log on) i *think* i'll get enoguh info [22:53] jhobbs: at this stage, the command that i see u run does not have --metadata-source... [22:54] uhh [22:54] weird [22:54] anastasiamac: http://paste.ubuntu.com/17668083/ [22:54] anastasiamac: i did the set-envs there and it didn't seem to change the output at all [22:55] sorry, set-model-config in this case [22:56] jhobbs: yeah, i can see in this output and now in mine (bootstrapped) that --metadtaa-source is not registering... i'll investigate further \o/ [22:56] jhobbs: thank you for input and pain ;) [22:57] anastasiamac: np, thanks for your patience on the logs and for looking into it [22:57] jhobbs: \o/ [23:08] jhobbs: no WAY! nice!!!! [23:09] jhobbs: was that documented anywhere? [23:12] no [23:12] bdx: well kind of - the juju output from metadata generate-images says to do one or the other, but you need to do both to get it to work === CyberJacob is now known as zz_CyberJacob [23:35] jhobbs: sad, did you file a bug for that specifically? [23:38] in juju 2.0 is there a way to specify the default series and an admin-secret when bootstrapping a maas 2.0 environment? [23:40] bdx: yeah bug #1594977 [23:40] Bug #1594977: juju-1 bootstrap forgets about metadata-source argument