siva | I am trying to deploy my charms and I find that the machine creation itself has not succeeded | 01:18 |
---|---|---|
siva | it is showing the 'pending' state | 01:18 |
siva | How do I debug this problem? | 01:18 |
siva | What can be the issue? | 01:19 |
siva | MAAS2.0 UI shows that the status of the machine is 'Deployed' | 01:19 |
siva | MACHINE STATE DNS INS-ID SERIES AZ 0 pending 192.168.1.252 4y3hkh trusty default | 01:20 |
siva | The above state is there for hours | 01:21 |
siva | Any help to debug this is much appreciated | 01:21 |
bildz | hey I'm having an issue with conjure-up on xenial. I have 5 machines allocated for MAAS (1) and Openstack (4). Maas works great and I can conjure-openstack, but im having an issue with it wanting another host and there isnt one for neutron. Do I need 5 openstack machines, instead of the 4 the previous one could work with? | 01:38 |
bildz | http://pastebin.com/ZfDciFgM | 01:42 |
stokachu | bildz: so 1 machine is always created for the 'controller' or 'juju admin' node | 01:42 |
stokachu | so if the bundle requires 5 machines that means you need at least 6 | 01:42 |
bildz | I'm assuming thats the bundle requirement | 01:43 |
stokachu | bildz: yea | 01:44 |
bildz | stokachu: where would i have read that? | 01:45 |
bildz | I can barely find install information for 2.0-rc1 | 01:46 |
stokachu | bildz: jujucharms.com | 01:46 |
stokachu | we use https://jujucharms.com/openstack-base/ for conjure-up | 01:47 |
stokachu | juju docs https://jujucharms.com/docs/stable/getting-started | 01:47 |
bildz | thank you | 01:47 |
bildz | thankfully, i have another to get me through the install, but i've been playing around with the cli commands and am getting familiar with managing the deployment | 01:48 |
stokachu | :) | 01:48 |
=== thumper is now known as thumper-dogwalk | ||
=== scuttlemonkey is now known as scuttle|afk | ||
=== thumper-dogwalk is now known as thumper-coffee | ||
=== thumper-coffee is now known as thumper | ||
=== saibarspeis is now known as saibarAuei | ||
=== saibarAuei is now known as saibarspeis | ||
=== saibarspeis is now known as saibarAuei | ||
=== aluria` is now known as aluria | ||
KpuCko | can somebody help me with container stuck in error/pending state? | 07:07 |
KpuCko | i have machine which hosting kvm container (this juju says) so i want to remove role from this server | 07:08 |
KpuCko | how to do that? how to fix services on the host machine which is in broken state? | 07:08 |
spaok | juju resolved, but if you want it retry the last hook use juju resolved -r then the name | 07:22 |
spaok | KpuCko: ^^ | 07:22 |
KpuCko | spaok http://pastebin.com/BBkFtA4D | 07:23 |
KpuCko | i don't know anythink for this machine: 11 | 07:23 |
KpuCko | how to recognize it? | 07:23 |
spaok | so, typically that happens when you run a add or deploy command it didn't match one of the other nodes | 07:25 |
spaok | it's trying to add a new one | 07:25 |
spaok | you can juju remove-machine 11 --force | 07:25 |
KpuCko | mhm, command runs fine, but result is the same | 07:26 |
KpuCko | oh sorry, i have to wait some time after im running juju status again | 07:27 |
spaok | ya | 07:27 |
spaok | its a queue like system | 07:27 |
KpuCko | thanks a lot | 07:27 |
spaok | np, that happens a lot :) | 07:27 |
spaok | if your testing stuff, using --to against one of your machine ID's will help | 07:28 |
KpuCko | yeah, yeah im working with juju gui | 07:29 |
KpuCko | but i'm trying to learn how to debug hooks | 07:30 |
KpuCko | how to fix them, etc | 07:30 |
KpuCko | many thanks for the help | 07:30 |
=== Guest60664 is now known as cyberjacob | ||
magicaltrout | alrighty, signed off to talk Juju at Pentaho Community event this year.... better sort out some more travel | 08:53 |
spaok | does anyone know a charm that uses the current python-basic template? trying to get a feel for writing a charm | 09:34 |
magicaltrout | there is a python basic template? | 09:38 |
magicaltrout | what you looking at spaok ? | 09:38 |
marosg | Hi, I am very new to Juju. I was following https://www.stgraber.org/2016/06/06/lxd-2-0-lxd-and-juju-1012/ and I am getting an error when bootstrapping controller | 09:39 |
marosg | test@lxd:~$ sudo apt install juju | 09:39 |
marosg | [sudo] password for test: | 09:39 |
marosg | Reading package lists... Done | 09:39 |
marosg | Building dependency tree | 09:39 |
marosg | Reading state information... Done | 09:39 |
marosg | The following additional packages will be installed: | 09:39 |
marosg | distro-info juju-2.0 | 09:39 |
spaok | magicaltrout: its part of the charm create command | 09:40 |
spaok | marosg: that should work, that's the error? | 09:41 |
spaok | This post assumes that you already have LXD 2.0 installed | 09:42 |
spaok | you got that? | 09:42 |
marosg | yes, LXD is working | 09:42 |
spaok | there's a bug with images, not sure if they fixed it | 09:42 |
spaok | might be that | 09:42 |
spaok | try sudo lxc image copy ubuntu:16.04 local: --alias ubuntu-xenial | 09:43 |
magicaltrout | ah just the default stuff spaok ? | 09:44 |
magicaltrout | most charms start with it | 09:44 |
magicaltrout | I've got some basic-ish stuff kicking around hold on | 09:44 |
spaok | ya, just reading about the services framework | 09:44 |
spaok | be nice to see a current example | 09:45 |
magicaltrout | https://github.com/buggtb/layer-mesos-master/blob/master/reactive/layer_mesos.py | 09:45 |
magicaltrout | thats a simple one | 09:45 |
spaok | cool, thanks | 09:45 |
magicaltrout | https://github.com/buggtb/layer-drillbit/blob/master/reactive/drillbit.py | 09:46 |
magicaltrout | there's a slightly more complex or messy one | 09:46 |
magicaltrout | depending how you look at it ;) | 09:46 |
marosg | spaok, thanks, looks like it did the trick | 09:48 |
spaok | marosg: np, I ran into that one also, I just pull the image now as ubuntu-xenial as part of my deployments | 09:48 |
marcoceppi | spaok: you need to update your version of charm | 10:10 |
marcoceppi | python-basic isn't the best template, python-reactive is | 10:10 |
spaok | ok | 10:11 |
marcoceppi | spaok: if you `sudo add-apt-repository ppa:juju/stable` and upgrade charm and charm-tools you'll get a better experience | 10:11 |
magicaltrout | ah that explains my confusion | 10:12 |
* magicaltrout didn't use python pre reactive | 10:12 | |
spaok | I looked at https://jujucharms.com/docs/stable/tools-charm-tools | 10:13 |
spaok | and http://pythonhosted.org/charmhelpers/getting-started.html | 10:13 |
magicaltrout | read 2.0 docs | 10:13 |
magicaltrout | they are | 10:13 |
magicaltrout | fail | 10:13 |
spaok | hah, ya, I've had a lot of problems with things in 2.0 docs | 10:14 |
marcoceppi | s/fail/in progress/ :P | 10:15 |
spaok | fair enough | 10:15 |
magicaltrout | sorry i was referring to my comment | 10:16 |
magicaltrout | not the docs | 10:16 |
magicaltrout | as I said read 2.0 without clicking the link as I read stable :) | 10:16 |
magicaltrout | said/read | 10:16 |
spaok | ya, the ones I was looking at don't really change between stable and dev | 10:16 |
magicaltrout | coffee clearly didn't have the desired effect this morning | 10:17 |
spaok | heh, its 3am here | 10:17 |
magicaltrout | yeah i don't think charm build has changed much if at all | 10:17 |
magicaltrout | west coaster | 10:17 |
spaok | yar | 10:17 |
magicaltrout | where abouts? | 10:17 |
spaok | san jose | 10:17 |
spaok | heart of silicon valley pretty much | 10:18 |
magicaltrout | ah very nice | 10:18 |
magicaltrout | took a drive up that way when I was out in Pasadena last week | 10:18 |
spaok | nice | 10:18 |
spaok | some nice beach roads | 10:19 |
magicaltrout | indeed | 10:19 |
magicaltrout | nice part of the world | 10:19 |
spaok | this current? | 10:20 |
spaok | charm 2.2.0-0ubuntu1~ubuntu16.04.1~ppa2 | 10:20 |
spaok | charm-tools 2.1.4 | 10:20 |
spaok | cause no template found for python-reactive | 10:21 |
magicaltrout | more current than mine | 10:21 |
magicaltrout | don't you just run "charm create"? | 10:21 |
spaok | wasn't sure, I saw different things about using -t python, or -t python-basic | 10:22 |
spaok | I can create it | 10:22 |
magicaltrout | na | 10:22 |
spaok | s/create/try/ | 10:22 |
magicaltrout | charm create mycharm | 10:22 |
magicaltrout | then i see | 10:22 |
magicaltrout | INFO: Using default charm template (reactive-python). To select a different template, use the -t option. | 10:22 |
magicaltrout | which is what you want | 10:22 |
spaok | kk, ya | 10:23 |
spaok | ok, now to figure out how to populate it | 10:23 |
magicaltrout | before you build it | 10:23 |
magicaltrout | make sure you have the environment vars set | 10:23 |
magicaltrout | else you'll end up build stuff wonky | 10:23 |
magicaltrout | JUJU_REPOSITORY etc | 10:24 |
spaok | for local repo? | 10:24 |
magicaltrout | yeah it'll stomp on your charm if you aren't careful | 10:24 |
spaok | ok | 10:24 |
magicaltrout | you have to build before deploying or pushing to the charmstore | 10:24 |
magicaltrout | so JUJU_REPOSITORY INFERFACE_PATH and LAYER_PATH | 10:25 |
magicaltrout | need setting | 10:25 |
spaok | cd | 10:25 |
spaok | ok, I'm going to mess with this some more in the day light, thanks for the help magicaltrout and marcoceppi, gives me a good starting place | 10:29 |
magicaltrout | no problem | 10:29 |
magicaltrout | swing back round in the afternoon and more clued up people are around | 10:29 |
MrDan | hello | 12:30 |
MrDan | if I shutdown physical mahcines on which i have juju units as LXDs, when I boot up the machines again, the LXDs should come up themselves, right? | 12:30 |
magicaltrout | I don't believe that is the case although I've never tried | 12:32 |
marcoceppi | MrDan: they should | 12:36 |
* marcoceppi powers off a node in maas to test | 12:38 | |
rick_h_ | MrDan: yes, they should come back up | 12:39 |
rick_h_ | MrDan: I do it on my laptop all the time, but not tried it on a maas server | 12:39 |
magicaltrout | booo i'm 0 for 2 today | 12:48 |
magicaltrout | someone else the other day said theirs weren't coming back up | 12:48 |
marcoceppi | magicaltrout: it could be a bug | 12:49 |
magicaltrout | surely not? | 12:49 |
marcoceppi | but the agent should bring back not just the workload on the machine, but also the containers (if any) | 12:49 |
rick_h_ | +1, definitely it should | 12:50 |
magicaltrout | fair enough | 12:50 |
magicaltrout | sounds sensible ;) | 12:50 |
magicaltrout | i'm still waiting for my kickstarter stuff to finally have a way to test MAAS properly | 12:51 |
magicaltrout | looking forward to it | 12:51 |
rick_h_ | kickstarter stuff? | 12:52 |
magicaltrout | http://www.udoo.org/udoo-x86/ | 12:52 |
magicaltrout | bought a bunch of these | 12:52 |
magicaltrout | altough they now have 32GB ram not 8 | 12:53 |
magicaltrout | if they arrive before i'm due to demo Juju at the Pentaho meetup this year I'm gonna take them and do a "MAAS" big data deployment | 12:54 |
rick_h_ | ah nice | 12:54 |
zeestrat | Cool stuff. Do they have some sort of management interface so they can be managed by MAAS? | 12:54 |
magicaltrout | you mean powerwise zeestrat ? | 12:55 |
zeestrat | magicaltrout: Yes. | 12:55 |
magicaltrout | dunno, good question | 12:56 |
MrDan | Hi, testing now beta18 for the neutron-gateway issues | 12:56 |
magicaltrout | the maas docs are a bit of a black hole | 12:57 |
rick_h_ | MrDan: k, the team's landing that fix today for the rc2 tomorrow as well so please watch for that to help you get back to rc | 12:57 |
magicaltrout | looks like there might be ipmi support zeestrat | 12:59 |
marcoceppi | magicaltrout zeestrat if there's IMPI it'll work for sure with maas | 13:00 |
zeestrat | magicaltrout: Then you are probably good to go. The MAAS docs aren't too bad. There's a list of power driver capabilities here: http://maas.io/docs/manage-power | 13:00 |
magicaltrout | yeah | 13:00 |
marcoceppi | magicaltrout: I know a guy who built a BMC with an arduino for raspberry pis. A lightweight restful api that would power on one of the GPIO pins to turn on or off an RPI | 13:00 |
magicaltrout | the problem with the docs is where they land on google | 13:00 |
magicaltrout | they don't seem to get indexed | 13:00 |
marcoceppi | magicaltrout: yeah, getting google juice has always been...a struggle for us on some of our properties | 13:01 |
marcoceppi | magicaltrout: I gave up googling and just go to the source, or use `site:jujucharms.com` or `site:maas.io` | 13:01 |
marcoceppi | maas.io is pretty new compared to maas.ubuntu.com | 13:02 |
magicaltrout | yeah but you google maas and you end up at maas.ubuntu.com but then maas.io is listed lower down but only the landing page | 13:02 |
magicaltrout | which is weird | 13:02 |
magicaltrout | anyway.... yeah these boards have an arduino built into them which is pretty crazy | 13:03 |
magicaltrout | interesting to see what you can do with an X86 board with an Arduino 101 on them | 13:03 |
MrDan | rick_h: it works, the neturon gateway services are installed | 13:03 |
rick_h_ | MrDan: cool, hope that unblocks you until rc2. | 13:04 |
MrDan | yep | 13:04 |
MrDan | tomorrow rc2 is out on ppa? | 13:04 |
rick_h_ | MrDan: yes, that's the plan | 13:04 |
MrDan | cool | 13:05 |
rock | Hi. We developed a "cinder storage driver" charm. In this charm we are using "subordinateconfigcontext" to pass configuration values to cinder.conf. [I mean to modify the cinder.conf charm]. we followed https://github.com/openstack/charm-cinder/blob/master/templates/mitaka/cinder.conf | 13:24 |
rick_h_ | marcoceppi: did we have the ability to list/etc plugins before? I don't see it in the command lists but curious how we missed this when doing the cli stuff | 13:26 |
marcoceppi | rick_h_: `juju-1.0 help plugins` | 13:26 |
marcoceppi | rick_h_: it was a help topic, but I don't see why it wouldn't be list-plugins today | 13:27 |
rock | Similary, Using the same charm we need to modify the nova.conf. We need to add use_multipath_option: true/false. | 13:28 |
rick_h_ | marcoceppi: gotcha, ok yea. I was trying to see why we didn't have any plugin based commands make it into the new cli planning | 13:28 |
rick_h_ | marcoceppi: but if it was under help we probably just missed it | 13:28 |
marcoceppi | rock: there is a way to do that without modifying the nova charm, but I don't remember what that is | 13:29 |
marcoceppi | rick_h_: I can build a `juju list-plugins` plugin, but it won't show up in help ;) | 13:29 |
rick_h_ | marcoceppi: lol yea | 13:29 |
marcoceppi | rick_h_: which brings up an interesting point | 13:29 |
marcoceppi | should `help commands` show plugins as well ;) | 13:29 |
rick_h_ | marcoceppi: yea, replying to your email thinking that through | 13:30 |
marcoceppi | rick_h_: cool, thanks | 13:30 |
rick_h_ | marcoceppi: if plugins is a noun in the juju-verse...what all does that mean. | 13:30 |
marcoceppi | you can list-plugins or `juju plugins` which are aliases | 13:30 |
rick_h_ | list, show, add/remove is done automatically via install. What's the command to get to a list of available plugins to install? | 13:30 |
marcoceppi | but I wonder about register-plugin | 13:30 |
rick_h_ | marcoceppi: that kind of stuff | 13:30 |
marcoceppi | where you whitelist the plugin so it shows up in commands | 13:30 |
rick_h_ | marcoceppi: hmm, not sure about that one | 13:31 |
marcoceppi | but that's a feature, listing plugins is parity | 13:31 |
marcoceppi | well, add-plugin, probably not register-plugin | 13:31 |
marcoceppi | fwiw, plugins still work | 13:31 |
rick_h_ | marcoceppi: yea, tbh as a whole this will have to fall into a 2.0.1 atm | 13:31 |
marcoceppi | pfft, PATCH RELEASE? | 13:31 |
rick_h_ | marcoceppi: exactly, we've got bugs of things that don't work that we need to get fixed for GA with it coming down to rc2 tomorrow and then a little gap for GA | 13:31 |
marcoceppi | jk, that sounds good | 13:31 |
rick_h_ | marcoceppi: so will file a bug and link it in my email reply | 13:32 |
marcoceppi | ta | 13:32 |
marcoceppi | rick_h_: I made sure to cc the -dev list ;) | 13:32 |
rock | marcoceppi: OK. I have a question. To modify nova.conf we are trying to use https://github.com/openstack/charm-nova-compute/blob/master/templates/mitaka/nova.conf. | 13:32 |
marcoceppi | rock: I imagine #openstack-charms would be a better room, a lot of good openstack charm knowledge there | 13:33 |
rock | marcoceppi: OK. Thank you. | 13:33 |
=== rye is now known as ryebot | ||
aisrael | tvansteenburgh: I know I asked this recently, but wrt python-jujuclient, can we get a new release with the 2.0 rc1 support? A PPA won't work, in this case, because it's being run from fedora | 14:35 |
tvansteenburgh | aisrael: i'll do that now | 14:36 |
aisrael | tvansteenburgh: thanks! | 14:36 |
tvansteenburgh | aisrael: uploaded | 14:41 |
tvansteenburgh | (v0.53.3) | 14:41 |
aisrael | excellent, thanks! | 14:41 |
kwmonroe | hey rick_h_, if i add a user to my controller, and grant that person (let's call him tvansteenburgh) 'write' privileges to a model, should he be able to 'juju ssh X' into units of that model? | 15:04 |
rick_h_ | kwmonroe: you'll need to add his ssh key | 15:05 |
rick_h_ | kwmonroe: see juju add-ssh-key and such | 15:05 |
rick_h_ | kwmonroe: since he's not admin he can't manage keys and without a key, no access | 15:05 |
kwmonroe | add-ssh-key?!?!? rick_h_, you're the best. i didn't know of such sorcery. | 15:06 |
rick_h_ | kwmonroe: let me know how it goes | 15:06 |
=== natefinch is now known as natefinch-afk | ||
pascalmazon | hi. Can I provide --resource foo=bar when deploying a bundle? how will it know what application it is for? | 15:23 |
rick_h_ | pascalmazon: sorry, bundles currently only support resources from the charmstore | 15:29 |
rick_h_ | pascalmazon: as you note, the bundle deployment scenario is a bit more complicated and was pushed past rev1 atm | 15:29 |
pascalmazon | rick_h: ok, thanks for the info! | 15:31 |
=== natefinch-afk is now known as natefinch | ||
kwmonroe | hey beisner, suchvenu emailed mbruzek and i with some questions about the cinder charm (stuff like cinder for Z and cinder in lxd). who is the best contact for those? or would it be better to ask on the juju ML? | 16:18 |
kwmonroe | (stand down beisner ^^; we got the routing sorted) | 16:30 |
beisner | woot thx kwmonroe | 17:33 |
=== dpm is now known as dpm-afk | ||
tvansteenburgh | rick_h_: is there any way to tell `juju ssh` which key to offer? | 18:10 |
rick_h_ | tvansteenburgh: hmm, with juju scp you can specify args to the underlying scp functionality with a --, but I think we redid juju ssh in a OS agnostic way. | 18:12 |
rick_h_ | tvansteenburgh: so honestly not sure, will have to ask around/look and see if I can see anything in the code | 18:13 |
tvansteenburgh | rick_h_: ok. right now i'm in a situation where i have to rename my keys to get juju ssh to work | 18:14 |
tvansteenburgh | rick_h_: should i file a bug about that | 18:15 |
tvansteenburgh | ? | 18:15 |
rick_h_ | tvansteenburgh: yes please. Will see if there's a flag that's just not in help/etc | 18:15 |
tvansteenburgh | rick_h_: roger, thanks | 18:15 |
rick_h_ | mgz: around? you were just in ssh-key land recently. any ideas on ^ | 18:16 |
rick_h_ | tvansteenburgh: looking at https://goo.gl/0q2KWq | 18:23 |
rick_h_ | tvansteenburgh: looks like not supported :( | 18:23 |
tvansteenburgh | rick_h_: ack, thanks for looking | 18:25 |
rick_h_ | bdx_: heads up, comment/etc coming in on https://bugs.launchpad.net/juju/+bug/1627554 | 18:28 |
mup | Bug #1627554: juju binary broken on sierra <juju:Triaged by jamesbeedy> <https://launchpad.net/bugs/1627554> | 18:28 |
rick_h_ | bdx_: let me know if you're up for seeing it through or want to punt. | 18:28 |
junaidali | Hi, i'm getting error 'ERROR unrecognized command:' for charm publish command. any idea? | 18:50 |
junaidali | charm * packages version: http://paste.ubuntu.com/23247897/ | 18:51 |
lazyPower | junaidali: the command changed to charm release | 18:59 |
junaidali | thanks lazyPower. I missed the latest update | 19:00 |
marcoceppi | lazyPower junaidali I'm adding a silent plugin to the next charm snap, where charm publish is just an alias to charm release | 19:12 |
beisner | wait what, is the next charm tools command for publishing changing? | 19:29 |
beisner | we've got automation on that, will need to know when that will affect stable users (deb) | 19:31 |
firl | lazyPower, congrats on getting kubernetes enterprise support ready! | 19:40 |
firl | I had a question around it when you get a moment | 19:41 |
lazyPower | firl: sure, i'm in/out but whats up? | 19:41 |
firl | I saw the kibana / es / filebeat configuration with kube, awesome! | 19:42 |
firl | is there a way to expose services yet like gce? | 19:42 |
lazyPower | firl: we ship with an nginx reverse proxy, there's a demo action of microbots | 19:44 |
lazyPower | it has limited support for sockets, i'm still working to make that easier | 19:44 |
firl | gotcha, I remember that being an issue | 19:44 |
lazyPower | i want to add an action that will let you pick from pre-configured ingress LB's and specify a namespace and it "just works" after that | 19:44 |
firl | Gotcha, yes that would be wonderful | 19:44 |
gQuigs | are there daily builds available for juju 1.25? | 19:45 |
lazyPower | so we're 1/2 way there with whats in there :) | 19:45 |
firl | yeah, sounds awesome | 19:45 |
lazyPower | gQuigs: not daily, 2.0 is what would be in daily | 19:45 |
lazyPower | firl: i'm highly interested in your return feedback this go around, please capture it for us | 19:45 |
firl | yeah, I will try it out in Openstack | 19:45 |
lazyPower | firl: if there's *anything* yinz need over there in the short term beta cycle, i want to get that captured. | 19:45 |
firl | I am considering if I need to try it on bare metal | 19:45 |
lazyPower | well your networking will be pokey | 19:45 |
gQuigs | lazyPower: right, but 1.25 is still maintained and I'd like to try a fix that's committed but not released | 19:46 |
lazyPower | overlay in an overlay and all that | 19:46 |
firl | do I need juju 2 for it? | 19:46 |
firl | yeah the docker config would need to have 1404 mtu | 19:46 |
lazyPower | firl: yeah, i def need that feedback of your network findings when deployed in openstack | 19:46 |
lazyPower | i'm probably going to have to expose some of the guts of the container rumtime for configuration there | 19:46 |
firl | cool. I have a few nodes to test it in | 19:46 |
firl | whats the repo / guide again to get it going and I will try to test it this week | 19:47 |
lazyPower | gQuigs: ah good point, try checking in #juju-dev | 19:47 |
lazyPower | firl: the readme - jujucharms.com/canonical-kubernetes | 19:47 |
gQuigs | will do, thanks | 19:47 |
firl | lazyPower perfect, I will try it with an openstack overlay. If it’s working enough I might put it on a couple bare metal machines. | 19:49 |
firl | does nodeport work through the juju overlay also? | 19:50 |
firl | or do i have to do a nginx reverse | 19:50 |
thumper | lazyPower: hey there | 19:52 |
lazyPower | if you want to do nodeport, you'll need to manually open the ports for now | 19:52 |
lazyPower | any worker can perform as a reverse proxy for both nodeport and for ingress | 19:52 |
lazyPower | thumper: yo | 19:52 |
lazyPower | firl: but our recommendation is to use the ingress controller so its encapsulated. for the workdloads that demand nodeport, like a znc bouncer, we'll have to brianstorm and make that better | 19:55 |
firl | im just thinking through if i wanted to have socket support | 19:55 |
firl | but this is all awsome, nice job! | 19:55 |
lazyPower | firl: https://github.com/kubernetes/contrib/tree/master/ingress/controllers/nginx | 19:56 |
lazyPower | its pretty extensive | 19:56 |
lazyPower | i mean you can throw kubelego as a pod on the ingress and get free tls with letsencrypt | 19:56 |
lazyPower | but thats the LB we shipped with for beta. it has the biggest portfolio of supported options without compromising support through advanced configuration. | 19:57 |
firl | yeah | 19:57 |
firl | currently I have an ssl secret with a single ssl proxy node that just acts as a termination point | 19:57 |
firl | just to get ssl so that might be a nice change of pace | 19:57 |
lazyPower | lots of options | 19:58 |
firl | ya, I will have to get a juju 2.0 environment working and test it out | 19:58 |
kwmonroe | lazyPower: how often do you refresh your boot2docker/docker-machine? my charmbox is kernel panic'ing more frequently these days, and i realize i haven't updated since 1.11.2 | 20:11 |
natefinch | lazyPower: is it me, or does the demo button here not work? https://jujucharms.com/canonical-kubernetes/ | 20:40 |
natefinch | lazyPower: I presume it's supposed to bring up the demo page with canonical-kubernetes deploying/deployed, right? for me it just brings up an empty demo window | 20:41 |
natefinch | lazyPower: it says "fetching bundle data" but then... nothing | 20:41 |
rick_h_ | natefinch: known issue, gui team is updating the gui to make it work | 20:43 |
natefinch | rick_h_: ahh, good to know. Sucks that it happened after the announcement | 20:47 |
rick_h_ | natefinch: yea | 20:48 |
=== natefinch is now known as natefinch-afk | ||
magicaltrout | Congratulations, and welcome to Apache: Big Data Europe! Your submission, "Highly Scalable Big Data Analytics with Apache Drill", has been accepted | 21:52 |
magicaltrout | Seville here I come | 21:52 |
magicaltrout | kwmonroe we have some interesting shit to work on | 21:53 |
magicaltrout | sorry stuff | 21:53 |
magicaltrout | not allowed to swear | 21:53 |
kwmonroe | right on magicaltrout! i've always wanted to do Highly Scalable Big Data Analytics with Apache Drill ;) | 21:56 |
magicaltrout | meh | 22:01 |
magicaltrout | i detect irony | 22:01 |
magicaltrout | well I can crack out the Big Top Stuff, Drill Stuff | 22:01 |
magicaltrout | at ApacheCon | 22:01 |
magicaltrout | I suspect it'll go down reasonably well and give me an excuse to have that upstream discussion | 22:02 |
kwmonroe | magicaltrout: do you know if there's any work behind https://issues.apache.org/jira/browse/BIGTOP-2001? it hasn't been updated in forever :/ | 22:05 |
kwmonroe | and from the drill side, even more foreverer: https://issues.apache.org/jira/browse/DRILL-114 | 22:06 |
magicaltrout | they both seem like odd requests | 22:06 |
magicaltrout | which is why they're probably both open | 22:07 |
kwmonroe | i'm guessing they're requests for bigtop to take over drill builds (or at least pin build versions for a particular bigtop release) | 22:07 |
magicaltrout | hmm, yeah but you can run Drill over anything in HDFS, Hive or HBase | 22:08 |
magicaltrout | its not vendor specific | 22:08 |
kwmonroe | at the summit, we talked about a hackathon with bigtop (c0s, specifically). when that happens, i'll see if there is any movement in bringing drill under the bigtop umbrella. | 22:10 |
magicaltrout | I don't see that happening | 22:13 |
magicaltrout | MapR have their claws in Drill | 22:13 |
magicaltrout | I also don't see the benifit but thats neither here nor there | 22:14 |
magicaltrout | especially as drill queries a bunch of non hadoop stuff | 22:14 |
spaok | magicaltrout: is there a list of layers or interfaces you can use? | 22:50 |
magicaltrout | indeed spaok | 22:51 |
magicaltrout | interfaces.juju.solutions | 22:51 |
spaok | interesting | 22:52 |
spaok | thanks | 22:52 |
spaok | magicaltrout: sorry to bug ya, just wondering, if I want to write an interface, do I just make something in charms/interfaces? not sure the difference between charms/interfaces and charms/deps/interface | 22:56 |
magicaltrout | no worries spaok i'm just trying to figure out where 150GB of non backedup science data has gone :) | 23:00 |
magicaltrout | yeah you just put it into charms/interfaces | 23:00 |
magicaltrout | no magic required there | 23:00 |
lazyPower | spaok: https://jujucharms.com/docs/stable/developer-layers-interfaces | 23:06 |
spaok | lazyPower: thanks | 23:07 |
spaok | https://jujucharms.com/docs/devel/developer-layer-example | 23:08 |
spaok | had a link to https://jujucharms.com/docs/devel/charms-layers-interfaces/ | 23:08 |
spaok | which is 4040 | 23:08 |
spaok | s/4040/404/ | 23:08 |
lazyPower | spaok: thanks for reporting that https://github.com/juju/docs/pull/1411 | 23:15 |
Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!