| === thumper is now known as thumper-afk | ||
| === thumper-afk is now known as thumper | ||
| === Ursinha is now known as Ursinha-afk | ||
| === Ursinha-afk is now known as Ursinha | ||
| leotr | hello | 05:22 |
|---|---|---|
| leotr | i added one admin address in postgresql service configuration | 05:22 |
| leotr | can't understand why juju doesn't add corresponding line to pg_hba.conf? | 05:23 |
| leotr | and when i add it manually it by some reason restores pg_hba.conf file to original state | 05:23 |
| leotr | i use ubuntu 14-04 | 05:23 |
| leotr | juju-local | 05:23 |
| leotr | any ideas? | 05:23 |
| leotr | or maybe i could do something to add my lines to pg_hba.conf or somewhere else? | 05:26 |
| leotr | ok, found solution | 05:30 |
| leotr | entered ip address in CIDR form | 05:30 |
| leotr | juju is so juju | 05:30 |
| === CyberJacob|Away is now known as CyberJacob | ||
| === vladk|offline is now known as vladk | ||
| === CyberJacob is now known as CyberJacob|Away | ||
| === CyberJacob|Away is now known as CyberJacob | ||
| === vladk is now known as vladk|offline | ||
| === vladk|offline is now known as vladk | ||
| pmatulis | yep | 09:06 |
| === vladk is now known as vladk|offline | ||
| === vladk|offline is now known as vladk | ||
| === urulama is now known as uru-afk | ||
| === uru-afk is now known as urulama | ||
| === CyberJacob is now known as CyberJacob|Away | ||
| === vladk is now known as vladk|offline | ||
| === vladk|offline is now known as vladk | ||
| === vladk is now known as vladk|offline | ||
| === vladk|offline is now known as vladk | ||
| === psivaa is now known as psivaa-lunch | ||
| === Ursinha is now known as Ursinha-afk | ||
| === Ursinha-afk is now known as Ursinha | ||
| === psivaa-lunch is now known as psivaa | ||
| jcastro | asanjar, lazyPower, you guys wanna talk bundles? | 14:24 |
| asanjar | sure | 14:25 |
| jcastro | can I whip one up in jujucharms.com? | 14:26 |
| jcastro | I was thinking of just following your readme, then bundle that as a start? | 14:26 |
| lazyPower | jcastro: that'll work - mbruzek is working on getting hive2 promoted into trusty for the store | 14:27 |
| jcastro | asanjar, where would hive fit in? | 14:28 |
| asanjar | hive is used as a bigdata warehousing plaform with SQL like interface. It fits in on top of hdfs for data access and distribution while using mapreduce to process the data based on HiveQL commands. | 14:31 |
| asanjar | so in our solution - we will have Hive connected to HDFS and YARN | 14:32 |
| jcastro | I assume you'll add that to the trusty/hadoop readme as well? | 14:33 |
| asanjar | also we have elasticsearch-hadoop.jar included in HIVE jar path to communicate with ES cluster for data indexing | 14:33 |
| === vladk is now known as vladk|offline | ||
| jcastro | ok, I'll try to make it less than 9 units | 14:34 |
| jcastro | so it's orangeboxable | 14:34 |
| lazyPower | jcastro: which part particularly? the notation about Hive being the big data warehousing? | 14:34 |
| jcastro | ok, why don't I get a bundle up and running that is the scale out hadoop example with ES | 14:35 |
| jcastro | and then you guys tell me what to connect next? | 14:35 |
| === vladk|offline is now known as vladk | ||
| asanjar | lazyPower: I found this last night http://www.elasticsearch.org/guide/en/elasticsearch/hadoop/current/hive.html | 14:36 |
| asanjar | there are HIVE sample command for ES | 14:37 |
| lazyPower | interesting | 14:38 |
| asanjar | would you try them, that would be a good demo for jcastro | 14:38 |
| asanjar | something as simple as this: | 14:38 |
| asanjar | CREATE EXTERNAL TABLE artists (...) | 14:38 |
| asanjar | STORED BY 'org.elasticsearch.hadoop.hive.EsStorageHandler' | 14:38 |
| asanjar | TBLPROPERTIES('es.resource' = 'radio/artists', | 14:38 |
| asanjar | 'es.index.auto.create' = 'false') ; | 14:38 |
| asanjar | would show hive is talking to ES | 14:38 |
| asanjar | jcastro: would that be sufficient | 14:39 |
| lazyPower | Sure, give me a bit to wrap what i'm working on | 14:39 |
| lazyPower | jcastro: ship me your bundle when you get it mapped up and i'll work through this ES tutorial | 14:39 |
| lazyPower | in validation mode | 14:39 |
| jcastro | I'll push it in about 10 minutes | 14:39 |
| asanjar | if jcastro could just do this aspart of the demo: | 14:40 |
| asanjar | CREATE EXTERNAL TABLE artists ( | 14:41 |
| asanjar | id BIGINT, | 14:41 |
| asanjar | name STRING, | 14:41 |
| asanjar | links STRUCT<url:STRING, picture:STRING>) | 14:41 |
| asanjar | STORED BY 'org.elasticsearch.hadoop.hive.EsStorageHandler' | 14:41 |
| asanjar | TBLPROPERTIES('es.resource' = 'radio/artists'); | 14:41 |
| asanjar | -- insert data to Elasticsearch from another table called 'source' | 14:41 |
| asanjar | INSERT OVERWRITE TABLE artists | 14:41 |
| asanjar | SELECT NULL, s.name, named_struct('url', s.url, 'picture', s.picture) | 14:41 |
| asanjar | FROM source s; | 14:41 |
| jcastro | asanjar, to add elasticsearch | 14:42 |
| jcastro | the readme doesn't mention how to add it to the scale out configuration | 14:42 |
| === vladk is now known as vladk|offline | ||
| jcastro | http://i.imgur.com/xqjY5r0.png | 14:44 |
| jcastro | this is what I have so far | 14:44 |
| jcastro | asanjar, ^^ what should I connect elasticsearch to? | 14:50 |
| lazyPower | jcastro: looks like we have a path for hive. export your bundle and ship it to me so i can deploy it | 14:51 |
| lazyPower | i'll start hacking on this ES demo | 14:51 |
| jcastro | ok on it | 14:52 |
| jcastro | lazyPower, https://code.launchpad.net/~jorge/charms/bundles/oscondemo/bundle | 14:59 |
| lazyPower | jcastro: got it, give me a few and i'll ping back with results. | 15:04 |
| asanjar | lazyPower: I am in a meeting, will join u as soon as I am done | 15:19 |
| lazyPower | asanjar: no rush, i'm just getting started | 15:20 |
| lazyPower | weird | 15:28 |
| lazyPower | jcastro: drag/drop of the bundle yielded deployment errors - only ES came through | 15:28 |
| jcastro | I only just pushed it | 15:28 |
| jcastro | I haven't tested it | 15:28 |
| lazyPower | ack - just a heads up | 15:28 |
| jcastro | Got pulled into a call | 15:28 |
| lazyPower | i'll get it working and push a patch once its deploying proper | 15:28 |
| jcastro | what errors did you get? | 15:29 |
| lazyPower | the gui didnt say, just errors on bundle | 15:29 |
| jcastro | I was just following the readme | 15:29 |
| jcastro | oh, use deployer man. :) | 15:29 |
| hazmat | :-) | 15:30 |
| lazyPower | :P | 15:30 |
| * lazyPower pushes the easy button | 15:30 | |
| lazyPower | same status result | 15:31 |
| jcastro | pastebin please | 15:31 |
| lazyPower | An error occurred while deploying the bundle: no further details can be provided about a minute ago | 15:31 |
| lazyPower | thats all i get. i'm going to nuke the relations and see if its the service definition | 15:32 |
| jcastro | oh I also forgot to rename envExport to something useful | 15:33 |
| lazyPower | yeah i took care of that | 15:34 |
| === vladk|offline is now known as vladk | ||
| asanjar | lazyPower: go back to "eco" | 17:01 |
| === urulama is now known as uru-afk | ||
| jcastro | wow | 17:12 |
| jcastro | lazyPower, my deployer straight up crashes on that bundle | 17:12 |
| lazyPower | jcastro: i'm rebuilding it by hand | 17:13 |
| lazyPower | this deployment configuration specified is also not one you'll see very commonly | 17:13 |
| lazyPower | we're going to hand off a ready made deployment W/ elasticsearch and hive | 17:14 |
| lazyPower | i'll post a bzr branch shortly for you to consume | 17:14 |
| jcastro | ack, I'm going to lunch | 17:16 |
| === ev_ is now known as ev | ||
| l1fe | anyone experiencing 409 conflicts with MAAS+JUJU when adding machines? | 18:04 |
| === scuttle` is now known as scuttle|afk | ||
| marcoceppi | l1fe: that usually means no MAAS nodes match your constraints (or you're out of available nodes) | 18:14 |
| l1fe | right, in this case, all nodes are ready | 18:18 |
| l1fe | (well aside from the bootstrapped node) | 18:18 |
| marcoceppi | l1fe: what architecture are they? | 18:18 |
| l1fe | amd64 and trusty | 18:18 |
| l1fe | the environment actually used to work | 18:18 |
| l1fe | and then after one too many destroy-environments | 18:18 |
| marcoceppi | l1fe: run with --debug see what juju is saying maas is complaining about | 18:18 |
| lazyPower | jcastro: http://i.imgur.com/DZfKZWb.png | 18:19 |
| marcoceppi | 409 is pretty much the MAAS API saying "I heard you, but I don't really feel like doing this because: ..." | 18:19 |
| lazyPower | i can ship you the bundle as is - we're still validating | 18:19 |
| l1fe | it's pretty generic | 18:19 |
| l1fe | 2014-07-11 16:58:06 ERROR juju.provisioner provisioner_task.go:421 cannot start instance for machine "1": cannot run instances: gomaasapi: got error back from server: 409 CONFLICT (No matching node is available.) | 18:19 |
| === roadmr is now known as roadmr_afk | ||
| l1fe | i've even reprovisioned the entire cluster | 18:20 |
| marcoceppi | l1fe: yeah, so maas says you don't have any nods available. What size ram are these? | 18:20 |
| l1fe | re-installed maas and juju | 18:20 |
| l1fe | all 16gb | 18:20 |
| marcoceppi | l1fe: are you on 1.20 ? | 18:20 |
| l1fe | 1.20.1 | 18:20 |
| * marcoceppi scratches head | 18:20 | |
| l1fe | and it literally was working 24 hours ago | 18:20 |
| l1fe | until after the last fateful destroy environment | 18:21 |
| marcoceppi | it's always that last destroy ;) | 18:21 |
| l1fe | :) | 18:21 |
| marcoceppi | And you've tried recommissioniong the nodes? | 18:21 |
| l1fe | yup | 18:21 |
| l1fe | full recommission | 18:21 |
| l1fe | reason why i even destroyed my environment | 18:21 |
| marcoceppi | it's weird | 18:21 |
| l1fe | was after i reported https://bugs.launchpad.net/bugs/1340261 | 18:22 |
| _mup_ | Bug #1340261: juju add-machine lxc:0 fails to start due to incorrect network name in trusty config <lxc> <network> <placement> <juju-core:Triaged> <https://launchpad.net/bugs/1340261> | 18:22 |
| marcoceppi | that it bootstraps | 18:22 |
| lazyPower | jcastro: lp:~lazypower/charms/bundles/oscondemo/bundle/ | 18:22 |
| l1fe | i wanted to re-setup my openstack environment with lxc...and figured i'd just start from scratch | 18:22 |
| l1fe | yeup, bootstraps...and if i bootstrap, it can even pick up random nodes | 18:22 |
| l1fe | that it would otherwise NOT pick up during add-machine | 18:22 |
| marcoceppi | l1fe: what does juju get-contstraints say? | 18:23 |
| l1fe | oh jesus | 18:23 |
| l1fe | you're kidding me | 18:23 |
| l1fe | as;dlkfjas;dlfkjlasdf | 18:23 |
| l1fe | thanks marcoceppi :) | 18:23 |
| marcoceppi | <3 | 18:24 |
| marcoceppi | l1fe: was it set to 32 bit? | 18:24 |
| l1fe | when i was bootstrapping, i was setting a constraint for 32G because i wanted to install on a particular maas node | 18:24 |
| marcoceppi | ah, yeah | 18:24 |
| marcoceppi | that sets it globally | 18:24 |
| l1fe | that's good to know | 18:24 |
| l1fe | omg | 18:24 |
| l1fe | i did a full reprovision twice | 18:24 |
| l1fe | thanks | 18:24 |
| marcoceppi | l1fe: no problem, you can use juju set-contstraints to put the mem limit lower | 18:25 |
| l1fe | yup | 18:25 |
| l1fe | i just did juju set constraints "" | 18:26 |
| l1fe | to clear it | 18:26 |
| marcoceppi | cool | 18:26 |
| l1fe | oh my god, i thought i was going insane | 18:26 |
| l1fe | i went down to 1.18 | 18:26 |
| l1fe | and 1.19 in devel | 18:26 |
| l1fe | and was like...this was working!! | 18:26 |
| jrwren | lazyPower: https://bugs.launchpad.net/charms/+source/mysql/+bug/965094 | 18:39 |
| _mup_ | Bug #965094: Provide config option to specify database encoding <mysql (Juju Charms Collection):Confirmed> <https://launchpad.net/bugs/965094> | 18:39 |
| lazyPower | jrwren: thanks for filing the bug. The charm maintainer will see it shortly. | 18:40 |
| jrwren | lazyPower: i want to fix it. i wanted to ask you about possible ways to fix. | 18:42 |
| marcoceppi | jrwren: the "problem" with setting it in the interface, is that redefines the interface | 19:15 |
| marcoceppi | jrwren: it' a good idea, but the new key will have to be 100% optional | 19:15 |
| marcoceppi | jrwren: so, you'd want to have a default-encoding configuration option, which is used when no encoding is sent via the relation | 19:15 |
| jrwren | agreed. going for 100% optional | 19:16 |
| jrwren | exactly what I was thinking. | 19:16 |
| jrwren | i spiked it and tested it out and it works, i'm just charm new, and wanted to ask before i submit a merge request | 19:17 |
| lazyPower | jrwren: :) we're really open to code reviews as MP's | 19:26 |
| lazyPower | it makes life great when its of good quality and we just merge it in | 19:26 |
| lazyPower | if not, its a source to track the conversation as the merge evolves | 19:26 |
| jrwren | lazyPower: ok, thanks. i'll get something up | 19:27 |
| l1fe | quick question regarding juju and lxc containers...does juju provide a unified network fabric for any services exposed over LXC? as in something on LXC:0 being able to communicate with LXC:1 | 19:32 |
| marcoceppi | l1fe: yes, something in LXC:0 will be able to talk to LXC:1 | 19:33 |
| l1fe | hmm | 19:33 |
| l1fe | so, i have a mysql lxc on 1/lxc/1 | 19:33 |
| l1fe | and nova-cloud-controller on 0/lxc/1 | 19:34 |
| l1fe | and i get errors where it can't connect to mysql | 19:34 |
| l1fe | via the ip address | 19:34 |
| l1fe | which makes sense since those 10.0.* ip addresses are only for each machine | 19:34 |
| l1fe | physical machine | 19:35 |
| l1fe | unless i'm misunderstanding something | 19:35 |
| l1fe | (given the constraints thing earlier, that's entirely possible) | 19:36 |
| lazyPower | l1fe: are you sure teh mysql service has started? its using a bridged virtual ethernet adapter to do the communication in LXC | 19:37 |
| lazyPower | LXCBR0 is the default virtual device for lxc connections, and all lxc containers use 10.0.3.x ip addresses by default as well | 19:37 |
| l1fe | yeah, it's started | 19:37 |
| l1fe | yeah, i raised a bug about how lxcbr0 wasn't actually being used, so i even manually went in and changed that in containers directory | 19:39 |
| l1fe | for lxc.cfg | 19:39 |
| lazyPower | strange, works on every other setup. The only time i had to change the br interface was when i wanted to bridge it with an actual ethernet device | 19:41 |
| lazyPower | and even then i think i configured it wrong and i had soem races where the interface wouldn't come online during boot and I had to manually initialize it every time the server was reset. | 19:41 |
| === vladk is now known as vladk|offline | ||
| lazyPower | but thats a story for another time.... | 19:41 |
| === vladk|offline is now known as vladk | ||
| lazyPower | so if you can reach the MySQL node, and you've verified the server is up - whats the log output that's stating you cannot connec to MySQL? | 19:42 |
| l1fe | https://bugs.launchpad.net/bugs/1340261 | 19:42 |
| _mup_ | Bug #1340261: juju add-machine lxc:0 fails to start due to incorrect network name in trusty config <lxc> <network> <placement> <juju-core:Triaged> <https://launchpad.net/bugs/1340261> | 19:42 |
| l1fe | for the templates not being setup properly for trusty | 19:42 |
| lazyPower | ahhh ok so this is not a local deployment | 19:43 |
| lazyPower | this is lxc, in a maas node for density based deployments | 19:43 |
| lazyPower | ack. | 19:43 |
| l1fe | unit-nova-cloud-controller-0: 2014-07-11 19:12:44 ERROR juju.worker.uniter uniter.go:486 hook failed: exit status 1 unit-cinder-0: 2014-07-11 19:12:55 INFO shared-db-relation-changed 2014-07-11 19:12:55.604 18833 CRITICAL cinder [-] OperationalError: (OperationalError) (2003, "Can't connect to MySQL server on '10.0.3.34' (113)") None None | 19:44 |
| l1fe | yeah, not local | 19:44 |
| l1fe | sorry forgot to mention that | 19:44 |
| l1fe | these are different physical maas nodes | 19:44 |
| lazyPower | its ok, its my own brain going immediately to local when someone mentions LXC | 19:44 |
| l1fe | i'm just trying to test out the best ways to deploy my environment :) | 19:44 |
| l1fe | and experimenting with encapsulating services | 19:45 |
| l1fe | i guess bridging networks between lxc containers across different physical nodes would be something that's not possible right now | 19:46 |
| lazyPower | i know that the networking when using juju deploy --to lxc:# can be a bit troublesome when it comes to networking configuration | 19:47 |
| l1fe | ah | 19:47 |
| l1fe | i don't see any DHCP or DNS changes in MAAS (which is controlling all of that stuff right now) | 19:47 |
| === roadmr_afk is now known as roadmr | ||
| lazyPower | There was a topic on this not long ago | 19:48 |
| l1fe | and juju is using the actual IP address vs a fqdn | 19:48 |
| lazyPower | there's either a script or a subordinate charm that helps this... i forget how long ago the thread was - but its in the juju mailng list. | 19:48 |
| lazyPower | *mailing list | 19:48 |
| l1fe | ah, i'll have to do some searching | 19:49 |
| === vladk is now known as vladk|offline | ||
| === cmagina_ is now known as cmagina | ||
| === scuttle|afk is now known as scuttlemonkey | ||
| === CyberJacob|Away is now known as CyberJacob | ||
| === CyberJacob is now known as CyberJacob|Away | ||
Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!
