=== CyberJacob is now known as zz_CyberJacob
=== redir is now known as redir_afk
ryebotI keep getting "ERROR invalid entity name or password" when I try to run `juju debug-log`, not sure how to get rid of it - any tips?02:47
ryebotI've tried a juju logout/back in and rebooting, but no luck.02:47
ryebotthis is on juju202:48
ryebotalso looked in the /var/log/juju logs in machine-0; nothing there either02:54
ryebotmaking a new controller fixed it03:49
txHey guys, I'm trying to bootstrap openstack from a fresh ubuntu 16.04 LTS install, it has been on the initializing model stage for about an hour now05:51
=== mup_ is now known as mup
=== mup_ is now known as mup
DirlerHi All. If I have kvm machines listed in “virsh list” but not listed in “juju status” for kvm environment, Can I add such machines in juju kvm environment?08:23
jamespagegnuoy, morning09:10
jamespagecan you take a look at https://review.openstack.org/#/c/318641/09:10
jamespagegnuoy, and https://review.openstack.org/#/c/319138/ :-)09:38
=== zz_CyberJacob is now known as CyberJacob
jamespagegnuoy, https://review.openstack.org/#/c/319172/ please (stable counterpart to the one before)10:31
jamespagegnuoy, ta10:32
gahanhow do I force cancelling of activity in landscape? 'Add hardware' is hanging on 'Add jujue machine...' for 24 hours now.11:14
magicaltroutis it acceptable to crackout the beers when you are at work with the cricket on the TV and everyone in the stadium is drinking?11:58
ryebotjuju2 - how do I remove a service that has a unit in an "agent is lost, sorry!" state?12:55
tvansteenburghryebot: you could try remove-machine12:57
ryebottvansteenburgh: excellent, thanks, I'll give that a shot12:58
=== chuck__ is now known as zul
jamespagegnuoy, https://review.openstack.org/#/c/318611/ and then I'll push master and stable branches to the charm store...13:37
lazyPowerkjackal - if you have the spare cycles, this is a pretty small review https://github.com/juju-solutions/layer-filebeat/pull/813:43
kjackalI will do it in a moment lazyPower13:49
kjackaljcastro ?13:49
kjackaljcastro, I hear we have an issue with the pluin?13:50
jcastrokjackal: yeah!13:59
jcastrolet me refire up the bundle, give me a moment13:59
jcastrothis is the bundle I am trying14:00
jcastrohmm, issues with the store today?14:00
kjackalyes, it seems so...14:00
kjackalin any case, which bundle are you deploying, jcastro? Can you show me the .yaml you are deploying?14:01
jcastrohah no, I was deploying right from the store14:01
kjackalAh, that might be an issue, we havent published the charms yet, so the referenced ones might to be working14:02
=== mup_ is now known as mup
kjackaljcastro, I guess you are deploying this: https://jujucharms.com/u/bigdata-dev/apache-processing-spark/bundle/114:08
jcastroI needed a bundle that used beats so I could write about it, and cory recommended this one14:09
jcastrobut I'm not wedded to any specific bundle, I just need to be able to deploy a workload, the post itself is about how to use beats to get metrics from a workload14:09
kjackalso the bundle.yaml there is suposed to use the production charm: cs:trusty/apache-spark)14:09
kjackalThis yaml is the one we want https://api.jujucharms.com/charmstore/v5/~bigdata-dev/bundle/apache-processing-spark-1/archive/bundle-dev.yaml14:10
jcastroyep, I see that14:10
jcastrooh ok, so deploy the -dev bundle instead is what you're saying?14:10
kjackalYes, this should work14:10
kjackalI wonder if there is a way to specify the yaml within the bundle when you reference the bundle from the store?14:11
jcastronot yet14:11
jcastrobut that's fine14:11
jcastrothough next time you guys publish please fill out the metadata url's so I can click through to the code, etc.14:12
jcastroright now "view code" is broken, etc. on the charm store page14:12
jcastrooh, nevermind, you did fill it out14:12
jcastroit's just "Home" and not "view code" on the page.14:12
kjackallazyPower, the patch looks good. Merged!14:22
lazyPowerta kjackal!14:22
suchvenuI have a query regarding naming the relation in  ibm-db2 charm. The relation name is "db2" defined in metadatafile as14:59
suchvenuprovides:  db:   interface: db214:59
suchvenusorry the relation name is "db"14:59
jcastrolazyPower: I think I found a bug in ~containers/kibana15:00
jcastrounit-kibana-0: 2016-05-20 14:55:26 INFO unit.kibana/0.install logger.go:40 groupadd: group 'kibana' already exists15:00
jcastrothat causes the install hook to fail15:00
suchvenuIs this name ok ? Or should it be renamed to some other name as db2-db or so ?15:00
lazyPowerah yeah15:00
lazyPowerthere's a patch to fix this, its known15:01
jcastroack, thanks15:01
jcastrois this the promulgated charm or is that somewhere else?15:01
lazyPowernegative, the promulgated charm is missing the dashboard action15:02
jcastrohmm, it sucks that the page doesn't make it obvious which is the promulgated charm15:02
jcastroother than looking at the URL I mean15:02
lazyPowerwhen we were riffing a couple weeks ago, i've delayed pushing the updates until i can get the config option to deploy *with* a dashboard15:02
lazyPowerand still need to trim the fat on the demonstration boards15:02
suchvenuI see the same relation name used by mysql charm. So will it cause confusion, when both are used in any bundle ?15:02
lazyPowersuchvenu - the relation-name is arbitrary. The interface is the important part of that equation.15:03
suchvenuwhen the states of the interface are used in the reactive layer15:03
suchvenuthe states from the db2 interface would be like db.available, db.ready etc15:03
lazyPowerwell, interestingly enough - do you use both mysql as well as db2 in the same charm? and is this a common deployment formation?15:04
lazyPowersorry, model formation15:04
suchvenui am not using both in any of the charm as of now15:04
suchvenuI just told an example15:05
lazyPowerjcastro - i'm in a lull with other things i have in flight, i can spend some time on that today. Do you mind being my primary stakeholder? I think i can get you an updated charm pushed @ containers just before standup15:05
jcastroyeah no worries15:05
lazyPowersuchvenu - ok. yeah. you should be fine. The onus of using those states and how its named is up to the charm author consuming DB215:05
jcastroI can't blog about beats without a dashboard, heh15:05
lazyPowerok, i'll fold in that patch and get you a revised dashboard15:05
suchvenudoes the relation name need to be unique so that the states also remain unique ?15:06
jcastrowhat's the tldr difference between this kibana and the promulgated one?15:06
lazyPowerthe dashboard loader action and test updates so it actually passes and flex's the deployment15:06
suchvenua decorater like this , is fine  ?15:08
lazyPowersuchvenu - That looks fine, but without context its hard to say15:10
lazyPowersuchvenu - you can mitigate any name collisions by setting the mysql database relation to something other than db when you add it ot metadata, like "mysql-db" or "mdb" or "notibmdb2"15:11
suchvenuI am thinking of the naming convention for the relation name . I have used db as the relation and if some other database (other than db2) also uses the same relation name , will it cause any issues15:12
suchvenuyou mean to say, the consumer charm of db2 or mysql can create a specific relation name as db2-db or msql-db and so it will not have issues ?15:13
suchvenuThanks lazyPower15:13
lazyPowerrelation-names are arbitrary. As the charm author consuming the interface, you define what that nomenclature is, and are responsible for the associated states.15:14
lazyPowernp happy to help suchvenu15:14
cory_fukwmonroe, kjackal: I submit https://github.com/juju-solutions/layer-apache-bigtop-base/pull/7 and https://github.com/juju-solutions/layer-hadoop-datanode/pull/1 for your consideration15:29
cory_fukwmonroe: Also, looks like we don't have to mess around with the bzr-owner hack any more: https://github.com/CanonicalLtd/jujucharms.com/issues/245#event-66646630015:31
kwmonroewoohoo!  nice cory_fu15:32
kwmonroecory_fu: for the dn PR, do you feel ok by returning from start_datanode with the status set as "starting datanode"?15:34
kwmonroei'm talking about if/when start_datanode hits the timeout and returns, of course15:34
cory_fuOh, no, I suppose not15:34
cory_fuGood catch15:34
magicaltroutrandom question big data folk: https://github.com/juju-solutions/layer-hadoop-client doesn't actually provide me with a hadoop executable15:37
magicaltroutI need my charm to do something like: hadoop jar $THRAX/bin/thrax.jar15:38
cory_fukwmonroe: Updated.  Though, I should hope that the status handling in the slave charm would actually override it15:39
cory_fumagicaltrout: Right.  Maybe we need to update the README.  The way that is intended to work is that you use layer:hadoop-client as a base layer, and then you connect it to a hadoop-plugin charm (e.g., https://jujucharms.com/hadoop-plugin/ or the older https://jujucharms.com/apache-hadoop-plugin/ for the non-Bigtop charms)15:41
bdxopenstack-charmers: will there be a charm rev'ing again before 16.07 ?15:41
cory_fumagicaltrout: The plugin is what provides the Hadoop libraries, and it ensures you get the correct libraries for the particular deployment of Hadoop to which you are connecting15:41
magicaltroutyeah cory_fu when I connect the plugin to my layer:hadoop-client enabled charm it all works15:42
magicaltroutbut "which hadoop" is empty and I can't figure out what its doing :)15:42
magicaltroutI thought it main role was to provide config information over the interface15:42
cory_fumagicaltrout: Can you give me a pastebin of your `juju status --format=tabular`?15:42
cory_fuIt provides config info over the interface but also installs the  client libs (the plugin; the client just manages the plugin relation for you to make it a bit easier)15:43
magicaltrouti tried that yesterday and got zip15:43
magicaltroutgive me 5 mins to spin up my charm and I'll get back to you15:43
cory_fumagicaltrout: Are you deploying the older apache-hadoop-X charms, or the newer Bigtop charms?15:44
magicaltroutthe older stuff15:44
magicaltrout juju deploy cs:bundle/apache-processing-mapreduce-015:44
cory_fuThe status messages ought to tell you what, if anything is missing, and if they all say ready then you should have the hadoop bin15:44
magicaltroutthat one to be exact15:44
magicaltroutall my status flags were green15:45
cory_fumagicaltrout: One question.  Are you trying to run the Hadoop binary from inside charm code?  There is environment data that doesn't seem to get populated in the hook context for whatever reason, so it requires a bit of additional work.15:45
kwmonroecory_fu: both PRs lgtm.  merged.15:46
magicaltroutnope cory_fu juju ssh'd in15:46
cory_fuThat should definitely work15:46
cory_fukwmonroe: Shall we do a JBD release, charm builds and publishes, and update the Bigtop PR?15:47
cory_fukwmonroe: Also, would you consider this a bugfix release for JBD or a minor feature release?15:48
cory_fuI'm inclined to go with bugfix15:48
kwmonroeyeah cory_fu, i've got jbd 7.1.2 building now15:54
Brochachocholcombe: Is there anyway to ignore ceph config parsing errors?16:32
cholcombeBrochacho, yeah put a newline at the end of the file.  We have a patch going in to fix that16:33
Brochachocholcombe: Thanks! Was driving me nuts, doesn't seem there's an option to silence that?16:36
cholcombeBrochacho, no and ceph is really bitchy about it16:36
cholcombei thought about submitting a patch to ceph because it's so annoying haha16:36
Brochachocholcombe: dang, also seems to be no way to avoid that 'dumped all in format json'?16:39
cholcombeBrochacho, when you say --format=json is it appending something to stdout?16:39
Brochachocholcombe: One sec16:43
Brochachocholcombe: No, wasn't grabbing stdout correctly -_-16:44
cholcombeah ok haha16:44
magicaltrouti'm talking $hit cory_fu18:04
magicaltroutI suspect i probably sshd in yesterday before the path was set18:04
magicaltroutand never noticed18:04
cory_fu:)  Glad it's working for you18:04
magicaltroutwell you'll have a bunch of happy linguists now who can build translation models on proper infrastructure and not standalone hadoop on their laptops18:05
cory_fuAwesome.  :)18:19
cory_fukwmonroe: Do you use Chrome?18:19
cory_fukwmonroe: Well, anyway.  You chastising me about approving the PR with trailing whitespace made me feel really bad, so I made this: https://github.com/johnsca/github-trailing-whitespace18:27
lazyPowercory_fu https://twitter.com/lazypower/status/73372662882885222418:30
cory_fulazyPower: ha.18:31
cory_fuIf I'd known you were going to tweet it, I would have ponied up the $5 to publish it to the Chrome Web Store18:31
kwmonroecory_fu: sometimes i wonder what greatness you could achieve if i didn't bug you about petty stuff.18:47
kwmonroetoo bad we'll never know18:47
kwmonroecory_fu: i literally think you're great.  i just spent the afternoon loading all kinds of unpacked extensions and now chrome works like firefox... in that it's mostly broken.19:20
cory_fuWhat other unpacked extensions did you load?19:20
kwmonroemostly stuff i wrote19:21
kwmonroebut let's not dwell on the past.  i have a bundle update for bigtop.  can you teach me git real quick so i can update pr 108?19:22
lazyPoweri can tell its Friday19:22
magicaltroutgit commit -a -m "mega commit"19:23
magicaltroutgit push -f19:24
magicaltroutnuke it all!19:24
kwmonroemagicaltrout: join the appear.in!19:24
lazyPowermagicaltrout - thats fun when it pipelines into prod on friday at 4:59pm19:25
kwmonroebeing that it's 2:25pm, i'm not that concerned19:25
magicaltrouti'm drinking beer and writing code for NASA at 8:30pm19:28
magicaltroutwhat could possibly go wrong?19:28
magicaltroutgit push -f19:28
magicaltroutoooh sugar19:28
lazyPowerjcastro cory_fu  - i have a branch ready which should squash the last issues with kibana and bring this up to spec for CWR efforts - https://code.launchpad.net/~lazypower/charms/trusty/kibana/add-dashboard-loader-action/+merge/29535920:03
lazyPowernot sure if you want to poke this now or leave it, but i thought I would ping with it regardless20:03
gennadiyhello everyone, can we deploy local charms to juju2 ? i try to use "juju deploy local:trusty/sipp" but got error "unknown schema for charm URL "local:trusty/sipp""20:57
tvansteenburghgennadiy: just give it a path20:59
magicaltroutgennadiy:  juju deploy --repository=/home/bugg/charms local:trusty/joshua-full joshua-full20:59
gennadiya lot of changes in version 2 :) thanks21:00
magicaltroutbit of that21:00
tvansteenburghmagicaltrout: that's juju121:00
tvansteenburghgennadiy: all the details at `juju help deploy`21:01
magicaltroutaww tvansteenburgh whatever21:02
marcoceppigennadiy magicaltrout `juju deploy /home/bugg/charms/trusty/joshua-full` is all you need21:02
magicaltrout 21:02
magicaltroutbugg@tomsdevbox:~$ juju --version21:02
marcoceppimagicaltrout: beta7 is the latest ;)21:03
magicaltrouti'm just a bit slow to update :P21:03
magicaltroutmarcoceppi: i get scared of updating these days :P21:03
marcoceppimagicaltrout: I understand, we're getting close to RC though! when we get that it'd be nice ot have you an dothers update to help shake the tree21:05
magicaltrouti'm only messing marcoceppi its just a dev environment, i'm not sure why its stuck on beta4, must come from a random ppa or something21:05
marcoceppimagicaltrout: probably, there were package renames and such for 2.021:06
arosaleskwmonroe: cory_fu are https://jujucharms.com/u/bigdata-dev/apache-processing-mapreduce/bundle/2 and https://jujucharms.com/u/bigdata-dev/apache-processing-spark/bundle/1 the latest or have these been promulagated?21:06
cory_fuarosales: Sorry, those are the most recent, yes21:23
gennadiyi use local env with lxd how to run machines in privileged mode? i need to change "ulimit -n"21:30
gennadiymaybe i can provide lxd profile for juju21:32
arosalescory_fu: oh need to apologize, just making sure I found  / using the current ones21:34
cory_fukwmonroe: You still around?22:42
arosaleskwmonroe: if your still around22:44
arosaleskwmonroe: http://paste.ubuntu.com/16538524/22:44
arosaleswaiting on plugin :-(22:44
arosalescory_fu: admcleod  ^22:45
cory_fuarosales: Hrm.  I think we had a fix for that, and I thought kjackal published it to bd-dev earlier today22:45
cory_fuarosales: Can you give me the [Services] section to see what charm revs you're  using?22:46
arosalesya, seems it cropped back up or didn't make to dev branch22:46
cory_fuarosales: I think the charm may have been updated but the bundle missed22:46
cory_fuarosales: You're not even using the bd-dev versions of the charms22:47
cory_fuThe main bundle.yaml must be pointing to the prod charms, which makes sense, but doesn't help for demo / testing22:48
cory_fuarosales: Do you need a fix asap?22:49
arosalescory_fu: just doing a lightning talk, but I can gloss over that as I am almost up22:50
arosalescory_fu: should I have been using a different bundle?22:50
cory_fuarosales: If you had time, I'd say `juju upgrade-charm --switch spark cs:~bigdata-dev/trusty/apache-spark` but it sounds like that would be risky and too close to the wire22:51
* arosales will try it :-)22:51
magicaltroutgo on arosales live life on the edge!22:51
arosaleserror: invalid service name "cs:~bigdata-dev/trusty/apache-spark"22:52
cory_fuarosales: the bundle-dev.yaml in that bundle should have worked, but we need to get those charms promulgated and sorted.  We switched focus all on to bigtop and they got left behind22:52
cory_fuarosales: My args were backwards22:52
arosalesmagicaltrout: keeps things exciting22:53
* arosales waves to magicaltrout22:53
cory_fujuju upgrade-charm cs:~bigdata-dev/trusty/apache-spark spark22:53
cory_fuI forgot the --switch22:53
arosaleserror: unrecognized args: ["spark"]22:53
cory_fuDoes anyone here remember the syntax for upgrade-charm --switch?  :p22:53
cory_fuAlso, does that even still work in 2.0?22:54
magicaltroutthe way things change i barely get through deploying with out doing juju help commands  :P22:54
cory_fujuju upgrade-charm spark --switch cs:~bigdata-dev/trusty/apache-spark22:55
cory_fuarosales: ^22:55
* arosales reading help22:55
arosalesya that looks better22:55
arosalescory_fu: thanks22:55
* cory_fu sits back and watches arosales's lightning talk burn.22:56
cory_fuEr, I mean, good luck!22:56
* kwmonroe has much faith22:56
arosalesI only have 39 machines running22:56
arosaleswhat could go wrong22:56
arosaleskube, swarm, spark, and 2 hadoop clusters22:56
* arosales had to make another ec2 support request to up limits22:57
kwmonroearosales: up from 50?!?!22:58
arosaleskwmonroe: well for us-west-2 as us-east-1 errored out on provisioning22:58
kwmonroeazure will go until your wallet empties without all this "limit" nonsense22:59
* magicaltrout makes a note not to use azure23:00
arosalesthe upgrade worked, but I lost ha on spark23:01
cory_fukwmonroe: The reason I pinged you, btw, is I wanted to know where to file bugs against layer:ibm-base23:01
kwmonroecory_fu: file them against cory.johns@canonical.com23:02
cory_fuarosales: Did you lose HA, or did the status message just change?  I think it reports it slightly differently23:02
cory_fuarosales: Can you give me the new status?23:02
cory_fukwmonroe: ha23:02
cory_fukwmonroe: Also, did you know that Monday is a holiday?  I didn't.23:03
kwmonroecory_fu: a holiday for who(m)?23:03
cory_fukwmonroe: All of the US23:03
cory_fukwmonroe: It's Memorial Day23:03
kwmonroeholy crap, is it May already?!?!23:04
cory_fuIndeed.  It's almost my birthday, even.  :p23:04
cory_fuarosales: "Fetching resources" isn't the most helpful status message23:04
cory_fuarosales: I think the status message may no longer include "HA" but it should mention which is master23:05
arosalessorry wrong buffer23:05
arosalesworkload status = standalone23:06
kwmonroearosales: that looks legit23:06
cory_fuarosales: Yep.  The one that says (standalone - master) means that it is in HA23:06
kwmonroeyou're in 'standalone' mode, which means any of those can take over at any moment23:06
arosalesah yes, I was looking at unit 123:06
arosalesunit 0 = Ready (standalone - master)23:06
arosalesgood stuff23:07
magicaltroutmidnight rolled around, it is my birthday, i beat you cory_fu :P23:07
arosalesHappy Birthday magicaltrout !23:07
cory_fukwmonroe: But for reals, clicking "Bugs" on https://code.launchpad.net/~ibmcharmers/layer-ibm-base/trunk goes to a 40423:07
kwmonroehappy bday magicaltrout!!!23:07
cory_fumagicaltrout: Really?!?  Happy birthday!!!!!23:08
magicaltrouti got my age wrong earlier23:08
magicaltrouti'm over it23:08
arosaleshappy early birthday cory_fu23:08
kwmonroecory_fu: please file a bug re: the bug url.  i'll get to it monday.23:08
cory_fumagicaltrout: I do that all the time.  Actually, my health insurance had my birth year as 1995, and I was super excited to have de-aged until they fixed the glitch23:09
cory_fukwmonroe: ha23:09
cory_fuarosales: :)  Thanks23:09
arosales1995, wow23:09
magicaltroutpeople born in the 90's and 2000's depress me23:09
arosalescory_fu: should have quickly filled for life insurance23:10
arosalespretty amazing how well upgrade charm worked23:10
arosalesand the operational knowledge distilled in those charms23:10
cory_fuarosales: In case you didn't notice (not at all because I forgot to put it in until just now), I asked for next Friday off for my (belated) birthday.  ;)23:11
arosalespretty amazing. On one controller I have 5 models with 40 machines with completly functioning kubernetes, swarm, spark cluster, hadoop cluster (ha), and hadoop cluster (bigtop)23:11
arosalescory_fu: most def you should take Friday off23:12
arosalesmagicaltrout: you too :-)23:12
magicaltroutaww thanks arosales23:12
cory_fuarosales: Well, upgrade-charm --switch is a little duplicitous, considering it changes the entire charm out from under the service.  But Spark is pretty resilient to being switched out, as long as jobs aren't actively running in standalone mode.23:12
kwmonroehehe, magicaltrout, i will give you next thursday off too.  keep being great.23:12
arosaleskwmonroe: is soo generous23:13
arosalescory_fu: ya upgrade implies you are already upgrading23:13
magicaltrouti get tomorrow off.....23:13
cory_fumagicaltrout: I was sad to realize I didn't quite make it under the wire to get hugs if I were born in the 80's23:13
arosalesha, but only tomorrow23:13
magicaltroutyeah sunday the mrs gets her mum day... so i'm lumped with the kids23:14
kwmonroemagicaltrout: are you still coding for nasa?  i feel like i don't want my rovers coded by you at this time of night.23:14
magicaltroutthats work... right?23:14
magicaltrouti am kwmonroe..... currently wondering why requirejs hates me so much :)23:15
arosalesmagicaltrout: indeed it is23:15
magicaltroutyeah thought so23:15
magicaltroutso i get 1 day off23:15
kwmonroenot because you're not a great coder, mind you, but because i fear your maliciousness when you don't get cake for your birthday.23:15
magicaltroutoh well, better than none23:15
magicaltroutdie rover! die!!!!23:15
arosalesmagicaltrout: you need another jubilee year23:16
cory_fumagicaltrout: Or worse: https://www.youtube.com/watch?v=Y6ljFaKRTrI23:16
kwmonroeaight arosales, i have to bid you farewell and good luck.  i'm sure the LT will go great because juju is frickin amazing.  lmk how you fare.  be well all!23:17
arosaleskwmonroe: have a good long weekend23:17
arosalesthanks for the last minute help here23:18
cory_fuarosales: np.  Are you up already for your lightning talk? Are you IRCing from the podium?23:18
arosalesnot up yet23:18
arosaleslike 15 min out23:18
magicaltroutmakes me feel as prepared as i was for my talk23:19
cory_fuI had the impression you were upgrading charms as you walked up to the stage.  :p23:20
cory_fuAlright.  Really must go, or my wife will murder me.  Have a good weekend, all!23:22
* cory_fu disconnects from the Matrix.23:22
cory_fuOh, and happy birthday again, magicaltrout!23:22
arosalescory_fu: have a good weekend23:22
arosalesjuju spoils one. ya I can spin up 5 clusters across 40 machines 1 hour before my talk :-)23:25
magicaltroutdepends if you write the charms yourself or not.... if you use the lovely bigdata dev charms thats probably true ;)23:26
* arosales would do the same with saiku :-)23:28

Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!