=== Tyrantelf_ is now known as Tyrantelf
=== danilos` is now known as danilos
[07:24] <Ankammarao> HI , frankban
[07:25] <Ankammarao> charm release giving error "ERROR cannot release charm or bundle: unauthorized: access denied for user "achittet""
[07:25] <Ankammarao> is this beacase of user issue , please help to resolve this issue
[07:33] <deanman> axw, ping!
[07:33] <axw> deanman: pong?
[07:39] <deanman> axw, thanks for looking into this bug :-). I still get the same error -> http://paste.ubuntu.com/24032595/
[07:39] <deanman> i need juju build as well >
[07:39] <deanman> ?
[07:39] <deanman> built*
[07:40] <axw> deanman: try passing "-u" to go get, to pull the latest source
[07:41] <axw> deanman: latest source for juju-tools that is
[07:43] <deanman> ok got i!
[07:45] <deanman> axw, do you need to have an lxd image running already or you could also do `juju deploy <centos7charm>` and it will work the same?
[07:50] <axw> deanman: it should not need to add-machine first
[07:50] <axw> deanman: juju deploy centos7-charm should work just as well
[08:01] <kjackal> Good morning Juju world!
=== frankban|afk is now known as frankban
[08:41] <frankban> Ankammarao: hey, what's the output of "charm whoami"
[08:48] <Ankammarao> frankban : User: achittet Group membership: ibmcharmers
[08:49] <frankban> Ankammarao: and what command are you using for releasing the charm?
[08:49] <Ankammarao> frankban : charm release cs:~ibmcharmers/xenial/ibm-db2-0
[08:51] <Ankammarao> frankban : with resource charm release cs:~ibmcharmers/xenial/ibm-db2-4 --resource ibm_db2_installer-0
[09:23] <frankban> mhilton: could you please help? ^^^ ibmcharmers have write permission on the charm, achittet is part of ibmcharmers but gets: "ERROR cannot release charm or bundle: unauthorized: access denied for user "achittet"" when trying "charm release cs:~ibmcharmers/xenial/ibm-db2-0"
[09:24] <frankban> Ankammarao: could you please try to log into jujucharms.com as achittet ?
[09:25] <Ankammarao> frankban: You're logged in as achittet
[09:26] <Ankammarao> frankban : logged in to jujucharms.com but no use , still i am getting the same issue
[09:27] <mhilton> Ankammarao, hi, what does charm whoami tell you? does the list of groups include idmcharmers?
[09:28] <frankban> mhilton: already asked, see above: User: achittet Group membership: ibmcharmers
[09:30] <Ankammarao> mhilton : yes listed include ibmcharmers
[09:33] <mhilton> frankban, Ankammarao: OK It looks like that charm is promulgated (at least the trusty one is) The current behaviour is that only members of ~charmers can publish promulgated charms to the stable channel (frankban please correct me if I've got that wrong)
[09:40] <frankban> mhilton: ah, yes, we should probably return a better error message
[09:41] <Ankammarao> mhilton : that charm ealrier pushed to charm store with different user and series(trusty)
[09:41] <Ankammarao> and now i am trying to push it for xenial series
[09:45] <mhilton> Ankammarao, yes the charmstore applies the same rules to all charms that have the same name and owner. The different series is just seen as a variation of the same charm. you should be able to publish to any other channel (such as development or edge). In order to publish to the stable channel however we will need to find somebody who has the permission to do it. frankban do you know who is allowed off the top of your head?
[09:46] <frankban> marcoceppi ^^^
[09:51] <Ankammarao> mhilton : even i have tried with the same user who pushed to the trusty charm earlier , but we are getting the same error
[09:57] <kjackal> Hi Ankammarao, frankban, mhilton, If you have ibm-db2 promulgated you should submit a new revision request to push an update
[09:57] <mhilton> Ankammarao, yes it's nothing wrong with your new user, just bacause the charm is deemed as being in the global namespace (promulgated) only trusted users are allowed to publish new ones.
[09:58] <mhilton> Ankammarao, depending on what you are trying to do (for example if you want to test the charm) you can still publish to other channels. see the --channel flag on publish.
[09:59] <kjackal> Ankammarao: frankban: mhilton: since the promulgated charm is released from the stable channel from your namespace, non-charmers lose any access rights they have on that channel (stable)
[09:59] <frankban> thanks kjackal
[10:13] <Ankammarao> mhilton, so now i can push the charm to other channels like edge
[10:15] <mhilton> Ankammarao, yes it's only the stable channel that becomes restricted on promulgated charms.
[10:19] <Ankammarao> mhilton, i pushed to cs with edge channel and now i am unable to view the charm in the charm store
[10:19] <Ankammarao> it still showing the trusty charm only
[10:23] <mhilton> Ankammarao, I think it is this one https://jujucharms.com/ibm-db2/xenial/4, IIRC the search will only return charms in the stable channel, as that's what most users are likely to want to find.
[10:24] <mhilton> Ankammarao, If the charm is ready to be pushed into the stable channel then you'll need to submit a new revision request as kjackal said earlier (although I don't know how you do that)
[10:27] <kjackal> Ankammarao: mhilton: Submitting a new review request is done through this interface: https://review.jujucharms.com/reviews
[10:27] <mhilton> kjackal, thanks
[10:30] <Ankammarao> mhilton , kjackal, frankban : thanks for your help
[10:31] <kjackal> Ankammarao: what is the charmname+revision of the charm you pushed to edge?
[10:32] <Ankammarao> kjackal : charmname = ibm-db2, charm-revision=4
[10:33] <kjackal> Ankammarao: this charm seems to be xenial only (https://jujucharms.com/ibm-db2/xenial/4) is this what you want?
[10:34] <Ankammarao> kjackal: yes, exactly
[10:34] <kjackal> Ankammarao: Wouldn;t you prefer to have a multiseries (trusty and xenial) for your promulgated charm?
[10:35] <kjackal> Ankammarao: Are you dropping the support for trusty?
[10:37] <Ankammarao> kjackal : no we need for both series
[10:40] <kjackal> Ankammarao: Have you considered the series field in your metadata.yaml? You can set there series: ['xenial', 'trusty'] so you get the same charm available in both trusty and xenial
[10:40] <kjackal> https://jujucharms.com/docs/2.0/authors-charm-metadata
[10:41] <Ankammarao> kjackal, yes i have mentioned both series in the metada.yaml
[10:44] <kjackal> Ankammarao:I am not sure why I see the charm only for xenial. I would expect to see both xenial and trusty as supported platforms.
[10:44] <kjackal> Ankammarao: Have alook at this charm: https://jujucharms.com/apache-hadoop-plugin/trusty/15
[10:45] <kjackal> Ankammarao: Right after "By bigdat-charmers" you see both trusty and xenial
[10:48] <Ankammarao> kjackal: i would have push to both series , i will try to push it again for both xenial and trusty
[10:49] <kjackal> Ankammarao: if you have a multiseries charm you do not need to specify a sieries when you do the push
[10:49] <kjackal> Ankammarao: it should eb just "charm push "
[10:50] <Ankammarao> kjackal: ok, got it thank you
[12:07] <magicaltrout> how are you   .
[12:07] <magicaltrout> nice
[12:07] <magicaltrout> stupid lag
[12:08] <magicaltrout> how are you supposed to move resources around, jujuresources.install() ?
[12:11] <magicaltrout> https://github.com/USCDataScience/sparkler/blob/master/sparkler-deployment/juju/sparkler/reactive/sparkler.py#L34 I have that
[12:11] <magicaltrout> and https://github.com/USCDataScience/sparkler/blob/master/sparkler-deployment/juju/sparkler/resources.yaml#L2
[12:11] <magicaltrout> and it doesn't do anything
[12:11] <magicaltrout> no errors or execution statements in the logs
[12:11] <magicaltrout> just silent failure
[12:17] <magicaltrout> kjackal: any idea?
[12:25] <kjackal> hello magicaltrout, reading... just a sec
[12:33] <magicaltrout> cory_fu should be awake by now, he can shed light on it :)
[12:34] <kjackal> magicaltrout: I think its public holidays in the US
[12:34] <magicaltrout> pfft
[12:34] <magicaltrout> stupid americans
[12:35] <kjackal> So... we are not talking about juju resources
[12:35] <kjackal> these are "old style resources"
[12:35] <magicaltrout> er right
[12:35] <magicaltrout> i googled juju charms resources
[12:35] <magicaltrout> thats what i got :)
[12:36] <magicaltrout> but, it also want something thats not bundled with the charm
[12:36] <magicaltrout> of cource I could just download it using standard python, i was just assuming I was sticking with the current ecosystem
[12:37] <magicaltrout> cource/course
[12:38] <kjackal> magicaltrout: do you want to give juju resources a try? Here is how they work: https://jujucharms.com/docs/2.0/developer-resources
[12:39] <magicaltrout> i have to bundle the resource with the charm though don't I kjackal ?
[12:39] <kjackal> You essentialy declare the resources in metadata.yaml and then at deployment time you can do a juju attach resource
[12:39] <magicaltrout> no then :)
[12:40] <magicaltrout> I'll just grab it using standard python stuff
[12:40] <kjackal> You have to upload a default (potentialy dummy) default resource
[12:41] <kjackal> So for the "old style" resources have a look here: https://github.com/juju-solutions/layer-apache-zeppelin/blob/master/lib/charms/layer/apache_zeppelin.py#L86
[12:42] <kjackal> Which I think is what you are doing
[12:42] <magicaltrout> don't worry about it, i have that stuff, but the install does nothing
[12:42] <magicaltrout> i'll just use urllib
[12:44] <kjackal> https://github.com/juju-solutions/jujuresources/blob/master/jujuresources/backend.py#L106
[12:45] <kjackal> magicaltrout: is it possible the verification is failing?
[12:46] <magicaltrout> maybe its not exactly high on the debug log entries though is it :)
[12:47] <magicaltrout> don't worry about it kjackal we'll move to new resources eventually
[12:47] <magicaltrout> I'll still with urllib
[12:47] <magicaltrout> stick
[13:30] <kklimonda> can I change LXD image server?
[13:30] <kklimonda> it seems to be defaulting to cloud-images.ubuntu.com
=== lazyPower changed the topic of #juju to: http://summit.juju.solutions || https://review.jujucharms.com/ || https://jujucharms.com/docs/ || http://goo.gl/MsNu4I || Youtube: https://www.youtube.com/c/jujucharms
=== hml_ is now known as hml
[15:36] <bdx> magicaltrout: https://gist.github.com/jamesbeedy/c5c2dc96dc517e03ebbde44da5fffa8f
[16:00] <eeemil> Hello, does anyone know how MaaS relates to "juju expose" and/or how I can set up networking on MaaS for enabling expose/unexpose functionality?
[16:01] <bdx> eeemil: good question, I'm not sure the 'expose' command does anything on the maas provider ....
[16:02] <bdx> eeemil: maas doesn't control any kind of port based security, so I'm guessing the 'expose' commnd doesn't have any function there
[16:17] <bdx> is cross-model-relations available under a feature flag somewhere?
[16:23] <kwmonroe> bdx: rick_h probably knows ^^, but in case he's out for the holiday, you might try asking in #juju-dev as well.
[16:28] <rick_h> bdx: yes, in 2.2 though. The in dev stuff.
[16:29] <rick_h> bdx: there might have been some bits in 2.1 but you want to stick around trunk.
[16:30] <erlon> folks, how do I make the bootstrap instance to use a proxy?
[16:30] <erlon> Ive used all proxy, since setting env to using juju model config, but I still getting this error: http://paste.openstack.org/show/599699/
[16:30] <bdx> rick_h: yeah ... just looking forward .... I want to prep my charms for XMR :-) :-) :-)
[16:31] <rick_h> bdx: heh, good news is should need anything. Just support the relation endpoints.
[16:32] <bdx> rick_h: so, if I pull down 'develop', its not behind a feature flag or anything?
[16:32] <bdx> rick_h: how do I "support the relation endpoints"?
[16:32] <kwmonroe> erlon: there's bootstrap config and model config.. for my proxy madness, i do this to get 'em all set:
[16:32] <kwmonroe> juju bootstrap --config http-proxy=$http_proxy --config https-proxy=$https_proxy --config no-proxy=$no_proxy --model-default http-proxy=$http_proxy --model-default https-proxy=$https_proxy --model-default no-proxy=$no_proxy localhost
[16:33] <kwmonroe> erlon: above, i have $[http|https|no]_proxy eniron vars already exported
[16:34] <erlon> kwmonroe: right that is for setting both the bootstrap and the model configs right?
[16:34] <erlon> kwmonroe: is there a way to set the proxy after the bootstrap is created?
[16:35] <kwmonroe> erlon: you could try "juju switch controller; juju model-config [http|https|no]-proxy=foo"
[16:36] <erlon> kwmonroe: I only have 1 controller
[16:37] <rick_h> bdx: sorry, off today running errands so erratic IRC times. I think it's still behind a flag in develop, and by support the relation endpoints I mean if the relation works in a model it'll work across them.
[16:37] <erlon> kwmonroe: ill try to create another controller using this command you suggested
[16:37] <rick_h> bdx: no special charm-fu at this time.
[16:37] <kwmonroe> right erlon, "juju switch controller" means "make the bootstrap/controller model active".  and then the subsequent "juju model-config x=y" would set the config on that active model.
[16:38] <erlon> kwmonroe: does this no-proxy configuration adds the juju-controller net automatically or I have to manually set?
[16:38] <bdx> rick_h: no worreies
[16:39] <bdx> rick_h: lets say I have mysql in model foo, and wordpress in model bar .... so `juju add-relation foo:mysql bar:wordpress` ?
[16:39] <kwmonroe> erlon: i manually set it in my env to the lxd subnet... it's ugly, but i do this in my .bashrc:  export no_proxy=`echo localhost 10.44.139.{1..255} | sed 's/ /,/g'`
[16:39] <kwmonroe> where 10.44.139.x is my lxd subnet
[16:39] <erlon> kwmonroe: if a run juju model-config  with no arguments it shows the list with the parameters that I need there, I believe that is enough to guarantee that it is using what I set right?
[16:39] <rick_h> bdx: basically yes. I think the model name on the local item can/is left off.
[16:40] <bdx> ahh, nice
[16:40] <kwmonroe> right erlon - if 'juju model-config' shows the right stuff, that's all you should need.
[16:40] <erlon> kwmonroe: that is in the juju client container? or inside the boostrap?
=== scuttlemonkey is now known as scuttle|afk
[16:43] <kwmonroe> erlon: that's from my laptop's term (my juju client).  here's my full rundown:  http://paste.ubuntu.com/24034672/
[16:44] <erlon> hmmm, kwmonroe, nice
[17:09] <eeemil> bdx: Thanks for the answer!
[17:09] <bdx> eeemil: np
[17:09] <bdx> rick_h: http://paste.ubuntu.com/24034870/
[17:09] <bdx> I guess I'm not sure exactly what the syntax is there
=== CyberJacob_ is now known as CyberJacob
[17:12] <bdx> rick_h: an application doesn't seem to know about applications in other models via juju-gui either  https://imgur.com/a/ILxcX
[17:12] <bdx> possibly it is behind a flag?
[17:15] <rick_h> bdx: yes behind the flag.
[17:16] <rick_h> Bdx the colon is the correct syntax
[17:16] <bdx> rick_h: whats the flag?
[17:17] <rick_h> bdx: let me check
[17:18] <bdx> rick_h: possibly "consume"? https://github.com/juju/juju/commit/73924d924b49b501b0249a612741f922d46224be
[17:18] <bdx> and/or find-endpoints
[17:19] <rick_h> bdx: see pm
[17:21] <erlon> kwmonroe: maan, love you!
[17:21] <erlon> kwmonroe: thanks very much!
[17:21] <erlon> kwmonroe: are you from ubuntu?
[17:33] <kwmonroe> erlon: glad it helped!  i do work for canonical, but my parents might take offense if i said i was 'from ubuntu' ;)
[17:35] <kwmonroe> erlon: you're not the first to get tripped up by the bootstrap vs model proxy settings.. i'll write something up for https://jujucharms.com/docs/2.0/models-config in case others could benefit from an example.
[17:36] <erlon> kwmonroe: ow,, sorry haha, did know that difference
[17:37] <erlon> kwmonroe: hmm, that would be nice. I would suggest that. Another thing that would be very valuable, is a global view of how the ubuntu deploying tools works
[17:37] <erlon> kwmonroe: like, autopilot, juju, maas, etc,
[17:38] <erlon> kwmonroe: it is very hard to find a way to put all those pieces toguether
[17:38] <kwmonroe> ahh, yup erlon, i'll make a note of that too for our docs.  thanks for the feedback!
[17:40] <erlon> kwmonroe: welcome
[17:42] <erlon> kwmonroe: for example, now the maas says the the nodes are deployed, but what do I do now? haha, do you know any link for a how to with juju? I was following an certification guide that doest not explain this part
[17:48] <kwmonroe> erlon: you lost me at "what do i do now?".  if juju is bootstrapped with your maas env, you can just start deploying charms/workloads, as described here: https://jujucharms.com/docs/stable/getting-started#deploy-applications
[17:49] <kwmonroe> erlon: can you link me to the certification guide you're following?  maybe that would help me understand what you're up to.
[17:49] <erlon> kwmonroe: , hmm, sorry, I went ahead of what you said, I did that, deployed the controller, and then after run a charm 'juju deploy cs:bundle/openstack-base-48 --show-log'
[17:51] <kwmonroe> ah - cool erlon!  did the deployment show any errors?
[17:52] <erlon> kwmonroe: Canonical & Juniper Contrail Lab Solutions Guide Prepared
[17:52] <erlon> kwmonroe: I received from our allience partner at Cananical
[17:54] <erlon> kwmonroe: theres no errors
[17:55] <kwmonroe> erlon: how about "juju status"?  does that look ok?
[17:55] <erlon> kwmonroe: hmm, interesting
[17:55] <erlon> kwmonroe: hold on
[17:56] <erlon> kwmonroe: http://paste.openstack.org/show/599709/
[17:57] <erlon> kwmonroe: vms (4) have only 4GB of ram, I wonder if that is enough
[18:01] <kwmonroe> erlon: you have reached the limit of my knowledge -- i'm really not sure about vm/ram reqs for the openstack-base bundle.    would you mind joining the #openstack-charms channel, posting your paste, and asking if your env meets the minimum specs?
[18:02] <erlon> kwmonroe: sure, I will, but juju status revealed a lot of useful information
[18:03] <erlon> kwmonroe: like, where will be the dashboard
[18:04] <kwmonroe> ah yes erlon -- 'juju status' is definitely the first place to check for details about a deployment
[18:15] <julen> Hi there! Is there any tutorial for installing juju on a server which already has MaaS on it?
[18:17] <julen> The "juju generate-config" is not there anymore, and everything I find is build on LXD. But following the instructions I get conflicts between LXD and MaaS (as they both run on the same machine)
[18:20] <erlon> kwmonroe: is there a clean way to stop all and re-start? there are some services that seems to be stuck: http://paste.openstack.org/show/599710/
[18:20] <erlon> kwmonroe: or should I wait more?
[18:22] <kwmonroe> erlon: checkout "juju debug-log --tail"
[18:23] <kwmonroe> erlon: if that doesn't produce any output for a few minutes, then it might be stuck.  if it looks busy, i'd wait more.
[18:25] <erlon> kwmonroe: yeah, still running, but Ill delete the deployment increase the memory and restart again after it finishes
[18:26] <kwmonroe> erlon: you don't need to be nice to it.  you can do "juju destroy-controller cloud-name-hds-4 --destroy-all-models" to tear it down even if it's not done.
[18:27] <kwmonroe> it doesn't feel pain :)
[18:27] <erlon> kwmonroe: hmm, nice, much better :)
[18:27] <erlon> kwmonroe: Im so happy it went off that I dont want to hurt it lol
[18:28] <kwmonroe> one thing to mention erlon
[18:28] <kwmonroe> i'm assuming you don't have any other models on the cloud-name-hds-4 controller
[18:28]  * kwmonroe should have asked that earlier
[18:28] <kwmonroe> the --destroy-all-models flag will tear down the entire controller.. just fyi.
=== GoosGoarch is now known as ayan
[18:39] <erlon> hmm, no I dont
[18:43] <erlon> kwmonroe: hmm, it seems it have finished now, what is the url of the dashboard?
[18:44] <erlon> kwmonroe: it says dashboard is in: openstack-dashboard/0*    active    idle   3/lxd/2  192.168.51.78   80/tcp,443/tcp  Unit is ready
[18:44] <erlon> kwmonroe: but I only can see a standard apache welcome
[18:45] <erlon> 192.168.51.78/dashboard also has nothing
[18:46] <erlon> kwmonroe: wait, nevermind, log says there still things being configured
[18:47] <kwmonroe> erlon: not 100% what URLs are available to you, but the openstack-base readme makes me think it's http://192.168.51.78/horizon - see towards the end of this page: https://jujucharms.com/openstack-base/
[18:48] <erlon> kwmonroe: hmmm, there it is the requirements, sorry about filling you up with questions, own you a beer
[18:49] <kwmonroe> :)  np erlon
[18:49] <erlon> kwmonroe: have you been to the OpenStack summit?
[18:50] <kwmonroe> i haven't erlon.  should i go to the next one to cash in on my free beer?
[18:51] <erlon> kwmonroe: haha, I asked because beers don't  have much value in the Summits, they are usually given at free for everyone, you should definitely go
[18:53] <kwmonroe> well, "free for everyone" is not the same as "free, hand-delivered by erlon".  i'll mark my calendar :)
[18:54] <erlon> kwmonroe: owww nice!! Ill be in the one in Sidney, meet me there :)
[18:59] <erlon> kwmonroe: one more question and I let you live, the bubble page says me to source novarc, but I don't know where juju save those scripts
=== ayan is now known as ayan-afk
[19:03] <kwmonroe> ah yes erlon, i see that too.  just above the section where it says "source novarc" it asks you to "download this bundle ... all commands are executed from within the expanded bundle".  to do that, see the upper-right box on that page -- the "Files" box, and click the "Download .zip" link to grab the bundle.
[19:04] <kwmonroe> erlon: alternatively, wget/curl the link, as in "curl https://api.jujucharms.com/charmstore/v5/openstack-base/archive -o ~/bundle.zip", then unzip it.
[19:04] <erlon> kwmonroe: hmm, sweet
[19:10] <kklimonda> I have a problem with deploying openstack charms keystone (so far) is not configured with mysql vip, and not all 3 controller nodes are granted access to keystone db.
[19:13] <kwmonroe> kklimonda: i'm not sure many openstack charmers are monitoring this channel today (it's a US holiday).  you might have more success in the #openstack-charms channel.
[19:14] <kklimonda> oh, didn't know there was a separate channel
[19:14] <kklimonda> thanks
[19:14] <kwmonroe> np
[20:53] <andrew-ii> I have a maas node that keeps getting stuck at "pendning" when allocating LXD containers. Is there a way to get LXD debug info when it (almost?) attempts to create a container?
[20:54] <andrew-ii> The command `cat /var/log/lxd/lxd.log` just shows a half-hearted attempt to get a xenial image and then... nothing ever happens after that.