[16:39] <ejat> hi .. how to restart juju machine/services once change the setting/configuration ?
[16:47] <bdx> ejat: juju help run
[16:48] <ejat> ok thanks
[16:49] <bdx> ejat: it would be something like - juju run --unit nova-compute/0 "service nova-compute restart"
[16:49] <ejat> im trying to deploy charm openstack base
[16:50] <ejat> once completed .. i cant login
[16:51] <bdx> ejat: ensure you are using the credentials you specified for admin-password in your charmconf.yaml
[16:52] <bdx> ejat: for keystone
[16:52] <ejat> i just use default
[16:53] <ejat> its redirected to /horizon/admin but timeout
[16:53] <bdx> so are you extrapolating the randomly generated password from the keystone service in that case?
[16:54] <bdx> ejat: your horizon default login should be un: admin, pw: <randomly generated password unless set>
[16:54] <bdx> https://jujucharms.com/keystone/trusty/28
[16:55] <ejat> the default pw : openstack right ?
[16:55] <bdx> ejat: why would you think that?
[16:55] <ejat> its in services config
[16:55] <bdx> ejat: oh...are you deploying a bundle than I take it?
[16:55] <bdx> ahh gotcha
[16:56] <bdx> can you paste the url you are using?
[16:56] <ejat> yeah the bundle
[16:57] <ejat> url for my deployment or the url for the bundle ?
[16:57] <bdx> ejat: also can you paste the url generated by running the command "juju status --format=tabular | pastebinit"
[16:57] <bdx> for the bundle
[16:57] <ejat> https://jujucharms.com/openstack-base/36
[16:58] <bdx> ejat: ok, I see what you are saying.
[16:58] <ejat> bdx : http://paste.ubuntu.com/12503691/
[16:59] <ejat> i have to set /etc/hosts in openstack-dashboard the internal IP
[17:00] <ejat> because inside openstack-dashboard putting the public dns to connect to keystone
[17:01] <bdx> ejat: can you run "juju resolved nova-cloud-compute/0 --retry" and then "juju status --format=tabular | pastebinit"
[17:01] <bdx> and paste me the url
[17:02] <ejat> i try several time to retry .. its up then down again
[17:03] <ejat> fenris@macbuntu:~$ juju resolved nova-cloud-compute/0 --retry
[17:03] <ejat> ERROR unit "nova-cloud-compute/0" not found
[17:03] <ejat> fenris@macbuntu:~$ juju status nova-cloud-compute
[17:03] <ejat> environment: azure2
[17:03] <ejat> machines: {}
[17:03] <ejat> services: {}
[17:04] <ejat> ok .. its up .. wait
[17:05] <ejat> http://paste.ubuntu.com/12503779/
[17:06] <ejat> tried to login dashboard : The connection was reset
[17:07] <bdx> ejat: ok, ...I just wanted to get rid of the error state of the hook on the nova-cloud-controller before we move forward
[17:08] <bdx> ejat: Your telling me you can login now?
[17:08] <ejat> nope
[17:08] <bdx> ok
[17:08] <ejat> still cant login
[17:09] <bdx> ejat: what if any config modifications are you making to openstack-dashboard concerning dns?
[17:10]  * ejat not modify .. just adding the host in /etc/hosts
[17:10] <ejat> keystone host in dashboard /etc/hosts
[17:11] <bdx> ok
[17:11] <ejat> keystone/0
[17:11] <ejat>     IP Address: 10.0.0.46
[17:11] <ejat>     Status: started
[17:11] <ejat>     Public Address: 10.0.0.46
[17:12] <ejat> its doesnt get the right public dns : juju-azure2-j41rc8jcm9.cloudapp.net
[17:14] <bdx> ah ok, so whats going on here is you also need to add the entry for openstack-dashboard in keystone
[17:14] <ejat> vise versa ?
[17:15] <bdx> so do the same thing you did for keystone but for openstack-dashboard, does that make sense?
[17:15] <bdx> ya
[17:15] <ejat> let me try
[17:16] <bdx> once you have made the change, run "juju run --unit openstack-dashboard/0 "service apache2 restart""
[17:20] <ejat> un : admin , pw: openstack ?
[17:20] <bdx> yea
[17:21] <ejat> still ... connecting ...........
[17:21] <ejat> redirected to /horizon/admin with The connection was reset
[17:22] <ejat> is it need to define all machine the local/internal IP ?
[17:22] <ejat> :(
[17:26] <ejat> bdx :
[17:26] <bdx> can you paste the url from "juju run --unit openstack-dashboard/0 "apt install pastebinit > /dev/null 2>&1 && cat /etc/openstack-dashboard/local_settings.py | pastebinit""
[17:26] <bdx> ejat: possibly
[17:27] <bdx> hopefully not, we will see
[17:28] <ejat> http://paste.ubuntu.com/12504087/
[17:29] <bdx> ejat: also, what happens if you run "juju run --unit openstack-dashboard/0 "host juju-azure2-j41rc8jcm9.cloudapp.net""
[17:30] <ejat> juju run --unit openstack-dashboard/0 "host juju-azure2-j41rc8jcm9.cloudapp.net"
[17:30] <ejat> juju-azure2-j41rc8jcm9.cloudapp.net has address 207.46.155.248
[17:32] <ejat> if i didnt add inside /etc/hosts ... dashboard cant ping to that address
[17:34] <bdx> ejat, hmmm did you add "207.46.155.248 juju-azure2-j41rc8jcm9.cloudapp.net" to /etc/hosts?
[17:34] <ejat> in dashboard ?
[17:34] <bdx> or the internal 10.0.0 ?
[17:35] <ejat> 10.0.0.x
[17:35] <bdx> yea?
[17:35] <ejat> not public
[17:35] <ejat> u want me to add that ?
[17:35] <bdx> ok, lets try using the public ip in the hosts files on for keystone and openstack-dashboard
[17:36] <ejat> one second
[17:37] <ejat> then restart apache2?
[17:37] <bdx> so in keystone /etc/hosts -> "23.99.125.58 juju-azure2-3a6jyl44mj.cloudapp.net"
[17:37] <bdx> and in openstack-dashboard /etc/hosts "207.46.155.248 juju-azure2-j41rc8jcm9.cloudapp.net"
[17:38] <bdx> once you have made the change, run "juju run --unit openstack-dashboard/0 "service apache2 restart""
[17:39] <ejat> done .. trying to login the dashboard
[17:40] <ejat> still the same
[17:40] <ejat> redirected to /horizon/admin then timed out
[17:41] <bdx> darn
[17:41] <bdx> if you are not opposed, we can redeploy the dashboard
[17:42] <ejat> mean .. destroy n redeploy?
[17:42] <bdx> "juju destroy-service openstack-dashboard && juju destroy-machine 26 --force"
[17:42] <bdx> yea
[17:42] <ejat> :)
[17:42] <ejat> ok
[17:42] <bdx> then
[17:42] <ejat> im ok ..
[17:43] <bdx> lets specify a charmconf.yaml for openstack-dashboard before we redeploy
[17:43] <ejat> ok ..
[17:44] <ejat> got sample for charmconf.yaml ?
[17:44] <bdx> yea omp
[17:44] <ejat> omp ?
[17:44] <ejat> on manual page ?
[17:45] <ejat> dont destroying the dashboard
[17:45] <bdx> this is a kilo deploy right?
[17:45] <ejat> done*
[17:45] <bdx> ok
[17:45] <ejat> not sure what r the charm get the version
[17:46] <bdx> http://paste.ubuntu.com/12504309/
[17:47] <bdx> put that into a file e.g. charmconf.yaml
[17:47] <bdx> then run these commands
[17:47] <bdx> "juju deploy openstack-dashboard --config charmconf.yaml && juju add-relation openstack-dashboard keystone"
[17:48] <bdx> or command I should say :-)
[17:50] <bdx> and then paste me the url from "juju status --format=tabular | pastebinit"
[17:52] <ejat> http://paste.ubuntu.com/12504424/
[17:52] <ejat> then expose ?
[17:52] <bdx> ummm
[17:52] <bdx> hold on
[17:53] <ejat> okie
[17:53] <bdx> run this again "juju status --format=tabular | pastebinit"
[17:54] <ejat> http://paste.ubuntu.com/12504545/
[17:54] <ejat> waiting agnet
[17:54] <ejat> agent*
[17:54] <bdx> ok
[17:55] <bdx> let me know when its finished and in idle state, then paste me the url
[17:55] <bdx> sorry
[17:55] <ejat> ok
[17:55] <ejat> no worries .. thanks for ya assistant :)
[17:56] <bdx> np
[17:57] <bdx> run this this time "juju status | pastebinit"
[18:01] <bdx> and?
[18:02] <ejat> http://paste.ubuntu.com/12504742/
[18:02] <ejat> not yet :)
[18:02] <bdx> darn...
[18:02] <bdx> ha
[18:03] <ejat>  installing charm software
[18:04] <ejat> basic instanse hware
[18:04] <ejat> its up
[18:04] <ejat> expose?
[18:04] <bdx> run this this time "juju status | pastebinit"
[18:05] <ejat> http://paste.ubuntu.com/12504940/
[18:05] <bdx> run this this time "juju status | pastebinit"
[18:05] <ejat> http://paste.ubuntu.com/12504948/
[18:06] <bdx> ok yea now expose
[18:07] <ejat> http://paste.ubuntu.com/12504988/
[18:07] <ejat> what is the pw ?
[18:08] <bdx> it should be un: admin, pw: openstack
[18:08] <bdx> but im trying to login atm...no dice
[18:09] <ejat> :(
[18:09] <bdx> what are the internal ips for keystone and openstack-dashboard?
[18:10] <ejat> openstack-dashboard/0
[18:10] <ejat>     IP Address: 10.0.0.52 | Ports 80/tcp, 443/tcp
[18:10] <ejat>     Status: started
[18:10] <ejat>     Public Address: juju-azure2-6yl2jtw74n.cloudapp.net | Ports 80/tcp, 443/tcp
[18:10] <bdx> also, lets try this in /etc/openstack-dashboard/local_settings.py I want you to modify OPENSTACK_HOST to be the 10.0.0.x ip for keystone
[18:10] <ejat> hmm .. now its get the right public dns
[18:10] <bdx> and keyston?
[18:10] <ejat> keystone/0
[18:10] <ejat>     IP Address: 10.0.0.46
[18:10] <ejat>     Status: started
[18:10] <ejat>     Public Address: 10.0.0.46
[18:11] <ejat> is it the bundle charm not updated ?
[18:11] <bdx> ok, so set OPENSTACK_HOST=10.0.0.46
[18:11] <ejat> ok
[18:11] <bdx> in local_settings.py
[18:11] <ejat> ok
[18:11] <bdx> and then "juju run --unit openstack-dashboard/0 "service apache2 restart""
[18:14] <bdx> yea?
[18:14] <ejat> done restart
[18:14] <bdx> ok
[18:15] <bdx> do you have the openrc file?
[18:16] <ejat> http://paste.ubuntu.com/12505116/
[18:16] <ejat> nope .. i think ..
[18:16] <ejat> havent get the openrc yet
[18:16] <ejat> and scp it
[18:16] <bdx> if so, can you "juju scp openrc keystone/0:~/ && juju run --unit keystone/0 "source openrc && keystone user-list"
[18:17] <bdx> ok so now on your local machine run "ssh-keygen -f "/root/.ssh/known_hosts" -R 10.0.0.52"
[18:17] <bdx> then try again
[18:18] <ejat> mkstemp: Permission denied
[18:18] <bdx> ssh-keygen -f "/root/.ssh/known_hosts" -R 10.0.0.52 ?
[18:19] <bdx> oooh
[18:19] <bdx> your are root?
[18:19] <bdx> err
[18:19] <bdx> ok
[18:19] <bdx> so
[18:19] <bdx> sudo ssh-keygen -f "/root/.ssh/known_hosts" -R 10.0.0.52
[18:19] <ejat> openrc i dont have :)
[18:19] <bdx> ohh ok
[18:20] <ejat> $ sudo ssh-keygen -f "/root/.ssh/known_hosts" -R 10.0.0.52
[18:20] <ejat> mkstemp: No such file or directory
[18:21] <bdx> ok
[18:21] <bdx> so lets do this as a normal user
[18:21] <bdx> not root
[18:22] <ejat> tenant name = admin
[18:23] <ejat> keystone = internal ip ?
[18:23] <ejat> for the openrc ?
[18:23] <bdx> ok
[18:23] <bdx> I know exactly whats up
[18:24] <bdx> if you want horizon to be able to use keystone's public url you have to expose it
[18:24] <bdx> lol
[18:24] <ejat> horizon expose ?
[18:24] <ejat> or keystone?
[18:25] <ejat> both ?
[18:25] <bdx> otherwise you need to set OPENSTACK_HOST=10.0.0.46 in local_settings.py on the openstack-dashboard
[18:25] <bdx> horizon is already exposed
[18:25] <bdx> but notice in local_settings.py it is trying to talk to keystone's public address which is not exposed
[18:25] <bdx> so
[18:26] <bdx> if you want public access to your api endpoints (which I'm sure you probably will) then expose keystone
[18:27] <ejat> $ juju scp openrc keystone/0:~/ && juju run --unit keystone/0 "source openrc && keystone user-list"
[18:27] <ejat> /tmp/juju-exec096226571/script.sh: line 1: openrc: No such file or directory
[18:27] <ejat> ERROR subprocess encountered error code 1
[18:27] <bdx> ok
[18:27] <ejat> the openrc is in the path
[18:27] <bdx> I know
[18:27] <bdx> its
[18:27] <bdx> ok
[18:27] <bdx> my bad
[18:28] <bdx> if so, can you "juju scp openrc keystone/0:/home/ubuntu/ && juju run --unit keystone/0 "source /home/ubuntu/openrc && keystone user-list"
[18:28] <bdx> but that wont work until you expose keystones public endpoint
[18:29] <ejat> \0/ user-list
[18:29] <bdx> did you expose keystone?
[18:29] <ejat> not yet
[18:29] <bdx> ok
[18:29] <ejat> hehe wait
[18:30] <bdx> are you reading what I'm telling you
[18:30] <bdx> if you look in the openrc it has keystone's public endpoint which is not accessable unless you expose it
[18:30] <bdx> so once you expose it
[18:31] <bdx> "juju run --unit openstack-dashboard/0 "keystone user-list""
[18:31] <bdx> and you should have success
[18:32] <ejat> http://paste.ubuntu.com/12505350/
[18:32] <ejat> :(
[18:32] <bdx> grrr
[18:33] <ejat> change back the internal in local_setting.py ?
[18:33] <bdx> yeah
[18:33] <bdx> if you exposed keystone
[18:33] <ejat> keystone/0
[18:33] <ejat>     IP Address: 10.0.0.46
[18:33] <ejat>     Status: started
[18:33] <ejat>     Public Address: 10.0.0.46
[18:33] <ejat> after exposed ..
[18:33] <ejat> its still the same
[18:34] <bdx> oooh
[18:34] <bdx> ok
[18:35] <ejat> i think something wrong with the bundle openstack base :(
[18:35] <bdx> is that the value of OPENSTACK_HOST? in local_settings.py?
[18:35] <ejat> yups
[18:35] <bdx> 10.0.0.46 ?
[18:35] <ejat> still the same
[18:35] <bdx> ok
[18:35] <bdx> now "juju ssh openstack-dashboard/0"
[18:35] <bdx> can you do that?
[18:35] <ejat> can
[18:35] <bdx> ok
[18:36] <bdx> inside of openstack-dashboard run "sudo service apache2 restart"
[18:37] <ejat> restarted
[18:37] <ejat> $ keystone user-list
[18:37] <ejat> Expecting an auth URL via either --os-auth-url or env[OS_AUTH_URL]
[18:37] <bdx> source openrc
[18:37] <bdx> then keystone user-list
[18:38] <bdx> ejat: inside of openstack-dashboard
[18:39] <bdx> "sudo cat /var/log/apache2/error.log | pastebinit && sudo cat /var/log/apache2/access.log | pastebinit"
[18:39] <ejat> http://paste.ubuntu.com/12505464/
[18:40] <ejat> http://paste.ubuntu.com/12505480/
[18:40] <ejat> http://paste.ubuntu.com/12505481/
[18:41] <bdx> ejat: http://paste.ubuntu.com/12505480/ line 347
[18:42] <bdx> now what I want you to try
[18:42] <bdx> form openstack-dashboard
[18:42] <bdx> edit openrc
[18:42] <bdx> to reflect keystone's internal ip
[18:42] <bdx> 10.0.0.46
[18:43] <bdx> http://paste.ubuntu.com/12505530/
[18:43] <ejat> already .46
[18:43] <ejat> in dashboard
[18:43] <bdx> ok
[18:43] <bdx> now edit the openrc to look like http://paste.ubuntu.com/12505530/
[18:43] <ejat> owh no regionone
[18:43] <bdx> huh
[18:44] <bdx> if you want to share a chrome desktop session with me I can fix this for you alot faster prob
[18:44] <bdx> chrome remote
[18:45] <bdx> do you know about that
[18:45] <ejat> ok
[18:45] <bdx> https://chrome.google.com/webstore/detail/chrome-remote-desktop/gbchcmhmhahfdphkhkmpfmihenigjmpp?hl=en
[18:45] <ejat> one moment
[18:49] <bdx> ready when you are
[18:52] <ejat> bdx : PM
[18:54] <ejat> ?
[18:54] <ejat> the code xpiring
[18:55] <bdx> yeah send it
[18:55] <bdx> pm me
[20:49] <ejat> bdx : thanks u so much ... for keep trying
[21:56] <blr> I have a web service subordinate to squid, with a requires interface defined, however when adding a relation between the web service and gunicorn, _both_ the web service and gunicorn declare that they are subordinate to eachother, which clearly isn't correct.
[22:06] <rick_h_> blr: the issue is that gunicorn is a subordinate: https://api.jujucharms.com/charmstore/v4/trusty/gunicorn-0/archive/metadata.yaml
[22:06] <rick_h_> blr: it's meant to be used to serve out/hook up with your app as a full charm vs a subordinate
[22:06] <rick_h_> blr: and juju does not support a subordinate on a subordinate
[22:06] <blr> rick_h_: oh it doesn't? :/
[22:06] <rick_h_> blr: no, because a subordinate is meant to stick onto something and apply to that machine
[22:07] <blr> I need squid both squid and my web service on the same unit however
[22:07] <blr> err hopefully you can parse that hah
[22:07] <rick_h_> blr: so there we'd suggest using colocation
[22:07] <rick_h_> blr: so I'd do your service as its own charm
[22:08] <rick_h_> and colocate it using the --to command
[22:08] <rick_h_> blr: juju supports containers and 'hulk smash' colocation methods you could use to have them all running on one machine
[22:09] <rick_h_> blr: https://jujucharms.com/docs/stable/charms-deploying#deploying-to-specific-machines-and-containers
[22:09] <blr> thanks rick_h_, I'll try that again.
[22:10] <rick_h_> blr: np, hope that helps some