16:04:19 <arturt> #startmeeting JOID weekly
16:04:19 <collabot`> Meeting started Wed May 25 16:04:19 2016 UTC.  The chair is arturt. Information about MeetBot at http://wiki.debian.org/MeetBot.
16:04:19 <collabot`> Useful Commands: #action #agreed #help #info #idea #link #topic.
16:04:19 <collabot`> The meeting name has been set to 'joid_weekly'
16:04:32 <bryan_att> #info Bryan Sullivan
16:04:39 <narinder> #info Narinder Gupta
16:04:58 <narinder> zahra_jahedi: if you are attending meeting please sign here
16:05:03 <bryan_att> Agenda item: takeaways from Apex Congress discussion for consideration in JOID
16:06:42 <arturt> #link https://etherpad.opnfv.org/p/joid
16:07:00 <arturt> #info Artur Tyloch
16:07:20 <arturt> #topic Colorado release status
16:07:30 <zahra_jahedi> yeah I'm attending
16:08:43 <arturt> zahra_jahedi: add your name to the meeting notes collected by collabot with this command:  #info NAME
16:09:41 <zahra_jahedi> #info zahra jahedi
16:09:50 <David_Orange> #info David
16:10:59 <David_Orange> narinder: we can ear you
16:11:13 <narinder> David_Orange: but i can not hear any of you
16:11:24 <David_Orange> narinder: we ear that :)
16:12:09 <narinder> David_Orange:  any queris on C release
16:12:26 <David_Orange> narinder: artur is explaining Converged arch
16:12:31 <narinder> ok
16:14:01 <bryan_att> What is the "converged arch"? Is this documented on the wiki?
16:14:43 <arturt> bryan_att - not yet
16:34:11 <rprakash_> #info prakash
16:46:53 <arturt> #topic takeaways from Apex Congress discussion for consideration in JOID
16:48:17 <rprakash_> shared the status of JOID with Release team stating that os-nosdn-nofeature-ha was using 14.04 currently and other scenerio'sin queue will be with ubuntu 16.04 LTS
16:52:13 <arturt> #topic OPNFV Summit Berlin
17:02:20 <David_Orange> narinder: arturt: and thanks for this work on converged architecture. it seems really interesting, even if i will have a lot of question with my telco background :)
20:13:27 <fdegir> narinder: ping
20:13:38 <narinder> fdegir: pong
20:13:46 <narinder> fdegir: how are you doing today>
20:13:52 <fdegir> narinder: good
20:13:55 <fdegir> narinder: how are you?
20:14:15 <narinder> fdegir: i am good thanks in Austin at Big communication event.
20:14:40 <fdegir> narinder: good for you
20:14:43 <fdegir> narinder: now
20:15:02 <fdegir> narinder: do you remember the mail I sent a week or two ago
20:15:05 <narinder> fdegir: yeah i am out and working on few stuff
20:15:15 <narinder> fdegir: regarding?
20:15:15 <fdegir> narinder: regarding changing how we run jobs on jenkins
20:15:35 <fdegir> narinder: you know, intel-pod5 runs brahmaputra jobs, intel-pod6 runs master for joid
20:15:40 <fdegir> narinder: we are changing it now
20:15:53 <fdegir> narinder: and your jobs will be regenerated
20:15:54 <narinder> fdegir: ok
20:16:07 <fdegir> narinder: like joid-os-nosdn-nofeature-ha-baremetal-master and brahmaputra
20:16:18 <narinder> that sounds good to me and i am working on config not part of joid
20:16:36 <fdegir> meaning that we will run any joid job on any pod joid has
20:16:36 <narinder> so that it will input to joid and you can deploy the stack dynamically
20:16:44 <fdegir> that's fine
20:16:51 <narinder> woe thats fantasic totally
20:16:54 <fdegir> but the jobs will roam on 2 pods randomly
20:17:06 <narinder> okey only two pods :)
20:17:48 <narinder> thats fine but can i define to have some way to not to reschedule it on a regular interval initally
20:17:49 <fdegir> narinder: this is the first step
20:17:53 <narinder> fdegir: ok
20:17:54 <fdegir> narinder: later - perhaps d release
20:18:01 <narinder> gotch you
20:18:02 <fdegir> narinder: we want to put all pods to common pool
20:18:05 <fdegir> anyway
20:18:08 <fdegir> I'll send a new patch
20:18:09 <narinder> i am totoally +1 to it
20:18:11 <fdegir> and explain details there
20:18:19 <narinder> sure sounds good to me
20:18:27 <fdegir> so please read the details and put your objections there in case if you have any
20:26:02 <narinder> fdegir: sure
04:25:33 <bryan_att> narinder: when you get back on, the charm and result from trying to deploy it are attached to the page https://wiki.opnfv.org/display/copper/Installer+Support
14:44:10 <bryan_att> narinder: do you or your charm experts have some time today to help me move forward on the Congress charm?
14:45:23 <narinder> bryan_att: yes i can sit with you after 2:00 PM CST
14:45:55 <narinder> bryan_att: and if needed then we can ask queries in #juju channel
14:47:25 <bryan_att> narinder: OK, thanks 2PM is great. The first issue I have is the error line "ERROR juju.state.unit unit.go:738 unit congress/0 cannot get assigned machine: unit "congress/0" is not assigned to a machine". Will the juju channel guys be able to help me with that?
14:48:28 <narinder> bryan_att: yes you can point towards your charm code and they can.
14:48:53 <narinder> bryan_att: can u run juju status and send me output?
14:49:22 <narinder> and how did you deploy?
14:49:41 <bryan_att> narinder: here it is https://www.irccloud.com/pastebin/WUHmkYZE/
14:50:42 <bryan_att> narinder: log is at https://wiki.opnfv.org/download/attachments/6821110/congress-charm.log?version=1&modificationDate=1464236675441&api=v2
14:53:20 <narinder> ok error you do not have enough hardware machine in maas to deploy congress on physical machine
14:53:44 <narinder> so either you have to use container or colocate the service.
14:54:27 <narinder> since you are using the maas provider so any deployment looks for a machine from maas
14:54:32 <bryan_att> narinder: ok, is there something I can copy as container config for this? Really I want to deploy it on the same container as Horizon.
14:55:05 <narinder> to colocate the service you can do juju deploy loca:trusty/congress --to 1
14:55:38 <bryan_att> narinder: I just walked thru the guide provided without really understanding it, so I'm sure there are things missing.
14:56:36 <bryan_att> narinder: is the command "juju deploy local:trusty/congress --to 1" ? The command I used was "juju deploy local:liberty/congress"
14:56:55 <narinder> sorry it could be liberty
14:57:07 <narinder> path depends upon where your local charm is
14:57:12 <bryan_att> what is the "--to 1" ?
14:57:31 <narinder> --to 1 means deploy service to machine 1 in MAAS
14:58:03 <narinder> in this case it is 1          started 1.25.5  node1-control.maas /MAAS/api/1.0/nodes/node-dc124fbe-222b-11e6-a6b6-525400525c72/ trusty  arch=amd64 cpu-cores=4 mem=16384M tags=control availability-zone=default
14:58:15 <bryan_att> ok, thanks. I can try that. How would I direct it to deploy on the same containter as Horizon?
14:58:17 <narinder> which is node1-control.maas
14:58:52 <narinder> for that you need to say --tp 1/lxc/6
14:59:06 <narinder> --to 1/lxc/6
15:00:29 <bryan_att> ok, let me try that. I don't want to bug you too much, so I'll followup on juju if that's the right place.
15:00:58 <narinder> you can ask me question and also on #juju channel they will be supportive
15:01:55 <bryan_att> ok, first is I just tried that and got "ERROR cannot add service "congress": service already exists" - how do I remove it to try again?
15:03:22 <bryan_att> maybe via "juju remove-service"?
15:04:42 <bryan_att> narinder: I got this error trying again https://www.irccloud.com/pastebin/RuSAEWdG/
15:07:32 <bryan_att> narinder: same error with "juju deploy local:liberty/congress --to 1"
15:09:14 <bryan_att> afk for a bit
15:50:41 <bryan_att> narinder: I'm going to ask for followup on juju. I've documented the current status at https://etherpad.opnfv.org/p/copper
16:07:12 <bryan_att> narinder: I changed the command to "juju deploy local:trusty/congress --to 1" and it seems to be getting further. The instructions caused me to think I should name the series based upon OpenStack release but apparently that's not how you named it for JOID.
16:08:12 <David_Orange> narinder: hi
16:34:14 <bryan_att> narinder: hook failed: "install" log lines for "juju deploy local:trusty/congress --to 1"  https://www.irccloud.com/pastebin/Hn5V1UgP/
16:41:01 <David_Orange> narinder: orange pod2 is down -> a big auth issue from nova with bad tokens
16:42:10 <David_Orange> narinder: i found something more than stange, on all clusters, all containers have the vip, making pacemaker restart solve the multi vip issue, but it seems not normal
16:43:44 <narinder> bryan_att:     unit-congress-3[1725099]: 2016-05-26 15:58:25 INFO unit.congress/3.install logger.go:40 TabError: inconsistent use of tabs and spaces in indentation
16:43:53 <narinder> which is the root cause
16:44:41 <narinder> David_Orange: vip should be configured in all right to get it services.
16:44:51 <narinder> may be cluster type is active-active
16:45:58 <David_Orange> this is not the case on pod1
16:46:09 <David_Orange> only one vip for 3 servers
16:46:32 <David_Orange> and restarting pacemaker restore 1 vip for 3 servers
16:47:16 <David_Orange> or it is a new behavior
16:48:30 <David_Orange> but after a moment, vip is back on each server
16:48:55 <David_Orange> it may not be linked but this is different between pod1 and pod2
16:53:29 <narinder> David_Orange: i need to understand the changes in pacemaker before i can comment further.
16:53:40 <David_Orange> ok
16:55:23 <David_Orange> it is the same for all clusters of pod2
16:55:54 <David_Orange> i stopped 2 pacemakers on keystone services, to check if auth problem is solved or not
16:58:23 <narinder> David_Orange: ok is it resolving?
16:58:56 <David_Orange> not for the moment
17:02:45 <David_Orange> still the same issue:
17:02:47 <David_Orange> (keystone.middleware.core): 2016-05-26 17:02:21,788 WARNING RBAC: Invalid token
17:02:47 <David_Orange> (keystone.common.wsgi): 2016-05-26 17:02:21,789 WARNING The request you have made requires authentication.
17:03:37 <David_Orange> i have to go, i will follow debug tomorrow if you don't have time for it
22:01:05 <bryan_att> narinder: quick question if you are around. How can I get the values typically found in the admin-openrc.sh without having to download it? Are there juju commands I can use for that?
22:02:32 <narinder> bryan_att: yes you can use juju get to get the charm configuration where value resides
22:03:34 <bryan_att> narinder: ok, do you have a link to where that is described? I see some related guidance at https://help.ubuntu.com/lts/clouddocs/en/Installing-OpenStack.html
22:04:19 <narinder> bryan_att: i think you are loking for vip for keystone service  and also regiona name and username name and password
22:04:38 <bryan_att> yes
22:05:51 <bryan_att> OS_AUTH_URL, OS_TENANT_ID, OS_TENANT_NAME, OS_PROJECT_NAME, OS_USERNAME, OS_PASSWORD, OS_REGION_NAME
22:06:11 <bryan_att> thats all I need to use the OpenStack CLI commands in the install scripts
22:06:55 <narinder> usually at the end we create it and placed under ci/cloud directory
22:07:25 <narinder> and script to create is in ./nosdn/juju-deployer/scripts
22:10:08 <narinder> and this scripts uses JUJU to get the ip of keystone and get it embedded in to the adminrc file
22:10:21 <narinder> script name is openstack.sh
22:11:16 <narinder> unfortunately both intel pod5 and pod6 have started a fresh deployment so can not give you exact commands
22:13:00 <bryan_att> ok, np, I have the file. It looks like the value are hard-coded. Is it acceptable for me to just call that script as part of the install process (bash-based)?
22:13:21 <narinder> but you need to change the networking
22:14:06 <narinder> part as it download the image and upload into cloud and also try to create the network also
22:15:01 <narinder> but if you are using deploy.sh creating adminrc already happening
22:15:20 <bryan_att> I can get the keystone IP address OK. I just needed to know what the other values are. Looks like they are all hard-coded. I can just use the same values, but my script will break if they ever change... but anyway we are moving to charm. I just wanted to simplify the bash install process. So I think I am OK with hard-coding them.
22:15:35 <bryan_att> where is the adminrc that is created?
22:15:55 <narinder> under joid/ci/cloud/
22:16:33 <bryan_att> ok, I see it now. I'll just call that in the script. Assumes I know where it is.
22:16:47 <narinder> ok
22:17:45 <bryan_att> Do you have a standard practice for where you put the joid repo, e.g. at ~/joid ? I have been putting it at ~/git/joid
22:18:00 <bryan_att> (under the ubuntu user)
22:18:15 <bryan_att> just want to make sure I reference it correctly
22:18:49 <narinder> if you are using jenkins then you can find openrc file under ~/joid_config
22:19:29 <narinder> otherwise it can go anywhere based on user choice but from ci directory it is under cloud directory always
22:20:51 <bryan_att> ok, that's a little tricky, not reliably the same folder. So for now I will hard-code the values except for the keystone IP. Simpler.
22:21:14 <narinder> bryan_att: i am sure we have juju way to find it
22:21:59 <narinder> as we can use juju get keystone and look for all config options and then decide dynamically
23:16:36 <bryan_att> narinder: if you
23:17:51 <bryan_att> narinder: are there... I need to access mysql on the horizon lxc but I'm getting "ERROR 1045 (28000): Access denied for user 'root'@'localhost' (using password: NO)" - previously I didn't get this error. Is there a password I need to use?
23:20:52 <narinder> bryan_att: can we pose this query on #juju
23:20:53 <narinder> ?
23:21:20 <bryan_att> sure, I just thought this was an OPNFV-specific thing
23:22:00 <narinder> bryan_att: no i am using generic mysql server and i am sure there willl be beter way to connect or run the query
23:22:41 <bryan_att> narinder: I'm trying to create the database for Congress
23:23:22 <bryan_att> but I can work around this I think by creating my own LXC for Congress, for now
23:23:23 <narinder> bryan_att: yeah for that you do not need to connect. I think you can send your sql command to mysql charm and they will take care of creating it
23:23:42 <bryan_att> narinder: yes, but I'm not there with the charm yet...
23:24:18 <narinder> bryan_att: we use percona-cluster and password set in bundle i think
23:24:30 <narinder> which bundle you deployed?
23:24:58 <bryan_att> narinder: not sure what you mean... this is a plain JOID install for liberty
23:25:12 <narinder> with odl or nosdn
23:25:13 <narinder> ?
23:25:19 <bryan_att> nosdn
23:25:19 <narinder> ans also ha or nonha?
23:25:24 <bryan_att> non-ha
23:26:09 <narinder> ok i do not set it in bundle for non ha
23:26:19 <narinder> but for now you can set it using juju
23:26:30 <narinder> let me give you command and then you can connect
23:28:21 <narinder> juju set mysql/0 root-password=anypassword
23:28:33 <narinder> then you can use anypassword to access the mysql
23:28:58 <narinder> if you need to sey sst-password
23:29:21 <narinder> then juju set mysql/0 sst-password=my-sst-password
23:29:44 <bryan_att> "ERROR "mysql/0" is not a valid service name"
23:29:55 <narinder> may be you have different number
23:30:19 <narinder> can you send me juju status --format tabular
23:30:37 <narinder> also try replacing mysql/0 to mysql
23:30:47 <bryan_att> https://www.irccloud.com/pastebin/zHexsNsB/
23:31:19 <bryan_att> that worked (just mysql)
23:32:03 <narinder> ok yeah it will change it to all instances running my bad
23:32:16 <narinder> it should be mysql only then
23:32:37 <narinder> now you can try to connect to mysql using this root password
23:32:52 <bryan_att> but looks like it didn't work - I still get denied even when I set it to "nopassword"
23:33:04 <narinder> just wait
23:33:18 <narinder> chec juju status mysql/0
23:33:26 <narinder> and see what the status says
23:33:53 <narinder> as it take time all other services has to repolicate and restart to adopt new password
23:34:07 <narinder> juju status --format=tabular should not show execting
23:39:30 <bryan_att> ok, I didn't see it change. let me try creating a new LXC and see if that works.
00:37:06 <narinder> ok
01:50:29 <narinder> fdegir: how do i enable build job for intel pod6 which is my dev server
01:50:35 <narinder> sorry dev pod
13:24:25 <narinder> David_Orange: hi good morning
13:24:34 <David_Orange> Hi
13:25:17 <narinder> David_Orange: i will be checking in the code for DPDK today.
13:25:31 <David_Orange> great
13:25:34 <narinder> David_Orange: and next week for LX| enable_dhcp       | False                                        |
13:25:34 <narinder> D
13:25:43 <narinder> next week for LXD
13:25:55 <David_Orange> good news
13:26:23 <David_Orange> i debug pod2 yesterday but had to reinstall it on stable today
13:26:26 <narinder> i know we still has to work out issues with ODL though
13:26:39 <narinder> is stable working well?
13:26:49 <David_Orange> yes, it blocks me on public api
13:26:54 <David_Orange> it seems
13:26:58 <narinder> ok
13:27:17 <narinder> so we need to see the difference i know one change in SR2 od ODL with latest
13:27:33 <narinder> and need to fogure it out other thing else well in terms of bundle
13:27:36 <David_Orange> we will see next week, ixia is coming here in France to install ixia virtual loader
13:27:44 <narinder> ok
13:27:55 <David_Orange> ok
13:28:15 <narinder> can u pass me bundles.yaml from ci ? so i can compare?
13:30:02 <narinder> i still feel it is SR1 vs SR2 where ODL SR1 works and SR2 has issues
13:32:28 <David_Orange> you need bundles from pod1 ?
13:32:34 <David_Orange> i had issues with both
13:32:39 <narinder> odl bundle where it works
13:32:47 <narinder> oh
13:32:59 <narinder> that means stable is not working?
13:33:12 <David_Orange> first with SR1 then i upgrade to SR2
13:33:38 <narinder> with master branch is it? what about stable branch?
13:33:44 <David_Orange> no, i think i used master bundle template before your switch to SR2
13:33:54 <David_Orange> i use master on pod1
13:34:00 <David_Orange> stable on pod2
13:34:00 <narinder> ok
13:34:12 <narinder> and both pod have issues is it?
13:34:22 <David_Orange> only pod1
13:34:37 <David_Orange> on pod2 it was another issue
13:34:43 <narinder> ok so i need bundles.yaml from pod2 stable
13:34:56 <David_Orange> pod2 is nosdn
13:34:59 <David_Orange> sorry
13:35:02 <narinder> oh ok
13:35:20 <narinder> nosdn works with master as well
13:35:29 <narinder> ok nosdn is not a problem
13:35:34 <narinder> looking for odl
13:36:03 <David_Orange> the only i have is pod1 and i modified it for public api
13:36:47 <narinder> ok
13:37:01 <narinder> no you can keep working on public api
13:37:11 <narinder> i will see if i can reuser intel pods
13:38:24 <narinder> btw i have changed the node naming concention in maas as well to match the physical location
13:40:04 <David_Orange> ok
17:35:42 <bryan_att> narinder: any idea on the issue I reported on #juju? gnuoy seems to be OOO
17:37:50 <bryan_att> narinder: how do I stop a juju deploy in progress? This having to wait for the deploy to reach the next "failed" state then "resolved" etc step by step thru the hopeless process is pretty time consuming. How can I abort a deploy and remove the service so I can start again?
17:38:28 <bryan_att> narindergupta: ping
19:35:13 <narinder> bryan_att: will you please tell me issue again sorry i was in meeting?
20:35:41 <bryan_att> narinder: how do I stop a juju deploy in progress? This having to wait for the deploy to reach the next "failed" state then "resolved" etc step by step thru the hopeless process is pretty time consuming. How can I abort a deploy and remove the service so I can start again?
20:36:08 <bryan_att> narinder: also  any idea on the issue I reported on #juju? gnuoy seems to be OOO
20:37:13 <bryan_att> narinder: related to these charm-congress deploy errors https://www.irccloud.com/pastebin/ozjLJ7Kw/
20:37:15 <narinder> bryan_att: you can always terminate it using juju destroy-service and then run terminate-machine
20:38:06 <narinder> bryan_att: in this looks like few packages are not there?
20:38:24 <bryan_att> narinder: destroy-service doesn't seem to work when the service is in hook-failed state
20:38:45 <narinder> bryan_att: then terminate-machine then the lxc name should do
20:39:14 <narinder> bryan_att: specially --force
20:39:16 <bryan_att> narinder: re the deploy, yes gnuoy claimed ability to deploy and posted a log. But when I try exactly the same sequence I get that error
20:39:35 <narinder> and you are tring the liberty
20:39:56 <bryan_att> narinder: yes, liberty
20:39:58 <narinder> can u run sudo apt-cache policy on congress-server
20:40:18 <narinder> sorry command is sudo apt-cache policy congress-server
20:41:33 <bryan_att> N: Unable to locate package congress-server
20:41:53 <narinder> that is problem
20:42:11 <narinder> can u send me out of la /etc/apt/source.list.d/
20:42:17 <bryan_att> but gnuoy was able to deploy without that error. How?
20:42:43 <bryan_att> ls: cannot access /etc/apt/source.list.d/: No such file or directory
20:42:45 <narinder> lets check. He may have the liberty repos added
20:43:10 <narinder> sorry it is sources.list.d
20:43:13 <narinder> not source
20:43:39 <narinder> cloudarchive-liberty.list should be listed
20:44:05 <bryan_att> yes it's there
20:44:13 <narinder> nope it is not in cloudsrchive-libert
20:44:25 <narinder> do you know where can we find this package?
20:44:33 <bryan_att> https://www.irccloud.com/pastebin/1iX1irSx/
20:45:01 <bryan_att> how do I get the machine-id for the terminate command?
20:45:07 <narinder> not on jumpshot on the system where install issue is there
20:45:33 <narinder> juju status --format=tabular should show the machine id
20:45:38 <narinder> if it is container
20:45:50 <narinder> then it might be 1/lxc/0
20:46:02 <narinder> or something liske that against service
20:46:34 <narinder> looks like congress-server package is not part of liberty repo in ubuntu
20:46:46 <bryan_att> then how did it work for gnuoy
20:47:05 <bryan_att> here is the status output - what is the machine ID for congress?
20:47:07 <bryan_att> https://www.irccloud.com/pastebin/ENy41si5/
20:47:08 <narinder> it must be in ppa some where
20:47:37 <narinder> bryan_att: i need output of tabular format
20:47:47 <narinder> as short only talks about services
20:47:57 <bryan_att> https://www.irccloud.com/pastebin/zl7zO2UH/
20:48:38 <narinder> thats bad looks like you are trying on bare metal itself rather than container
20:48:52 <narinder> so it will be difficult to terminate the machine'
20:49:01 <bryan_att> well it's per the charm-congress, right?
20:49:05 <narinder> as other services are running
20:49:20 <narinder> yes by default with maas provider it is bare metal machine
20:49:33 <bryan_att> The command I used was "juju deploy local:trusty/congress --to 1"
20:49:43 <narinder> that is for bare metal
20:49:55 <bryan_att> what should I use for a container?
20:50:53 <narinder> in bundle we use lxc:nodes=0
20:50:59 <bryan_att> I guess I could use the Horizon container ID? But what if I wanted a new container for Congress?
20:51:33 <narinder> or lxc:nodes=1 depend upon which machine you want
20:51:44 <narinder> i need to check correct command
20:51:58 <narinder> but you can try --to lxc:nodes=0
20:52:20 <bryan_att> what does "nodes=0" mean?
20:52:48 <narinder> i deployed ubuntu services and give name as  nodes
20:52:56 <narinder> and uses two macines
20:53:30 <narinder> nodes=0 means deploy me my server in lxc container on the machine with instance id 0 and service nodes is deployed
20:54:14 <bryan_att> ok, once I get this stopped and removed I will try again with that
20:54:32 <bryan_att> but I need to resolve the package issue firsrt
20:57:27 <bryan_att> I guess everyone is bugging out for the holiday weekend. Understandable, I just want to continue making progress... anyway I'll try again once I get it cleaned up. If you find out anything about the packages or how gnuoy overcam that, let me know if you can.
20:57:40 <bryan_att> gotta go get my lunch out of the oven
20:58:13 <narinder> gnuoy is based in UK so he will be off by now
20:58:28 <narinder> bryan_att: ok but let me check the congress package first
20:59:54 <bryan_att> narinder: sure, np and thanks for your help
21:00:08 <narinder> bryan_att: i figure it out the congress package issue
21:00:34 <narinder> bryan_att: it seems congress is enabled with mitaka only on xenial not with trusty
21:00:48 <narinder> bryan_att: check this https://launchpad.net/ubuntu/+source/congress/3.0.0+dfsg1-1
21:01:07 <narinder> it builds only for xenial
21:02:04 <narinder> bryan_att: so only option for you to use is mitaka on xenial
21:03:09 <narinder> which you can deploy it with joid master branch  ./deploy.sh -o mitaka -s nosdn -t nonha -d xenial -l <your lab name>
21:03:45 <bryan_att> narinder: do I have to have xenial installed on the jumphost?
21:04:07 <narinder> no not neede
21:04:11 <narinder> d
21:04:17 <narinder> you can deploy using current tool
21:04:35 <bryan_att> narinder: ok, I will try that.
21:04:48 <narinder> as maas is ready to deploy the xenial as well now. When did you isntall MAAS?
21:05:51 <bryan_att> a couple of days ago at least
21:06:02 <bryan_att> I can reinstall maas if needed.
21:06:16 <narinder> no need to reinstall MAAS
21:06:25 <narinder> just login into maas page and click on images
21:06:47 <narinder> check whether it has image downloaded for both 14.04 and 16.04?
21:07:07 <narinder> if it does then do not do anything
21:07:21 <narinder> otherwise check the 16.04 and click on apply
21:07:25 <bryan_att> 16.04lts is there
21:07:37 <narinder> and images how many images are there?
21:08:11 <bryan_att> do you mean buttons I can select or something in a list?
21:08:51 <narinder> unders the button selection do you see 16.04 and 14.04 both or only 14.04?
21:09:08 <bryan_att> there is a 16.04 button
21:09:14 <narinder> is it checked?
21:09:31 <bryan_att> I just checked it and hit apply
21:09:49 <bryan_att> it says "Step 1/2: Region importing ."
21:09:59 <narinder> ok sounds good lets wait for completion
21:10:02 <narinder> then clean it
21:10:15 <narinder> and run deploy according to my instructions above
21:10:32 <bryan_att> ok, will do. fingers crossed!
21:10:44 <narinder> i know as this is mitaka with xenial
21:11:05 <narinder> i know mitaka works with 14.04 fine
21:11:17 <narinder> but looks like congress package is not there in trust
15:53:09 <David_Orange> narinder: hi
15:53:26 <narinder> David_Orange: hi David how are you?
15:53:32 <David_Orange> narinder: fine and you ?
15:53:46 <narinder> David_Orange: yeah i am doing good thanks
15:54:02 <David_Orange> i worked today on public api
15:54:49 <David_Orange> i set a script to create a reverse proxy on jumphost for CI/Tests only (where we can not set NAT on the default router)
15:55:24 <David_Orange> it seems to works well on pod1 but now i wish i can try to run functest
15:56:30 <David_Orange> but first i have to reinstall pod1 and be sure i wont have the "No valid host was found" error. Did you progress on that N
15:56:31 <David_Orange> ?
15:59:10 <David_Orange> narinder: i also write an how-to: https://wiki.opnfv.org/display/joid/How+To#HowTo-ConfigureaccesstoPublicAPIonHAmode
16:02:02 <narinder> David_Orange: please try nosdn that should work
16:02:15 <narinder> it seems odl issue is more complicated
16:02:22 <narinder> and i am not getting much help on that
16:02:41 <David_Orange> ok, i will do that tomorrow
16:02:57 <David_Orange> and let you know if functest pass
16:09:59 <narinder> sure that will be great
16:10:37 <narinder> also i am changing few position of variables etc in yaml files to include more support of lxd and other storage options.
16:11:15 <narinder> so in worst case you might need to resumbit the patch after rebasing i am hoping it won't conflict
19:12:51 <fdegir> narinder: ping
19:12:55 <fdegir> narinder: regarding pod6
19:13:01 <fdegir> narinder: you said it is your dev pod
19:13:10 <fdegir> narinder: but it is a CI pod
19:13:12 <narinder> fdegir: hi
19:13:38 <narinder> fdegir: i use it for both and once dev is done then it will be used for CI
19:13:48 <fdegir> ok
19:13:51 <narinder> otherwise i will left with no dev pod
19:13:58 <fdegir> but since we switched using labels
19:14:08 <fdegir> jobs might start on both pods
19:14:34 <narinder> yeah i see that and my dev delpoyment usually cleaned so that may be reason
19:26:58 <fdegir> well, you need to bring this to infra wg
19:33:00 <narinder> fd ok
15:26:31 <narinder> David_Orange: thanks for correcting the password. Can u try to deploy it and see whether it works or not?
15:27:00 <narinder> narinder: also i have implemented to use pass mutiple comma seperated feature during deployment
15:27:02 <David_Orange> narinder: hi, deploy is actually in progress with my publicapi changes
15:27:37 <narinder> ok cool
15:27:44 <David_Orange> narinder: ok great
15:28:34 <narinder> David_Orange: further i will work on removing the node specific data like putting interface name in ext-port i am planning to replace with MA
15:28:38 <narinder> C address
15:29:06 <narinder> and also once maas-deployer will be update then i can remove the auto enabling of interface data in joid
15:30:06 <David_Orange> narinder: once i finish that i will test dpdk, but i think i will also contribute to you new deploy script (add password management, pep8 syntax...) and also think of how to merge multiple sdn controller template, but it will be later
15:30:47 <narinder> David_Orange: thanks for that David and really appreciated the work you are doing
15:30:53 <narinder> with joid
15:32:04 <David_Orange> thanks narinder, the transatlantic sync is not always easy, but i like to contribute on joid :)
15:32:22 <narinder> :)
16:03:42 <collabot`> arturt: Error: Can't start another meeting, one is in progress.  Use #endmeeting first.
16:03:50 <arturt> #endmeeting