#opnfv-pharos: OpenStack 3rd Party CI
Meeting started by fdegir at 14:01:02 UTC
(full logs).
Meeting summary
- Roll Call (fdegir, 14:01:27)
- Julien (Julien-zte,
14:01:32)
- Chris Price (ChrisPriceAB,
14:01:36)
- Markos (hwoarang,
14:01:50)
- Jack Morgan (jmorgan1,
14:01:56)
- Agenda (fdegir, 14:02:19)
- Welcome (fdegir,
14:02:24)
- Background Info (fdegir,
14:02:28)
- Progress: bifrost (fdegir,
14:02:32)
- Progress: puppet-infracloud (fdegir,
14:02:38)
- Infra/HW Needs (fdegir,
14:02:41)
- AOB (fdegir,
14:02:45)
- http://docs.openstack.org/developer/bifrost/readme.html
(Julien-zte,
14:02:50)
- https://wiki.opnfv.org/display/INF/OpenStack+3rd+Party+CI
(Julien-zte,
14:03:05)
- Welcome (fdegir, 14:03:15)
- I want to welcome everyone and thanks for
joining this effort! (fdegir,
14:03:29)
- Background Info (fdegir, 14:04:50)
- There might be questions regarding why we are
working on this (fdegir,
14:05:04)
- We tried to give a summary regarding this
activity on Wiki so please feel free to read, comment, update
(fdegir,
14:05:25)
- https://wiki.opnfv.org/display/INF/OpenStack+3rd+Party+CI
(fdegir,
14:05:35)
- The primary tools we are looking into are
bifrost and puppet-infracloud (fdegir,
14:05:56)
- https://github.com/openstack/bifrost
(fdegir,
14:06:00)
- http://git.openstack.org/cgit/openstack-infra/puppet-infracloud
(fdegir,
14:06:04)
- Progress: bifrost (fdegir, 14:06:41)
- We have been evaluating/trying out bifrost and
fixing what we think could be fixed/improved (fdegir,
14:07:01)
- Yolanda and Markos have been pretty active with
locating and fixing nasty bugs (fdegir,
14:07:25)
- the objective of this activity is to make sure
we can use bifrost to deploy an OPNFV scenario on a "pharos"
lab (ChrisPriceAB,
14:07:54)
- using openstack native CI tools (ChrisPriceAB,
14:08:22)
- We can now provision 3 VMs using bifrost
(fdegir,
14:09:43)
- Ubuntu as OS and the 3 VMs serve as jumphost,
controller, and compute (fdegir,
14:09:57)
- Work to introduce Centos and SuSe support is
ongoing (fdegir,
14:10:08)
- Introducing SuSe support is work in progress.
we are missing the ipxe package which i need to work on. but it's
pretty close after that (fdegir,
14:11:26)
- openSUSE Tumbleweed at first. then Leap. then
SLES (fdegir,
14:11:38)
- The work we've done with bifrost is put on
OPNFV Jenkins and we are provisioning VMs over and over
automatically (fdegir,
14:12:23)
- https://build.opnfv.org/ci/job/infra-provision-virtual-daily-master/lastBuild/consoleFull
(fdegir,
14:12:29)
- This is the script that is run by
Jenkins (fdegir,
14:15:50)
- https://gerrit.opnfv.org/gerrit/gitweb?p=releng.git;a=blob;f=jjb/infra/infra-provision.sh
(fdegir,
14:15:52)
- bifrost contributions since 2016-06-01
(fdegir,
14:18:46)
- http://bit.ly/2c4tgC7 (fdegir,
14:18:47)
- Progress: puppet-infracloud (fdegir, 14:19:31)
- We have been evaluating/trying out
puppet-infracloud and fixing what we think could be fixed
(fdegir,
14:19:56)
- The main pain with the puppet-infracloud is
that the scripts they have for trying stuff have many hardcoded
values in them (fdegir,
14:20:23)
- Yolanda spent quite some time to get rid of
them to make it work for OPNFV (fdegir,
14:20:44)
- Based on her work, we can now install OpenStack
on 2 node setup from stable/mitaka on Ubuntu (fdegir,
14:21:20)
- Putting bifrost and puppet-infracloud together:
How to provision VMs and install openstack (fdegir,
14:23:09)
- Provision VMs using bifrost (via script:
https://gerrit.opnfv.org/gerrit/gitweb?p=releng.git;a=blob;f=jjb/infra/infra-provision.sh)
(fdegir,
14:23:37)
- Once the 3 VMs are provisioned, login to
controller and do puppet apply (fdegir,
14:23:54)
- And then to compute and do puppet apply
(fdegir,
14:24:02)
- Once puppet apply is done on controller and
compute (fdegir,
14:26:00)
- The installation can be verified from
jumphost (fdegir,
14:26:10)
- The puppet-infracloud process for OPNFV has
been documented on Gerrit Releng (fdegir,
14:26:30)
- https://gerrit.opnfv.org/gerrit/gitweb?p=releng.git;a=blob;f=prototypes/puppet-infracloud/README.md
(fdegir,
14:27:08)
- Fix puppet apply --manifests/site.pp
--modulepath=/etc/puppet/modules:/opt/releng/prototypes/puppet-infracloud/modules
(fdegir,
14:28:07)
- puppet apply manifests/site.pp
--modulepath=/etc/puppet/modules:/opt/releng/prototypes/puppet-infracloud/modules
(fdegir,
14:28:14)
- I did the same thing with puppet-infracloud and
here is the link to contributions since 2016-06-01 (fdegir,
14:33:22)
- http://bit.ly/2bRJWQ5 (fdegir,
14:33:30)
- Some might not be for OPNFV so it is up to you
to read them (fdegir,
14:33:51)
- We started with VMs since it is easier to work
with them - at least for me personally (fdegir,
14:37:49)
- But we will soon go to baremetal (fdegir,
14:37:59)
- As noted above, we currently deploy
stable/mitaka on Ubuntu (fdegir,
14:38:14)
- Once the Centos support is introduced and we
can deploy from latest, work to deploy on baremetal will
start (fdegir,
14:38:36)
- Infra/HW Needs for OpenStack 3rd Party CI (fdegir, 14:45:34)
- Our aim is not to fork anything but certain
things can't be upstreamed either (fdegir,
14:46:00)
- So we decided to keep them in OPNFV releng
repo (fdegir,
14:46:10)
- And have a basic flow regarding how to combine
upstream with releng: cp -R (fdegir,
14:46:33)
- https://gerrit.opnfv.org/gerrit/gitweb?p=releng.git;a=tree;f=prototypes
(fdegir,
14:46:42)
- As you see, we have 2 folders there; bifrost
and puppet-infracloud (fdegir,
14:46:56)
- Each of these directories follow same directory
structure as upstream so we can add new files or overwrite existing
files as needed (fdegir,
14:47:19)
- does that process allow OPNFV projects to add
things easily? Can SDNVPN for instance add ODL with a specific
config? (ChrisPriceAB,
14:47:36)
- is that a TBD item, or do they create a
structue in their repo? (ChrisPriceAB,
14:47:54)
- TBD item (fdegir,
14:48:10)
- The current structure just allows us to move on
with not much hassles (fdegir,
14:48:32)
- But needs to be improved definitely when things
become serious (fdegir,
14:48:46)
- We also have some hardware to use for
development work (fdegir,
14:49:13)
- We currently have 4 machines in Intel POD4 to
use for development purposes using VMs (fdegir,
14:49:34)
- hwoarang submitted a jira ticket to Pharos/Jack
to get access to Intel POD4 and he will get pod4-node2 (fdegir,
14:49:55)
- Yolanda uses pod4-jumphost and me
pod4-node1 (fdegir,
14:50:08)
- We also will hopefully get a full POD for
development on baremetal (fdegir,
14:50:43)
- LF POD4 or POD5 - to be discussed during Infra
WG meeting after this one (fdegir,
14:51:03)
- AoB (fdegir, 14:52:05)
- Finally, I want to thank everyone again to see
this happening! (fdegir,
14:54:06)
Meeting ended at 14:54:41 UTC
(full logs).
Action items
- (none)
People present (lines said)
- fdegir (194)
- Julien-zte (36)
- ChrisPriceAB (21)
- jmorgan1 (21)
- hwoarang (12)
- collabot (6)
Generated by MeetBot 0.1.4.