Friday, 2017-05-19

*** thorst_afk has joined #openstack-ansible00:03
*** piliman974 has quit IRC00:06
*** fguillot has joined #openstack-ansible00:07
*** piliman974 has joined #openstack-ansible00:08
*** schwicht has joined #openstack-ansible00:08
*** marst has quit IRC00:10
*** manheim_ has quit IRC00:11
*** thorst_afk has quit IRC00:23
*** schwicht has quit IRC00:23
*** woodard has quit IRC00:30
*** woodard has joined #openstack-ansible00:31
*** thorst_afk has joined #openstack-ansible00:44
*** thorst_afk has quit IRC00:48
*** cuongnv has joined #openstack-ansible00:57
*** portante has quit IRC01:03
*** dmellado has quit IRC01:04
*** portante has joined #openstack-ansible01:08
*** SerenaFeng has joined #openstack-ansible01:08
*** mikal has joined #openstack-ansible01:14
mikalHey, how much do we care about mitaka builds being broken? There's a pip dependancy issue which is stopping them from installing at the moment.01:14
*** cathrich_ has joined #openstack-ansible01:15
*** cathrichardson has quit IRC01:15
*** qiliang28 has joined #openstack-ansible01:20
*** qiliang27 has quit IRC01:21
*** qiliang28 is now known as qiliang2701:21
*** dmellado has joined #openstack-ansible01:22
*** esberglu has joined #openstack-ansible01:22
*** esberglu has quit IRC01:23
*** SerenaFeng has quit IRC01:30
*** zul has joined #openstack-ansible01:31
*** galstrom_zzz is now known as galstrom01:31
*** hw_wutianwei has joined #openstack-ansible01:31
*** SerenaFeng has joined #openstack-ansible01:35
cloudnullmikal: we care.01:40
cloudnullis there something you can point me at that i can go fix?01:40
openstackgerritKevin Carter (cloudnull) proposed openstack/openstack-ansible stable/mitaka: [DNM] Testing builds  https://review.openstack.org/46614201:42
*** fguillot has quit IRC01:44
*** thorst_afk has joined #openstack-ansible01:48
*** redaudit has quit IRC01:49
*** thorst_afk has quit IRC01:53
*** thorst_afk has joined #openstack-ansible01:58
*** thorst_afk has quit IRC01:58
*** smatzek_ has joined #openstack-ansible02:02
*** cpuga has quit IRC02:12
*** cpuga has joined #openstack-ansible02:13
*** markvoelker has joined #openstack-ansible02:15
*** zul has quit IRC02:16
mikalcloudnull: sorry, wandered off for a bit02:22
mikalcloudnull: I would have thought that EOL == not caring any more, but glad you want to fix it02:22
*** duallist has joined #openstack-ansible02:22
mikalcloudnull: http://molokai.stillhq.com/lab/logs/20170518-1430-ostrich-mitaka-a166681/000049-os-cinder-install.gz is a log of a failed install02:23
*** duallist has quit IRC02:23
*** duallist has joined #openstack-ansible02:23
*** smatzek_ has quit IRC02:24
mikalcloudnull: http://paste.openstack.org/show/609976/ is the relevant bit though02:24
mikalpython-keystoneclient fails to install because it wants oslo.i18n, which wants Babel02:24
*** duallist has quit IRC02:25
mikalNo matching distribution found for Babel!=2.4.0,>=2.3.4 (from oslo.i18n>=2.1.0->python-keystoneclient)02:25
mikalCould not find a version that satisfies the requirement Babel!=2.4.0,>=2.3.4 (from oslo.i18n>=2.1.0->python-keystoneclient) (from versions: 2.2.0)02:25
mikalBecause in mitaka Babel is pinned in upper-constraints to 2.2.002:25
mikalIt seems like we're trying to install the master version of python-keystoneclient and it all goes wrong02:26
mikalI can give you access to the environment if you want it02:26
mikalOh, and it worked last week of course02:27
*** smatzek has joined #openstack-ansible02:28
*** winggundamth has joined #openstack-ansible02:28
mikalcloudnull: hmmm, looks like your zuul job is going to pass though, it certainly installed cinder successfully02:29
cloudnullohai.02:30
* cloudnull watching movies.02:30
cloudnullwasn't paying attention.02:30
mikalFair enough02:30
cloudnullbut yea, so long as we have a branch for it I'm down to maintian it / fix it02:30
cloudnull+ I suspect a lot of our deployers are running it02:31
* cloudnull looking at the logs02:31
*** xc9 has joined #openstack-ansible02:33
cloudnullmikal: python_keystoneclient-2.3.2-py2.py3-none-any.whl -- looks like it built uisng that02:33
mikalIn which environment? The gate?02:33
cloudnullthe gate02:34
cloudnullfrom here http://logs.openstack.org/42/466142/1/check/gate-openstack-ansible-openstack-ansible-aio-ubuntu-trusty/e6fed7b/logs/aio1_repo_container-f0b28d82/repo_venv_builder.log02:34
cloudnullwhich tag was the failed build on? or was it the head of the branch ?02:34
mikalOne sec.02:35
mikalfbafe397808ef3ee3447fe8fefa6ac7e5c6ff14402:36
cloudnullthat looks like the head of the branch02:37
mikalYeah, I think it is but I wasn't 100% sure02:37
mikalI do a similar tagging thing that OSA does with the openstack bits02:38
mikalSo, I have a 2.3.2 in my repo server too02:38
mikalroot@aio1-repo-container-1b0397a3:/var/www/repo# find . | grep keystoneclient02:38
mikal./links/python_keystoneclient-2.3.2-py2.py3-none-any.whl02:38
cloudnullso that looks like it converges in the gate job02:41
cloudnullis that mitaka build using the same set of sha's ?02:42
mikalAs in do I tweak the SHAs that OSA uses?02:43
cloudnullyes02:43
cloudnullfrom here https://github.com/openstack/openstack-ansible/blob/fbafe397808ef3ee3447fe8fefa6ac7e5c6ff144/playbooks/defaults/repo_packages/openstack_services.yml02:43
mikalI just use whatever shas are in the one sha I pin to02:43
cloudnullok.02:43
mikalkeystone_git_install_branch: 27793e6df9ddf2ce727b05cce31a000b97897c44 # HEAD of "stable/mitaka" as of 13.04.201702:44
mikalBut oslo.i18n (which is the thing bringing in Babel) isn't pinned in OSA, right?02:44
cloudnullno02:45
cloudnulllooks like in the gate we had babel-2.2.0-py2.py3-none-any.whl02:45
cloudnulland oslo.i18n-3.5.0-py2.py3-none-any.whl02:45
mikalroot@aio1-repo-container-1b0397a3:/var/www/repo# find . | grep i18n02:46
mikal./links/oslo.i18n-3.5.0-py2.py3-none-any.whl02:46
mikalYeah, same there too02:46
*** galstrom is now known as galstrom_zzz02:46
*** qiliang28 has joined #openstack-ansible02:48
mikalI note that oslo.i18n has a unreleased change which removes Babel entirely as a runtime requirement02:49
mikalSo obviously someone else has seen this too02:49
*** qiliang27 has quit IRC02:51
*** qiliang28 is now known as qiliang2702:51
*** thorst_afk has joined #openstack-ansible02:59
cloudnulllooks like these where the wheels used to build eveything http://paste.openstack.org/show/609977/02:59
cloudnullin the gate that is02:59
*** markvoelker has quit IRC03:10
*** markvoelker has joined #openstack-ansible03:11
*** smatzek has quit IRC03:15
cloudnullmikal: do you have the constraint file ?03:15
cloudnullwithin the repo server03:15
*** markvoelker has quit IRC03:15
cloudnullmaybe the absolute requirements file?03:16
cloudnulli wonder if we've got something pinned elsewhere that is allowing the gate to work but not your build?03:16
*** thorst_afk has quit IRC03:19
*** winggundamth has quit IRC03:21
*** winggundamth has joined #openstack-ansible03:23
*** jamielennox is now known as jamielennox|away03:46
*** piliman974 has quit IRC03:47
*** jamielennox|away is now known as jamielennox03:53
mikalSorry, got hauled into a meeting04:01
mikalThe contraints file is archived at http://molokai.stillhq.com/lab/logs/20170518-1430-ostrich-mitaka-a166681/upper-contraints.txt04:01
mikalWhere would I find the absolute requirements file on the repo server?04:02
mikalcloudnull: ^--04:02
*** SerenaFeng has quit IRC04:07
*** jamesdenton has joined #openstack-ansible04:14
*** thorst_afk has joined #openstack-ansible04:15
*** jamesden_ has joined #openstack-ansible04:18
*** jamesdenton has quit IRC04:19
*** thorst_afk has quit IRC04:20
*** jamesden_ has quit IRC04:27
*** jamesdenton has joined #openstack-ansible04:28
mikalcloudnull: what is developer mode? Is that enabled for the gate?04:30
cloudnullthe absolute requirements file is on the repo server under the /var/www/repo/os-releases/$VERSION04:31
cloudnulldeveloper mode is enabled in the individual role gates.04:31
cloudnullnot within the integrated one.04:32
cloudnullthat option tells the role to skip the repo server, pull from source ,and build everything locally04:32
*** jamesdenton has quit IRC04:33
*** chhavi has joined #openstack-ansible04:34
*** winggundamth has quit IRC04:52
*** vnogin has joined #openstack-ansible04:54
openstackgerritKevin Carter (cloudnull) proposed openstack/openstack-ansible-os_nova master: [WIP] Testing using a common SystemD init role  https://review.openstack.org/46617905:04
*** woodard has quit IRC05:12
*** cuongnv has quit IRC05:15
mikalcloudnull: is there a good example somewhere of how to add a new role which just makes an empty container by the way?05:15
mikalcloudnull: I'm having troubles spelunking through the plumbing05:15
*** thorst_afk has joined #openstack-ansible05:16
*** cuongnv has joined #openstack-ansible05:19
cloudnullthe containers are created here https://github.com/openstack/openstack-ansible/blob/master/playbooks/lxc-containers-create.yml05:19
cloudnulland that is done using items from inventory05:19
cloudnullwe doc'd the role create parts here https://docs.openstack.org/developer/openstack-ansible/developer-docs/additional-roles.html05:20
cloudnullif you need a new role05:20
*** thorst_afk has quit IRC05:20
*** retreved has joined #openstack-ansible05:22
cloudnullthere's some docs for env.d and conf.d (maybe in the ops guide) I forget where it is05:22
* cloudnull is tired and cant remember05:22
cloudnullalso in googleing i found https://github.com/midonet/midonet-openstack-ansible -- which is cool -cc jamesdenton05:23
cloudnullidk if there are any midonet people in the channel but it'd be great to chat05:23
cloudnullsome time when im more awake :)05:23
cloudnullanyway, im off.05:24
cloudnullmikal: ping if I can help in any way.05:24
cloudnulli hope you're doing well.05:24
cloudnullttys05:24
*** udesale has joined #openstack-ansible05:24
*** retreved has quit IRC05:26
*** shausy has joined #openstack-ansible05:29
*** pranavb has joined #openstack-ansible05:34
*** SerenaFeng has joined #openstack-ansible05:39
*** pbandark has joined #openstack-ansible05:41
*** pbandark has quit IRC05:45
*** jbadiapa has quit IRC05:53
evrardjpgood morning05:53
kyssemournings05:54
*** cpuga has quit IRC06:02
*** cpuga has joined #openstack-ansible06:02
*** Oku_OS-away is now known as Oku_OS06:04
*** fergus has quit IRC06:06
*** fergus has joined #openstack-ansible06:06
*** jbadiapa has joined #openstack-ansible06:08
*** cpuga has quit IRC06:08
*** markus_z has joined #openstack-ansible06:11
*** thorst_afk has joined #openstack-ansible06:17
*** thorst_afk has quit IRC06:21
*** armaan has joined #openstack-ansible06:23
*** vnogin has quit IRC06:28
*** retreved has joined #openstack-ansible06:34
*** chas has joined #openstack-ansible06:38
openstackgerritJean-Philippe Evrard proposed openstack/openstack-ansible-ops master: [WIP] Make possible to start leap from K  https://review.openstack.org/46365906:38
*** retreved has quit IRC06:38
*** pcaruana has joined #openstack-ansible06:39
*** tobberydberg has joined #openstack-ansible06:43
*** pcaruana is now known as pcaruana|tve|06:44
*** pcaruana|tve| is now known as pcaruana|rtve|06:44
*** weezS has joined #openstack-ansible06:47
*** weezS has quit IRC06:51
*** shardy has quit IRC07:16
*** shardy has joined #openstack-ansible07:17
*** thorst_afk has joined #openstack-ansible07:17
*** pbandark has joined #openstack-ansible07:21
*** thorst_afk has quit IRC07:22
*** armaan has quit IRC07:24
*** chas has quit IRC07:31
mpranjicmorning07:33
*** alezzandro has joined #openstack-ansible07:38
*** chas has joined #openstack-ansible07:41
VAhlGood Morning. do hope this will be a good day when we can solve alot of problem07:42
*** chas has quit IRC07:45
*** retreved has joined #openstack-ansible07:46
*** chas has joined #openstack-ansible07:47
*** manheim_ has joined #openstack-ansible07:51
*** retreved has quit IRC07:51
*** manheim_ has quit IRC07:51
*** manheim_ has joined #openstack-ansible07:52
*** chas has quit IRC08:00
*** chas has joined #openstack-ansible08:00
jrosser_odyssey4me: you around?08:01
*** Disova has quit IRC08:11
*** Disova has joined #openstack-ansible08:11
*** vnogin has joined #openstack-ansible08:17
*** thorst_afk has joined #openstack-ansible08:18
*** alezzandro has quit IRC08:20
*** chas has quit IRC08:23
*** thorst_afk has quit IRC08:23
*** chas has joined #openstack-ansible08:23
*** armaan has joined #openstack-ansible08:29
*** pranavb has quit IRC08:35
*** alezzandro has joined #openstack-ansible08:35
*** pranavb has joined #openstack-ansible08:47
*** alezzandro has quit IRC08:52
pjm6good morning all08:56
*** alezzandro has joined #openstack-ansible09:06
openstackgerritJean-Philippe Evrard proposed openstack/openstack-ansible-ops master: Make possible to start leap from K  https://review.openstack.org/46365909:07
odyssey4memorning all09:09
*** chas has quit IRC09:09
odyssey4mehwoarang whoa! https://review.openstack.org/453130 :)09:11
odyssey4mejrosser_ here now09:11
*** fergus has quit IRC09:12
*** fergus has joined #openstack-ansible09:13
*** electrofelix has joined #openstack-ansible09:13
*** retreved has joined #openstack-ansible09:14
openstackgerritJean-Philippe Evrard proposed openstack/openstack-ansible-ops master: Make possible to start leap from K  https://review.openstack.org/46365909:15
*** fergus has quit IRC09:15
odyssey4meandymccr considering the stable branch, I'm thinking that https://review.openstack.org/#/c/466099/1 should still implement the keymap if it's been overridden, your thoughts?09:18
*** retreved has quit IRC09:18
*** thorst_afk has joined #openstack-ansible09:19
*** fergus has joined #openstack-ansible09:19
odyssey4meI've put my comment in review09:21
*** SerenaFeng has quit IRC09:33
*** thorst_afk has quit IRC09:38
*** os-slack has joined #openstack-ansible09:39
hwoarangodyssey4me: yeah! almost there...09:44
*** fergus has quit IRC09:53
*** fergus has joined #openstack-ansible09:54
*** openstack has joined #openstack-ansible10:00
openstackgerritMarkos Chandras (hwoarang) proposed openstack/openstack-ansible-repo_build master: defaults: Replace spaces with underscores in distro repository  https://review.openstack.org/46624910:02
openstackgerritMarkos Chandras (hwoarang) proposed openstack/openstack-ansible-repo_build master: Add SUSE support  https://review.openstack.org/46625010:02
odyssey4mehwoarang quick edit suggestion for https://review.openstack.org/46624910:05
hwoarangaha10:07
hwoarangok10:07
*** admin0 has joined #openstack-ansible10:08
openstackgerritMarkos Chandras (hwoarang) proposed openstack/openstack-ansible-repo_build master: defaults: Replace spaces with underscores in distro repository  https://review.openstack.org/46624910:08
openstackgerritMarkos Chandras (hwoarang) proposed openstack/openstack-ansible-repo_build master: Add SUSE support  https://review.openstack.org/46625010:08
admin0hi .. one of my controllers crashed, and the lxc containers start but without iP .. rebooting the containers or the server is not helping10:08
admin0i recall there being some rebuild command..anyone remebers that10:09
*** retreved has joined #openstack-ansible10:10
*** retreved has quit IRC10:14
*** stuartgr has joined #openstack-ansible10:17
*** piliman974 has joined #openstack-ansible10:17
jrosser_odyssey4me: you asked yesterday about testing "Support delegation in strategy plugin" on stable/ocata10:17
odyssey4meyep10:17
jrosser_i did that in an aio and it went OK10:17
jrosser_not sure if that is "good enough" test though?10:18
jrosser_not much host-host ssh going on maybe....10:18
odyssey4mewell, I expect that an ocata deployment through a bastion currently won't work because of the service playbook delegation of the db create and rabbitmq vhost creation delegation from the service containers to the galera/rabbitmq containers10:19
odyssey4mewith that plugin, it should work10:19
odyssey4meso if you have an environment/lab that can validate that then that'd be the best test to validate that it not only doesn't break anything, but it adds value10:19
odyssey4methe AIO test tells me that it doesn't break anything, which is great10:20
jrosser_we've got ocata/bastion working, but there might well be a couple of things we've hacked10:24
jrosser_just a mo10:24
*** chhavi has quit IRC10:24
admin0looks like i have to delete all containes there and re-create10:24
andymccradmin0: might be better to try debug why they arent getting ips, using one of them and lxc-attach or some others.10:26
andymccrit sounds like perhaps some service didnt start up properly on the host or the ordering was wrong - may indicate a bug too10:26
admin0it crashed due to out of RAM ..   c11_utility_container-09f0f600              RUNNING 1         onboot, openstack -    -10:26
admin0the IPs are missing in the end10:27
admin0i have restatred the containers and the base server .. but does not seem to help10:27
andymccrafaik there isnt a recreate, im not sure what that would look like :) you could re-run the lxc-create play and it wouldnt create them, it'd just check the confs and update appropriately10:27
admin0stop and start on the log make it zero bytes10:27
andymccris there a diskspace or some other issue?10:27
admin0not a diskspace issue10:27
odyssey4mejrosser_ well, my primary motivation for proposing a backport would be to help cut down your tech debt as your team is the only bunch I know of using the bastion configuration... if it's not a driver for your team then I'm happy to leave it be10:28
admin0enough disk space and inodes ( 23% used on inodes, 30G free on disk space)10:28
admin0and not a read only file system10:28
andymccradmin0:  hmm. can you lxc-attach into a container? even though they dont have ips10:28
admin0yep10:28
admin0network restart inside seems to bring up the IP :D10:29
andymccrand the network config files etc? do they give any info + things like syslog when looking at starting the networking service10:29
andymccrweird that it didnt start automatically though10:29
admin0hmm.. stop start does not bring it up10:30
*** zul has joined #openstack-ansible10:31
admin0hmm..  devices missing10:31
admin0ip link shows for a lot no eth110:31
*** bauruine has quit IRC10:34
*** bauruine has joined #openstack-ansible10:36
*** shardy has quit IRC10:36
*** fergus has quit IRC10:37
*** fergus has joined #openstack-ansible10:37
*** cuongnv has quit IRC10:37
*** shardy has joined #openstack-ansible10:38
jrosser_odyssey4me: will roll our lab setup and test that delegation patch on ocata10:40
odyssey4methanks jrosser_10:45
*** tobberydberg has quit IRC11:09
*** chas has joined #openstack-ansible11:09
*** shardy has quit IRC11:12
*** chas has quit IRC11:13
*** os-slack1 has joined #openstack-ansible11:15
*** os-slack has quit IRC11:15
*** chas has joined #openstack-ansible11:15
*** admin0 has quit IRC11:16
*** zul has quit IRC11:19
*** shardy has joined #openstack-ansible11:19
*** udesale has quit IRC11:19
*** smatzek has joined #openstack-ansible11:24
*** armaan has quit IRC11:25
openstackgerritAndy McCrae proposed openstack/openstack-ansible-os_nova master: [WIP] Move to use UWsgi + upstream NGinx for Nova  https://review.openstack.org/45142511:28
*** schwicht has joined #openstack-ansible11:28
openstackgerritMerged openstack/openstack-ansible master: Add var for environment used only inside playbooks  https://review.openstack.org/46595011:29
*** hamzy has joined #openstack-ansible11:32
*** schwicht has quit IRC11:35
*** thorst_afk has joined #openstack-ansible11:35
*** pmannidi has quit IRC11:37
*** pmannidi has joined #openstack-ansible11:37
*** oneswig has joined #openstack-ansible11:38
*** armaan has joined #openstack-ansible11:38
*** schwicht has joined #openstack-ansible11:39
*** retreved has joined #openstack-ansible11:40
*** retreved has quit IRC11:41
*** retreved has joined #openstack-ansible11:42
*** retreved has quit IRC11:42
*** retreved has joined #openstack-ansible11:43
openstackgerritJesse Pretorius (odyssey4me) proposed openstack/openstack-ansible-ops master: Implement CLI test picker  https://review.openstack.org/46610811:47
openstackgerritMerged openstack/openstack-ansible-ops master: Move keystone and glance to a single file  https://review.openstack.org/46606211:48
*** piliman974 has quit IRC11:49
*** shausy has quit IRC11:50
*** piliman974 has joined #openstack-ansible11:51
*** shausy has joined #openstack-ansible11:53
*** zul has joined #openstack-ansible11:54
VAhlWhen running the task: Lxc Container restart. There is an pre-start vet-cleanup.sh and a post-stop that also runs the veth-cleanup.sh script. That script disable the network interface. Why is that running as a post-stop mode? This cause the next Handler to halt and failed. Since that Handler tries to reach the lxc with ssh.11:55
VAhlThe next task is lxc_container_Create : Wait for ssh to be available11:55
openstackgerritAndy McCrae proposed openstack/openstack-ansible-os_nova master: [WIP] Move to use UWsgi + upstream NGinx for Nova  https://review.openstack.org/45142512:00
mgariepymorning everyone12:01
*** faizy has quit IRC12:04
*** schwicht has quit IRC12:05
*** pmannidi has quit IRC12:07
mhaydenbuenos dias12:11
*** askb has quit IRC12:11
evrardjpgoeiemorgend mhayden mhayden12:14
mgariepyfunny.. https://mail.python.org/pipermail/python-ldap/2017q2/003894.html12:16
*** schwicht has joined #openstack-ansible12:16
*** shausy has quit IRC12:20
*** shardy has quit IRC12:27
*** shardy has joined #openstack-ansible12:27
ArchiFleKshi, evrardjp in https://review.openstack.org/#/c/457629/ you mention a gate fix ? has it been backported to Ocata ? i'm trying to backport this https://review.openstack.org/#/c/465064/12:36
*** fguillot has joined #openstack-ansible12:37
odyssey4meArchiFleKs ah, the stable/ocata gate isn't working for any patches - see https://review.openstack.org/#/q/project:openstack/openstack-ansible-os_designate+branch:stable/ocata12:37
odyssey4mein fact, I think the upstream designate ocata gate may be broken12:38
evrardjpArchiFleKs: there were 2 fixes12:38
evrardjpone was the designate upstream broken12:38
openstackgerritAndy McCrae proposed openstack/openstack-ansible-os_nova master: [WIP] Move to use UWsgi + upstream NGinx for Nova  https://review.openstack.org/45142512:38
evrardjpthe other one was our inventory change12:38
odyssey4meif the upstream gate is broken, then we could try and pin to the last known good SHA12:38
evrardjpwe had to change a become12:38
evrardjpnot sure I backported the become change let me have a look12:39
odyssey4meevrardjp the become change shouldn't be needed - it was for ansible 2.3 which doesn't apply to the ocata branch12:39
evrardjpyes it was for 2.312:39
evrardjpwe don't have 2.3 in ocata?12:39
odyssey4menope - 2.212:39
evrardjpN 2.1 O=2.2 master=2.3 ?12:39
evrardjpoh12:39
odyssey4meyup12:39
evrardjpThat's worth backporting :p12:39
odyssey4mehaha, *stable*12:40
evrardjpI guess 2.2 is still supported indeed :p12:40
odyssey4meodd fail: http://logs.openstack.org/55/460155/1/check/gate-openstack-ansible-os_designate-ansible-func-ubuntu-xenial/c7cd877/console.html.gz#_2017-05-05_07_39_33_33289412:40
evrardjp05 may12:41
evrardjpworld has changed since then12:41
evrardjp:p12:41
odyssey4medoing rechecks - let's get some fresh results for those pending patches12:41
*** pranav has joined #openstack-ansible12:43
odyssey4meit appears that the designate ocata branch is OK, so it's not upstream: https://review.openstack.org/#/q/project:openstack/designate+branch:stable/ocata12:44
*** admin0 has joined #openstack-ansible12:44
*** mgariepy has quit IRC12:46
*** mgariepy has joined #openstack-ansible12:48
openstackgerritMajor Hayden proposed openstack/openstack-ansible-os_nova stable/ocata: Remove default for nova_console_keymap  https://review.openstack.org/46609912:50
*** kylek3h has joined #openstack-ansible12:50
*** aludwar1 has joined #openstack-ansible12:51
*** aludwar1 has quit IRC12:51
mhaydenodyssey4me: good catch on the keymap stuff -- i think i've incorporated those changes12:53
*** admin0 has quit IRC12:54
odyssey4methanks mhayden LGTM12:55
evrardjpwhat was the thing with the keymap?12:57
evrardjpIt's interesting for I18n12:57
openstackgerritMajor Hayden proposed openstack/openstack-ansible master: Add memcached vip for keystone and other services  https://review.openstack.org/45817213:02
*** schwicht has quit IRC13:03
openstackgerritMajor Hayden proposed openstack/openstack-ansible-os_nova master: Pass list for package install  https://review.openstack.org/45895213:05
mhaydenevrardjp: so it was being set to en-us by default13:06
mhaydenbut something was lost in translation between the web interface and the VM13:07
evrardjpmhayden: for the memcached thing we need to pay attention13:07
evrardjpI had patches for handling the deprecation13:07
mhaydenupstream novnc folks said let libvirtd sort it out automagically13:07
mhaydenevrardjp: oh we want to pull memcached entirely?13:07
evrardjpnope13:07
*** jamesdenton has joined #openstack-ansible13:08
evrardjpmemcached is still pretty muche in use for caching what's not token13:08
evrardjpbut we shouldn't configure it that way13:08
evrardjpwe should configure a shared pool with oslo13:08
*** schwicht has joined #openstack-ansible13:08
evrardjpwell your patch doesn't hurt, but I think it's not a good idea13:08
mhaydenso i don't know much from the python/oslo side and using memcached there13:09
mhaydenbut security wise, i was struggling to see a horrible result from what the patch is proposing13:09
evrardjpwell we have to have consistent hashing13:09
*** tobberydberg has joined #openstack-ansible13:09
evrardjpit seems ok13:11
evrardjpI may have to dig deeper.13:11
odyssey4meyeah, for that memcache thing I registered https://launchpad.net/bugs/169075613:11
openstackLaunchpad bug 1690756 in oslo.cache "cache 'backend' argument description is ambiguous" [Undecided,Confirmed]13:11
odyssey4meand lbragstad has pushed https://review.openstack.org/#/c/465631/ to help clarify13:11
odyssey4mequite honestly I'm not convinced that using a LB for memcache is a good idea at all13:11
evrardjpwell I had a look, I had pending patches, sorry if you had to discover this again on your own :/13:12
odyssey4meI'd rather have a library capable of talking to back-ends individually do so if possible, than introduce another HA thing which is not really for HA13:12
odyssey4meyep, evrardjp I saw your patch and reviewed it recently13:12
*** shardy has quit IRC13:12
odyssey4meyou're on the right track there13:12
*** woodard has joined #openstack-ansible13:12
odyssey4meif only you'd update and finish it ;)13:12
evrardjpodyssey4me: there is already a lib that does the spooling13:12
evrardjpwell it takes a significant amount of investment to test this properly too13:13
evrardjpand I'm not running out of work :p13:13
*** shardy has joined #openstack-ansible13:13
odyssey4mehttps://review.openstack.org/455625 is the one you're talking about13:13
*** armaan has quit IRC13:13
*** tobberydberg has quit IRC13:13
odyssey4mealso, considering that the original bug has had no confirmation via testing I'm not convinced about the solution presented either13:14
odyssey4meit needs someone to take the time to try the various options13:14
openstackgerritJean-Philippe Evrard proposed openstack/openstack-ansible master: Move to a all/ group var folder  https://review.openstack.org/46627813:14
*** armaan has joined #openstack-ansible13:14
openstackgerritAndy McCrae proposed openstack/openstack-ansible-os_nova master: [WIP] Move to use UWsgi + upstream NGinx for Nova  https://review.openstack.org/45142513:14
evrardjpI think we need to take time to cleanup this memcached13:14
evrardjphave it properly configured once and for all :p13:14
odyssey4meagreed13:15
evrardjpodyssey4me: gift for you: https://review.openstack.org/#/c/466278/13:15
odyssey4mealso, using two vars for the same thing is confusing as hell13:15
evrardjpyeah but in this case it's not the same thing13:16
evrardjpbut it's confusing.13:16
evrardjpI want to clarify all this13:16
odyssey4meagreed13:16
evrardjpwell it should be the same thing.13:16
odyssey4meif one thing's value is the same as another, but the config implemented in the template is different depending on the context, then we should only have one var13:16
evrardjpyou're talking about the memcached server right?13:17
*** woodard has quit IRC13:17
odyssey4meyes, the patch to put it behind a vip puts the memcache server address into the backend arguments13:17
odyssey4meso we have to vars for different things13:18
odyssey4mequite honestly we don't need most of these vars - we need to hard code the best practise and retire the old vars13:18
*** woodard has joined #openstack-ansible13:18
evrardjpagreed13:18
evrardjplet the deployer override if he wants to13:18
*** schwicht has quit IRC13:19
openstackgerritMerged openstack/openstack-ansible-os_molteniron master: Use package module to install distro packages  https://review.openstack.org/46174313:23
*** piliman974 has quit IRC13:25
*** cathrichardson has joined #openstack-ansible13:26
openstackgerritAndy McCrae proposed openstack/openstack-ansible-os_nova master: [WIP] Move to use UWsgi + upstream NGinx for Nova  https://review.openstack.org/45142513:27
*** cathrich_ has quit IRC13:27
openstackgerritMerged openstack/openstack-ansible-os_nova master: Streamline task execution  https://review.openstack.org/46607213:27
*** zul has quit IRC13:27
*** piliman974 has joined #openstack-ansible13:27
*** foutatoro has joined #openstack-ansible13:28
openstackgerritMajor Hayden proposed openstack/openstack-ansible-galera_client master: Cleanup Galera client role  https://review.openstack.org/46628313:29
*** schwicht has joined #openstack-ansible13:30
*** pranav has quit IRC13:32
*** pranavb has quit IRC13:32
*** esberglu has joined #openstack-ansible13:35
openstackgerritJean-Philippe Evrard proposed openstack/openstack-ansible master: Remove snowflaked keepalived config  https://review.openstack.org/46628413:35
*** piliman974 has quit IRC13:38
*** smatzek has quit IRC13:39
evrardjpodyssey4me: ^ you'll like this one too13:39
*** piliman974 has joined #openstack-ansible13:40
odyssey4meevrardjp very cool - already reviewed :)13:41
openstackgerritMajor Hayden proposed openstack/openstack-ansible-galera_client master: Cleanup Galera client role  https://review.openstack.org/46628313:43
evrardjpneeds time to ensure I don't break stuff with the var_files | default omit.13:43
odyssey4meevrardjp yeah, I was wondering about that one13:44
odyssey4mewhy not use with_first_found, skip:  true13:45
evrardjpnot sure13:45
*** esberglu has quit IRC13:45
*** esberglu has joined #openstack-ansible13:47
*** cpuga has joined #openstack-ansible13:56
*** smatzek has joined #openstack-ansible14:00
*** weezS has joined #openstack-ansible14:00
*** cpuga has quit IRC14:02
*** cpuga has joined #openstack-ansible14:02
*** ThiagoCMC has joined #openstack-ansible14:04
*** galstrom_zzz is now known as galstrom14:04
ThiagoCMCHey guys, I'm trying to use OSA, however, the following command: "openstack-ansible setup-hosts.yml", output "failed", for the "*_neutron_agents_container-*", how to debug this?14:05
ThiagoCMCAll the others are good!14:05
ThiagoCMCMy neutron_agents_container are STOPPED14:06
*** admin0 has joined #openstack-ansible14:06
ThiagoCMCSo, I'm looking for a way to debug this14:06
*** shardy has quit IRC14:08
*** shardy has joined #openstack-ansible14:08
*** schwicht has quit IRC14:09
*** rcarrillocruz has quit IRC14:09
openstackgerritJesse Pretorius (odyssey4me) proposed openstack/openstack-ansible-plugins master: Add group_vars to py_pkgs lookup  https://review.openstack.org/46629514:11
odyssey4meevrardjp cloudnull ^14:11
*** BjoernT has joined #openstack-ansible14:14
evrardjpanswered14:15
cloudnullThiagoCMC: if you do an lxc-ls -f14:18
cloudnullyou see that the contianer is stopped?14:18
ThiagoCMCyep, it is stopped... working on it, got help!   =)14:19
ThiagoCMCthanks!14:19
cloudnullcool14:19
openstackgerritMerged openstack/openstack-ansible-repo_build stable/newton: Make venv build process idempotent  https://review.openstack.org/46553814:20
openstackgerritKevin Carter (cloudnull) proposed openstack/openstack-ansible-os_neutron stable/ocata: Improve the execution time  https://review.openstack.org/46629914:26
openstackgerritKevin Carter (cloudnull) proposed openstack/openstack-ansible-os_nova stable/ocata: Streamline task execution  https://review.openstack.org/46630114:27
mhaydenso the nova role is failing on CentOS 7 due to a kernel call failure :|14:30
evrardjpwow14:30
evrardjpI'm surpised, do you have a link?14:31
mhaydenhttps://gist.github.com/major/8ade5d1187b0831f9ec1abf1cf479f0314:31
odyssey4meevrardjp cloudnull answered the questions in https://review.openstack.org/46629514:32
mhaydengoing to try an older kernel14:34
mhaydenkernel-3.10.0-514.16.1.el7.x86_64 came out the week of Apr 1714:35
openstackgerritAndy McCrae proposed openstack/openstack-ansible-os_nova master: [WIP] Move to use UWsgi + upstream NGinx for Nova  https://review.openstack.org/45142514:37
openstackgerritAndy McCrae proposed openstack/openstack-ansible-os_nova master: [WIP] Move to use UWsgi + upstream NGinx for Nova  https://review.openstack.org/45142514:38
*** schwicht has joined #openstack-ansible14:38
*** manheim_ has quit IRC14:39
mgariepyanyone have ldap domain auth and upgraded from mitaka to newton ?14:48
cloudnullsadly not i mgariepy :(14:51
*** cjloader_ has joined #openstack-ansible14:51
mgariepyit;s weird i'm unable to list my users, first it was a lib issue but now i'm not too sure what is going on..14:52
*** cjloader_ has quit IRC14:52
*** cjloader_ has joined #openstack-ansible14:52
cloudnullmaybe the lib is breaking compatibility ?14:53
cloudnullcan you do a list manually?14:53
odyssey4mealso seeing centos fails due to epel unavailability: http://logs.openstack.org/95/466295/1/check/gate-openstack-ansible-plugins-ansible-func-centos-7/f91f179/console.html#_2017-05-19_14_13_49_11881214:53
mgariepyi can do ldap search..14:53
mgariepyodyssey4me, did they remove epel from their mirror ?14:54
cloudnullodyssey4me: maybe the cent mirror is out of sync/broken again ?14:54
odyssey4meah, it's not just us - asked in infra14:55
openstackgerritJean-Philippe Evrard proposed openstack/openstack-ansible master: Remove snowflaked keepalived config  https://review.openstack.org/46628414:55
*** gouthamr has joined #openstack-ansible14:56
openstackgerritJean-Philippe Evrard proposed openstack/openstack-ansible master: Remove snowflaked keepalived config  https://review.openstack.org/46628414:56
evrardjpodyssey4me: yeah everything centos is broken it seems :p14:57
cloudnullevrardjp: you mean "everything is awesome!"14:58
*** exoduswtf has joined #openstack-ansible14:58
odyssey4meevrardjp cloudnull I've posed a though that is perhaps wise in https://review.openstack.org/46629515:00
evrardjpodyssey4me: won't change my vote, lookup is like a black box for deployers. We should keep it that way and not have a weird non-ansibly behavior15:00
*** schwicht has quit IRC15:00
evrardjpnah15:00
odyssey4meevrardjp it is already weird behaviour15:00
evrardjpgroup_vars/repo_build15:00
*** markus_z has quit IRC15:01
evrardjpI think we should keep it group_vars/15:01
*** manheim_ has joined #openstack-ansible15:01
evrardjpor maybe we should drop the lookup completely15:01
odyssey4meif you have a way of doing so, then I'd love to see the proposal15:03
evrardjpit involves a lot of RY15:04
*** exodusftw has quit IRC15:07
admin0is there a way to run/complete setup-openstack.yml but ignore one dead controller .. so that it works only with the remaining 215:07
*** ianychoi has quit IRC15:07
*** armaan has quit IRC15:08
evrardjplimit15:08
mgariepylogan-, are you using ldap ?15:09
evrardjpadmin0: but you should probably fix your controller instead15:09
admin0evrardjp that is true .. but i am also on a deadline to have a working OSA+OVS setup15:10
admin0which I find very hard15:10
*** ianychoi has joined #openstack-ansible15:11
admin0the containers need a LB  right .. tey cannot plug-in to an OVS bridge ?  .. beause neutron contailers fail15:11
evrardjpodyssey4me: what I meant is, instead of doing this live with a lookup we can generate the file from the defaults/ vars/ etc. into group_vars/ for OSA usage.15:11
admin0as the ansible tries to plugin to a bridge that is OVS15:11
evrardjpI don't get a word you say15:12
odyssey4meevrardjp I don't get a word you say ;)15:12
evrardjp:p15:12
evrardjpI'm from a different planet.15:12
odyssey4meBelgium15:13
odyssey4me:)15:14
*** karimb has joined #openstack-ansible15:17
*** schwicht has joined #openstack-ansible15:18
*** exoduswtf has quit IRC15:19
*** gouthamr has quit IRC15:19
*** marst_ has joined #openstack-ansible15:19
*** exodusftw has joined #openstack-ansible15:19
*** armaan has joined #openstack-ansible15:20
*** armaan has quit IRC15:24
*** armaan has joined #openstack-ansible15:24
*** armaan has quit IRC15:25
*** armaan has joined #openstack-ansible15:26
*** vnogin has quit IRC15:26
cloudnullthe land of waffles15:28
*** gouthamr has joined #openstack-ansible15:28
jamesdentonOVS!15:31
cloudnullthe future is now!15:32
mgariepy^^ it's actually the past..15:32
*** armaan has quit IRC15:34
*** manheim_ has quit IRC15:35
*** manheim_ has joined #openstack-ansible15:36
cloudnullwhat's old is new again15:41
*** BjoernT has quit IRC15:42
logan-mgariepy: nope. just using sqlbackend and orchestrating users/projects via keystone api15:47
*** Oku_OS is now known as Oku_OS-away15:48
mhaydencloudnull: we need to make openstack-ansible-hoof15:50
mhaydenfor mattt15:50
odyssey4meaoh15:51
odyssey4menot be be confused with aodh15:51
*** foutatoro has quit IRC15:52
*** esberglu has quit IRC15:57
*** esberglu has joined #openstack-ansible15:58
*** armaan has joined #openstack-ansible16:00
*** esberglu has quit IRC16:03
*** manheim_ has quit IRC16:06
*** admin0 has quit IRC16:08
openstackgerritJesse Pretorius (odyssey4me) proposed openstack/openstack-ansible-os_designate master: Perform an atomic policy file change  https://review.openstack.org/46109516:12
openstackgerritJesse Pretorius (odyssey4me) proposed openstack/openstack-ansible-os_monasca stable/ocata: Use package module to install distro packages  https://review.openstack.org/46582116:12
openstackgerritJesse Pretorius (odyssey4me) proposed openstack/openstack-ansible-os_aodh stable/ocata: Perform an atomic policy file change  https://review.openstack.org/46167016:12
openstackgerritJesse Pretorius (odyssey4me) proposed openstack/openstack-ansible-os_keystone master: Use uri module for git sourced configs  https://review.openstack.org/46339016:12
openstackgerritJesse Pretorius (odyssey4me) proposed openstack/openstack-ansible-os_ceilometer master: Source template files from git or deploy host  https://review.openstack.org/46398816:12
*** oneswig has quit IRC16:17
*** oneswig has joined #openstack-ansible16:17
*** vnogin has joined #openstack-ansible16:18
*** vnogin has quit IRC16:22
*** oneswig has quit IRC16:22
*** hogwash has joined #openstack-ansible16:24
*** manheim_ has joined #openstack-ansible16:26
*** manheim_ has quit IRC16:27
*** weezS has quit IRC16:34
*** weezS has joined #openstack-ansible16:34
*** gouthamr_ has joined #openstack-ansible16:36
*** rcarrillocruz has joined #openstack-ansible16:38
*** gouthamr has quit IRC16:38
*** pcaruana|rtve| has quit IRC16:42
*** fergus has quit IRC16:44
*** fergus has joined #openstack-ansible16:44
*** hw_wutianwei1 has joined #openstack-ansible16:46
*** hw_wutianwei has quit IRC16:49
*** hw_wutianwei1 is now known as hw_wutianwei16:49
*** esberglu has joined #openstack-ansible16:49
*** weezS has quit IRC16:50
*** weezS has joined #openstack-ansible16:51
*** alezzandro has quit IRC16:51
*** markvoelker has joined #openstack-ansible16:52
*** cpuga has quit IRC16:53
*** hw_wutianwei1 has joined #openstack-ansible16:57
*** hw_wutianwei has quit IRC17:00
*** hw_wutianwei1 is now known as hw_wutianwei17:00
openstackgerritMerged openstack/openstack-ansible-os_aodh stable/ocata: Perform an atomic policy file change  https://review.openstack.org/46167017:01
openstackgerritMerged openstack/openstack-ansible-os_designate master: Perform an atomic policy file change  https://review.openstack.org/46109517:05
*** chas has quit IRC17:06
*** chas has joined #openstack-ansible17:07
*** electrofelix has quit IRC17:10
*** chas has quit IRC17:11
openstackgerritMerged openstack/openstack-ansible-os_keystone master: Use uri module for git sourced configs  https://review.openstack.org/46339017:14
*** armaan has quit IRC17:16
*** admin0 has joined #openstack-ansible17:16
xdfilhey logan- are the group_vars meant to be modified in their current location or should I be putting modifications in user_variables.yml ?17:19
*** retreved has quit IRC17:22
logan-modifying group_vars in tree can become a maintenance pita because you'll have to merge your local changes every time you pull in updated osa code. user_variables.yml scoping is mostly similar to modifying group_vars/all.yml, but if you want to modify specific groups' vars besides "all" and want to maintain it long term then id look at other options than doing in-tree modifications17:22
logan-you can sort-of set group-scoped vars into inventory using env.d overrides. another option is overlaying osa and using a separate inventory directory that ties back into the osa tree17:25
evrardjplogan-:  xdfil there are now user level group vars17:26
evrardjpyou can create them in /etc/openstack_deploy/group_vars17:26
evrardjphowever, currently, a limitation prevents from overriding existing group_vars17:27
xdfillogan- ty. yeah I was asking because currently I put everything under /etc/openstack_deploy in my local git repo17:27
*** shausy has joined #openstack-ansible17:27
evrardjpso if you need to put these on top of existing group vars, that's fine17:27
evrardjpelse you should use user_* vars17:27
xdfilevrardjp: now? which tag did that start with?17:27
evrardjpI don't remember when I did that17:28
xdfilis it in N ?17:28
evrardjplatest N I think yes17:28
xdfilty17:28
evrardjpearly N definitely no17:28
evrardjphave a look at your openstack-ansible.rc17:28
evrardjpyou'll find the group_var folders :p17:29
evrardjpit's not a public feature yet, because in general user_* overrides are enough17:29
evrardjpI think user_* overrides should only be used whenever necessary. To really show "I know what I am doing".17:30
xdfilexport GROUP_VARS_PATH="${GROUP_VARS_PATH:-/etc/openstack_deploy/group_vars/}"17:30
xdfillooks like I am good to go17:30
evrardjpYou should try at least :p17:30
evrardjpremember the limitation for now.17:30
evrardjpcan't override existing group_vars vars and values.17:31
evrardjpyou can only add yours for now.17:31
evrardjpBut that may change depending on the user experience :)17:31
xdfilso I can copy /opt/... into /etc/... and at least I can keep them in git just need to remember to delete them when I pull17:32
evrardjpIf you like it, give us a shout.17:32
evrardjpwhy would you want that?17:32
xdfiloh maybe I am misunderstanding17:32
*** shausy has quit IRC17:32
xdfilor you mean why do I want to modify them?17:33
evrardjpI thought you wanted to have your own group vars17:33
evrardjpso, in-tree stays in-tree17:33
xdfilI need to modify " openstack_host_distro_packages:17:33
evrardjpand /etc/openstack_deploy are yours alone :)17:33
xdfilI need to prevent lvm2 from being installed17:34
evrardjpok17:34
evrardjpso17:34
evrardjpopenstack_host_distro_packages isn't in code group vars right now.17:34
evrardjptwo choices:17:34
xdfilI am thinking of submitting a bug about it17:35
evrardjpyou use a user_* variable (what we have always documented)17:35
evrardjpor you use a group_vars/<your group name, for example hosts>/filename.yml17:35
xdfilI could be wrong, but I think having lvm2 installed on compute hosts is bad17:36
evrardjpwell if we give a wrong variable for everyone, that should be a bug17:36
evrardjpwhy ?17:36
evrardjpok I have to go for today17:36
openstackgerritMerged openstack/openstack-ansible-os_ceilometer master: Source template files from git or deploy host  https://review.openstack.org/46398817:36
evrardjpbut please register a bug if you think it can help the community17:36
xdfilbecause when guest volumes are attached via iscsi, the compute host will "import" those lvm volumes17:37
evrardjpand your concern can be done with user_* for now :p17:37
xdfilI tried to blacklist, but I can't come up with a strategy to blacklist cinder volumes effectivley17:37
lbragstadrunning this by some of the folks here again - https://review.openstack.org/#/c/438701/1717:38
xdfilso what ends up happening if I have an lvm volume on a guest, things like migration start to have problems17:38
xdfilalso apt upgrades fail because grub triggers see multiple boot directories and things hang17:39
xdfilsame thing with ansible fact gathering17:39
xdfilswap volumes17:40
*** fguillot has quit IRC17:40
xdfilthe host thinks the guest swap volumes are fair game17:40
xdfilbut at the same time i am skeptical that I am the one "discovering" this17:42
xdfilmakes me think I'm being dumb about something17:42
*** vnogin has joined #openstack-ansible17:42
*** karimb has quit IRC17:43
*** lbragstad has quit IRC17:47
cineramawhat's the best way to do a git clone of all the roles?17:50
logan-thats awesome evrardjp17:51
logan-missed that commit till now17:53
odyssey4melogan- slowly but surely we will get rid of the *requirement* to use extra vars17:54
*** gouthamr_ has quit IRC17:55
*** armaan has joined #openstack-ansible17:55
logan-yep thats a great change. it looks like the vars plugin does a key-based merge of overlapping files?17:55
logan-ie dont have to override the whole file17:55
*** galstrom is now known as galstrom_zzz17:56
odyssey4melogan- it's unfortunately a lower precedence than the inventory group_vars, so you can add overrides for any defaults... but not anything in group_vars already17:59
odyssey4meevrardjp suggested moving our group_vars out of the inventory and into the repo root, allowing us to change that precedence18:00
logan-ah18:00
logan-interesting18:00
odyssey4mesee http://docs.ansible.com/ansible/playbooks_variables.html#variable-precedence-where-should-i-put-a-variable18:00
*** poopcat has joined #openstack-ansible18:01
*** vnogin has quit IRC18:01
*** galstrom_zzz is now known as galstrom18:01
odyssey4me'inventory INI or script group vars' is a lower precedence than group_vars/*18:01
logan-so the vars plugin prcedence falls under "script group vars"18:01
logan-gotcha18:01
odyssey4meyup18:01
odyssey4meso it merges with the dynamic inventory, effectively18:02
odyssey4meso this is a neat replacement for any of the env.d nonsense18:02
odyssey4mewe haven't made a noise about it yet because we're still doing a trial with RPC18:02
odyssey4mebut we discussed yesterday actually doing the group definitions using the INI format as far as possible18:03
*** cjloader_ has quit IRC18:03
logan-well at least it seems like no precedence change would take place by moving group_vars out of the inventory path and into a vars_plugin owned path18:04
*** vnogin has joined #openstack-ansible18:04
logan-has anyone proposed the patch yet? :)18:05
evrardjpI wanted to do it lgo18:10
evrardjplogan-: *18:10
evrardjphowever I think someone wanted to wait for next cycle18:10
*** vnogin has quit IRC18:10
evrardjpelse I can do it now, I have already a large cleanup in my mind :p18:10
evrardjpinteresting ping of lbragstad18:11
logan-i was just thinking of pushing a WIP to see how it works. id be curious to see if we can get inventory/group_vars moved to defaults/group_vars and how relative paths work with your vars plugin18:11
logan-i guess the path should be relative to the inventory18:12
evrardjpno18:14
evrardjpfor readability I think it's user to put it at the same level as playbooks, on the root :)18:14
*** Jeffrey4l has quit IRC18:14
evrardjpwhat I mean by no is the path relative to the inventory18:14
evrardjplet me rephrase that, it's not clear :)18:15
evrardjpas of ansible 2.1 (and 2.2), the loading of group_vars happened after the vars_loading18:15
evrardjpthey are looking into inventory/group_vars18:15
evrardjpso if we put our "group_vars" folder into something else than inventory/group_vars, it would make our life easier18:16
evrardjpwe just need to change the plugin if we change the default path18:16
evrardjpor just change the env var18:16
logan-yeah i'm thinking move the group_vars to some other path in the osa tree and introduce that path into openstack-ansible.rc as a relative path. ideally the plugin will be made to parse the env var paths relative to the ansible inventory path18:17
evrardjpmmm that's technically possible too.18:18
logan-i dont want to see the in-tree group_vars become an absolute path when openstack-ansible.rc is rendered, thats my main concern18:18
evrardjpbut why not setting the env var directly?18:18
*** cjloader_ has joined #openstack-ansible18:18
logan-the env var is defaulted in the .rc18:18
evrardjpand?18:19
odyssey4meyeah, group_vars in the repo root is very visible and easy to find18:19
logan-if we're moving group_vars out of the default path we need to introduce it into the .rc as part of the default paths right?18:19
odyssey4meit'll cause less confusion18:19
evrardjplogan-: yes18:19
odyssey4meyep18:19
*** gouthamr has joined #openstack-ansible18:19
logan-so the new default path would be something like ../defaults/group_vars:/etc/openstack_deploy/group_vars/18:19
evrardjplogan-: doing something like https://github.com/openstack/openstack-ansible/blob/master/scripts/bootstrap-ansible.sh#L14518:20
odyssey4methen we can also change the precedence - ie load the in-tree group_vars, then load the user group_vars18:20
logan-(ie if the in-tree default group vars path was now osa/playbooks/defaults/group_vars)18:20
odyssey4meyes18:20
*** Jeffrey4l has joined #openstack-ansible18:20
logan-looking the vars plugin its not clear to me whether it searches for paths relative to the inventory if it only accepts absolute paths18:21
evrardjpwhat if we speak about that over next week's meeting?18:21
odyssey4melogan- afaik it's absolute paths18:21
*** cpuga has joined #openstack-ansible18:22
evrardjpwe can definitely change that if you want18:22
odyssey4meI think I prefer absolute paths. It's unambiguous.18:22
*** retreved has joined #openstack-ansible18:22
odyssey4meDid you see what I did there? ;)18:22
logan-lol18:22
*** cpuga has quit IRC18:23
odyssey4melogan- if you look at https://github.com/rcbops/rpc-openstack it's easy to see where the group vars are18:23
*** cpuga has joined #openstack-ansible18:23
odyssey4meso anyone coming in blind, knowing ansible, will easily find it and know what it's for18:23
logan-yeah18:24
odyssey4methe placement will be odd, but the purpose is clear18:24
odyssey4mehaving to dig into playbooks/defaults/ to find them seems like a bad user experience for me18:24
logan-i like that too18:24
odyssey4mewe have a hard enough time trying to explain how the env.d stuff works18:24
logan-yeah im open to either. as long as it isn't playbooks/inventory/group_vars or playbooks/group_vars we're good. :D18:24
evrardjpgood :)18:25
odyssey4meevrardjp that's a green light to propose the path18:25
odyssey4me*path18:25
odyssey4medammit18:25
odyssey4mepatch*18:25
evrardjpfullpath?18:26
evrardjphaha ok18:26
evrardjpwell if possible, that would be SUPER cool to merge the other move of vars18:26
odyssey4meit's not even milestone 2 yet, I'll hold your beer18:26
evrardjphaha ok18:27
odyssey4meyou mean https://review.openstack.org/#/c/466278/ ?18:27
evrardjpI'll stop eating my chicken and patch.18:27
*** redaudit has joined #openstack-ansible18:27
* odyssey4me gets a beer18:27
evrardjpyeah and https://review.openstack.org/#/c/466284/18:27
logan-yeah so only big downside i have with putting it in /group_vars or etc/group_vars is deployers thinking about modifying it in tree18:27
*** retreved_ has joined #openstack-ansible18:27
evrardjpwhat!18:27
evrardjpI did something wrong apparently.18:27
*** retreved has quit IRC18:27
odyssey4melogan- that's a problem we have anyway18:28
evrardjpI'll chmod it 444 I promise18:28
evrardjp:p18:28
*** galstrom is now known as galstrom_zzz18:30
openstackgerritLogan V proposed openstack/openstack-ansible master: [WIP] Relocate group vars out of inventory path  https://review.openstack.org/46637818:30
openstackgerritMajor Hayden proposed openstack/openstack-ansible-galera_client master: Cleanup Galera client role  https://review.openstack.org/46628318:31
logan-now to get a depends-on for relative paths support in -plugins :P18:31
openstackgerritLogan V proposed openstack/openstack-ansible master: [WIP] Relocate group vars out of inventory path  https://review.openstack.org/46637818:32
*** armaan has quit IRC18:33
*** stuartgr has left #openstack-ansible18:34
odyssey4melol, evrardjp logan- beat you to it18:34
*** galstrom_zzz is now known as galstrom18:34
openstackgerritJean-Philippe Evrard proposed openstack/openstack-ansible master: [WIP] Use root folder group_vars  https://review.openstack.org/46637918:34
evrardjpnetwork connection18:34
evrardjpmy paaaaaatch!18:35
*** Jeffrey4l has quit IRC18:36
odyssey4merace of the patches18:37
openstackgerritJean-Philippe Evrard proposed openstack/openstack-ansible master: [WIP] Use root folder group_vars  https://review.openstack.org/46637918:38
logan-haha18:38
evrardjpnow with reno18:38
logan-noooo the absolute path18:40
evrardjpI will remove it later18:40
evrardjpdon't worry :p18:40
evrardjpor please review :)18:40
*** zul has joined #openstack-ansible18:41
logan-sure18:41
odyssey4meyou can sed it like we do for the inventory location18:42
openstackgerritJean-Philippe Evrard proposed openstack/openstack-ansible master: [WIP] Use root folder group_vars  https://review.openstack.org/46637918:42
evrardjpodyssey4me: yeah that's what I wrote a little above18:42
odyssey4meie https://github.com/openstack/openstack-ansible/blob/master/scripts/bootstrap-ansible.sh#L14518:43
odyssey4meexcept use https://github.com/openstack/openstack-ansible/blob/master/scripts/bootstrap-ansible.sh#L4718:43
*** Jeffrey4l has joined #openstack-ansible18:43
logan-personally would prefer to see relative paths supported in the plugin18:43
odyssey4melogan- what happens if you're providing inventory via -i18:44
*** ThiagoCMC has quit IRC18:45
logan-the correct thing imo ;)18:45
logan-hmm18:46
logan-maybe you're right18:47
odyssey4methe fixed path in openstack-ansible.rc in-tree is bad, but if we set it when ansible is bootstrapped like we do the inventory then I think that's good18:48
evrardjpI'll let you think of it, in the meantime I'm cleaning up the rest. I don't remember how I wrote the plugin, but I think it's loading the path with what you give him, so relative paths should work18:48
evrardjpodyssey4me: I agree.18:48
logan-yeah18:48
odyssey4melogan- FYI your use of 'default' as the parent directory is something I'm warming to18:48
logan-i agree now too18:48
evrardjpwant a quick fix, or want to reach consensus with the ../ ?18:48
odyssey4meevrardjp logan- see my review of logan-'s patch18:48
evrardjpok I'll add docs18:49
mpranjicevening18:50
odyssey4meevrardjp I mean the comment on the path itself18:50
odyssey4methe docs are important, yes, but I think logan- may be on to something with the path18:50
mpranjicodyssey4me: looking at https://review.openstack.org/#/c/466284/ ...18:50
mpranjic2017-05-19 15:01:58.848238 | fatal: [localhost]: FAILED! => {"changed": true, "cmd": ["grep", "/openstack/swap.img", "/proc/swaps"], "delta": "0:00:00.003478", "end": "2017-05-19 15:01:58.831878", "failed": true, "rc": 1, "start": "2017-05-19 15:01:58.828400", "stderr": "", "stderr_lines": [], "stdout": "", "stdout_lines": []}18:50
odyssey4meI'm on the fence.18:50
mpranjic2017-05-19 15:01:58.873638 |18:51
odyssey4mempranjic pastebin things like that please18:51
mpranjicodyssey4me: yeah, sorry about that...18:51
odyssey4menote that's a normal failure in bootstrap-aio if it's a first time execution18:51
*** cjloader_ has quit IRC18:51
odyssey4meyou'll notice a try/rescue block18:51
odyssey4mewe should ideally do something else there that doesn't cause a failed task18:51
*** cjloader_ has joined #openstack-ansible18:52
evrardjpI'm not sure to understand what you mean odyssey4me18:52
odyssey4meevrardjp so the path of <root>/defaults/group_vars conveys that the values in there are default values18:53
odyssey4meit's subtle18:53
odyssey4mea part of me likes it18:53
evrardjpwell then you don't know what this represents from reaching the path18:53
odyssey4mehowever a part of me thinks that anything in the git tree is a default value, surely people realise that18:53
logan-yeah i figured it was a good compromise from burying them in playbooks18:53
odyssey4mebut part of me also thinks that putting them into a sub folder makes them harder to find18:53
evrardjpI can add a disclaimer into each file, which I'd prefer :p18:54
odyssey4meSure, I'm down with that.18:54
odyssey4mesomething similar to https://github.com/openstack/openstack-ansible/blob/master/etc/openstack_deploy/user_variables.yml#L16-L1918:54
*** cathrichardson has quit IRC18:55
*** cathrichardson has joined #openstack-ansible18:55
openstackgerritJean-Philippe Evrard proposed openstack/openstack-ansible master: [WIP] Use root folder group_vars  https://review.openstack.org/46637918:58
*** smatzek has quit IRC18:58
openstackgerritJean-Philippe Evrard proposed openstack/openstack-ansible master: [WIP] Use root folder group_vars  https://review.openstack.org/46637919:01
openstackgerritMerged openstack/openstack-ansible-repo_build master: defaults: Replace spaces with underscores in distro repository  https://review.openstack.org/46624919:02
*** jeh has quit IRC19:03
openstackgerritMajor Hayden proposed openstack/openstack-ansible-security master: Initial Debian 8 support  https://review.openstack.org/46638619:05
openstackgerritJean-Philippe Evrard proposed openstack/openstack-ansible master: [WIP] Use root folder group_vars  https://review.openstack.org/46637919:06
*** marst_ has quit IRC19:09
*** marst_ has joined #openstack-ansible19:10
*** oneswig has joined #openstack-ansible19:12
*** marst_ has quit IRC19:13
*** marst_ has joined #openstack-ansible19:13
*** oneswig has quit IRC19:17
*** jeh has joined #openstack-ansible19:19
odyssey4meevrardjp I've validated that everything is there in https://review.openstack.org/466278 so once there've been a few more reviews I'm happy to make it go19:20
logan-nice19:21
evrardjpok19:21
evrardjplet me finish my docs first.19:21
evrardjp:p19:21
*** galstrom is now known as galstrom_zzz19:21
*** vnogin has joined #openstack-ansible19:24
openstackgerritJean-Philippe Evrard proposed openstack/openstack-ansible master: [WIP] Use root folder group_vars  https://review.openstack.org/46637919:27
*** galstrom_zzz is now known as galstrom19:31
openstackgerritJean-Philippe Evrard proposed openstack/openstack-ansible master: Move to a all/ group var folder  https://review.openstack.org/46627819:36
openstackgerritJean-Philippe Evrard proposed openstack/openstack-ansible master: Move to a all/ group var folder  https://review.openstack.org/46627819:36
mhaydenso those kernel panics aren't specific to Xen19:39
mhaydenit happens on KVM, too19:39
mhayden(the os_nova centos 7 failures)19:39
openstackgerritJean-Philippe Evrard proposed openstack/openstack-ansible master: Remove snowflaked keepalived config  https://review.openstack.org/46628419:40
odyssey4memhayden oh nice :)19:40
mhaydengdb shows that it's blowing up just as it does the final handoff to qemu19:41
openstackgerritJean-Philippe Evrard proposed openstack/openstack-ansible master: Avoid jinja2 templating in when warnings  https://review.openstack.org/46639119:45
*** shardy has quit IRC19:46
evrardjpoh a -i that stayed in my keyboard.19:46
*** kylek3h has quit IRC19:49
*** kylek3h has joined #openstack-ansible19:50
odyssey4meok, I'm out for the night and probably for the w/end19:50
odyssey4mehave a great w/end all :)19:50
evrardjphave a good week-end!19:51
cloudnulllater odyssey4me19:51
evrardjpI am doing the same.19:52
evrardjpHave a good week-end!19:52
logan-you too19:52
cloudnullcheers evrardjp.19:52
cloudnullhave a stella and chilax evrardjp, you earned it19:52
cloudnull:P19:52
evrardjphaha19:52
mhaydenwell played, cloudnull19:52
* cloudnull takes a bow19:53
* mhayden plans to snag some bourbon on the way home :)19:53
evrardjpI'm gonna stab you with a bow, yes!19:53
evrardjphaha19:53
evrardjpyou mean poor man's whisky?19:53
odyssey4melol19:53
evrardjpI say whisky and you're already back odyssey4me.19:54
evrardjpSurprise surprise19:54
odyssey4meyou're speaking the language of my people19:54
mhaydenhah19:56
evrardjpThat's weird I can't find the initial reason I thought the string was needed in when clauses19:57
evrardjpmaybe it's just for asserts than.19:57
evrardjpStill I gave an example19:57
logan-sometimes it is needed19:57
logan-like if you're doing when: 'test' in some_list19:57
logan-it has to be quoted19:57
evrardjpIIRC we should always do it19:57
evrardjpyeah " 'itemname' in listname"19:57
evrardjpI gave a link to an example in the official docs19:58
evrardjpbut nothing mentions if we should do19:58
*** cpuga has quit IRC19:58
evrardjpwhen: 'itemname' in listname or when: "'itemname' in listname"19:59
logan-first one will error i think19:59
evrardjpyeah probably19:59
evrardjpanyway. Always make it a string in ansible. It all goes through the "Horrible machine"20:00
evrardjpbut I'm off for this week.20:00
evrardjpSee you next week!20:00
*** galstrom is now known as galstrom_zzz20:01
*** alezzandro has joined #openstack-ansible20:01
*** hogwash has quit IRC20:06
*** galstrom_zzz is now known as galstrom20:09
*** tobberydberg has joined #openstack-ansible20:10
*** lbragstad has joined #openstack-ansible20:10
lbragstadthis might be something that interests deployers https://review.openstack.org/#/c/450415/920:11
mhaydenORLY20:13
mhaydenthis piques my interests20:13
*** tobberydberg has quit IRC20:15
*** jeh has quit IRC20:15
*** redaudit has quit IRC20:16
*** mikal has quit IRC20:25
*** jamesdenton has quit IRC20:25
*** mikal has joined #openstack-ansible20:27
*** zul has quit IRC20:32
*** zul has joined #openstack-ansible20:35
*** redaudit has joined #openstack-ansible20:40
*** yifei has quit IRC20:53
*** galstrom is now known as galstrom_zzz20:54
*** yifei has joined #openstack-ansible20:54
*** retreved_ has quit IRC20:56
*** lbragstad has quit IRC20:59
*** marst_ has quit IRC21:02
*** hw_wutianwei1 has joined #openstack-ansible21:02
*** oneswig has joined #openstack-ansible21:03
*** hw_wutianwei has quit IRC21:05
*** hw_wutianwei1 is now known as hw_wutianwei21:05
*** zul has quit IRC21:05
*** cpuga has joined #openstack-ansible21:07
*** oneswig has quit IRC21:07
*** smatzek has joined #openstack-ansible21:14
*** zul has joined #openstack-ansible21:15
*** yolanda has quit IRC21:18
*** cjloade__ has joined #openstack-ansible21:19
*** yolanda has joined #openstack-ansible21:20
*** esberglu has quit IRC21:20
*** cpuga has quit IRC21:21
*** cpuga has joined #openstack-ansible21:22
*** cjloader_ has quit IRC21:22
*** cjloade__ has quit IRC21:23
*** smatzek has quit IRC21:25
*** cpuga has quit IRC21:26
*** zul has quit IRC21:27
*** hamza has joined #openstack-ansible21:33
*** cpuga has joined #openstack-ansible21:36
*** schwicht has quit IRC21:36
*** lbragstad has joined #openstack-ansible21:42
*** esberglu has joined #openstack-ansible21:53
*** esberglu has quit IRC21:55
*** schwicht has joined #openstack-ansible22:02
*** thorst_afk has quit IRC22:06
openstackgerritMarkos Chandras (hwoarang) proposed openstack/openstack-ansible-tests master: test-prepare-host: Remove useless filter for RedHat network ifcfg files  https://review.openstack.org/46642122:06
openstackgerritMarkos Chandras (hwoarang) proposed openstack/openstack-ansible-tests master: SUSE: Remove post-up/post-down scripts from default ifcfg files  https://review.openstack.org/46642222:06
*** ianychoi has quit IRC22:10
*** ianychoi has joined #openstack-ansible22:13
*** lbragstad has quit IRC22:19
*** thorst_afk has joined #openstack-ansible22:19
*** thorst_afk has quit IRC22:23
*** thorst_afk has joined #openstack-ansible22:23
*** thorst_afk has quit IRC22:24
*** schwicht has quit IRC22:32
*** oneswig has joined #openstack-ansible22:32
*** oneswig has quit IRC22:36
*** hamza has quit IRC22:38
*** vnogin has quit IRC22:41
*** chas has joined #openstack-ansible22:41
*** chas has quit IRC22:46
*** lbragstad has joined #openstack-ansible22:49
openstackgerritKevin Carter (cloudnull) proposed openstack/openstack-ansible-os_nova master: [WIP] Testing using a common SystemD init role  https://review.openstack.org/46617922:57
*** jamielennox is now known as jamielennox|away23:00
*** cpuga has quit IRC23:00
*** hw_wutianwei has quit IRC23:05
*** hw_wutianwei1 has joined #openstack-ansible23:05
*** weezS has quit IRC23:07
*** hw_wutianwei1 is now known as hw_wutianwei23:07
*** weezS has joined #openstack-ansible23:07
*** alezzandro has quit IRC23:23
*** admin0 has quit IRC23:24
*** thorst_afk has joined #openstack-ansible23:24
*** thorst_afk has quit IRC23:29
*** weezS has quit IRC23:32
*** jamielennox|away is now known as jamielennox23:40
*** zul has joined #openstack-ansible23:46
jgriffithsTrying to install newton, but getting this error "Collecting gnocchiclient<3.0.0,>=2.2.023:46
jgriffiths  Could not find a version that satisfies the requirement gnocchiclient<3.0.0,>=2.2.0 (from versions: 3.1.1)23:46
jgriffithsNo matching distribution found for gnocchiclient<3.0.0,>=2.2.023:46
jgriffiths"23:46
jgriffithsDoes anybody have any idea why23:46
*** lbragstad has quit IRC23:54
*** esberglu has joined #openstack-ansible23:59

Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!