Thursday, 2019-02-28

*** macza_ has quit IRC00:03
*** sdake has joined #openstack-ansible00:04
*** dave-mccowan has quit IRC00:07
*** dxiri has joined #openstack-ansible00:15
*** sm806 has quit IRC00:17
*** sm806 has joined #openstack-ansible00:18
*** sm806 has quit IRC00:28
*** sm806 has joined #openstack-ansible00:29
*** cmart has quit IRC00:31
*** sdake has quit IRC01:08
*** DanyC_ has joined #openstack-ansible01:10
*** DanyC has quit IRC01:12
*** sm806 has quit IRC01:21
*** sm806 has joined #openstack-ansible01:21
*** hwoarang has quit IRC01:23
*** dxiri has quit IRC01:27
*** hwoarang has joined #openstack-ansible01:29
*** cyberpear has quit IRC01:29
*** gyee has quit IRC01:38
*** sdake has joined #openstack-ansible02:09
*** cshen has joined #openstack-ansible02:37
*** cshen has quit IRC02:41
*** sdake has quit IRC02:46
jamesdentonevenin'02:52
*** ansmith has quit IRC02:54
*** ansmith_ has joined #openstack-ansible02:54
*** dave-mccowan has joined #openstack-ansible03:06
*** DanyC has joined #openstack-ansible03:11
*** udesale has joined #openstack-ansible03:15
*** DanyC has quit IRC03:15
*** sdake has joined #openstack-ansible03:16
*** sdake has quit IRC03:17
*** sdake has joined #openstack-ansible03:22
*** sdake has quit IRC03:23
*** sdake has joined #openstack-ansible03:28
*** dave-mccowan has quit IRC03:32
*** sdake has quit IRC03:40
*** markvoelker has joined #openstack-ansible03:48
*** spsurya has joined #openstack-ansible04:14
*** markvoelker has quit IRC04:20
*** cshen has joined #openstack-ansible04:47
*** cshen has quit IRC04:51
*** gkadam has joined #openstack-ansible04:55
*** raukadah has quit IRC05:03
*** chandankumar has joined #openstack-ansible05:04
*** markvoelker has joined #openstack-ansible05:17
*** shyamb has joined #openstack-ansible05:19
*** shyamb has quit IRC05:28
*** sdake has joined #openstack-ansible05:30
*** cmart has joined #openstack-ansible05:37
*** markvoelker has quit IRC05:51
*** shyamb has joined #openstack-ansible05:53
*** eumel8 has joined #openstack-ansible05:53
*** sdake has quit IRC06:02
*** sdake has joined #openstack-ansible06:22
*** sdake has quit IRC06:26
*** rgogunskiy has joined #openstack-ansible06:26
*** gkadam has quit IRC06:27
*** ivve has joined #openstack-ansible06:36
*** sdake has joined #openstack-ansible06:43
*** markvoelker has joined #openstack-ansible06:48
*** shyamb has quit IRC06:51
*** shyamb has joined #openstack-ansible06:55
*** miloa has joined #openstack-ansible07:00
miloaGood morning :)07:00
*** sdake has quit IRC07:15
*** rgogunskiy has quit IRC07:18
*** markvoelker has quit IRC07:20
*** ChosSimbaOne has quit IRC07:23
*** ChosSimbaOne has joined #openstack-ansible07:23
*** cmart has quit IRC07:32
*** kopecmartin|off is now known as kopecmartin07:34
*** cshen has joined #openstack-ansible07:37
*** shyamb has quit IRC07:39
*** shyamb has joined #openstack-ansible07:40
jrosserMorning07:46
ChosSimbaOneMorning :-)07:47
fnpanicMorning07:50
*** aedc has quit IRC07:53
*** hamzaachi has joined #openstack-ansible07:57
*** shyamb has quit IRC07:58
*** gkadam has joined #openstack-ansible08:02
*** maxbab has joined #openstack-ansible08:03
*** markvoelker has joined #openstack-ansible08:18
*** DanyC has joined #openstack-ansible08:23
*** pcaruana has joined #openstack-ansible08:32
*** DanyC has quit IRC08:33
*** tosky has joined #openstack-ansible08:39
CeeMacmorning08:45
miloaHi, On OSA aio stable/rocky, tag 18.14, install_method: distro, on a CentOS 7 VM, when I deploy, I've got a failed test in tempest. The cirros instance seems to start, but tempest cannot connect with ssh.08:49
*** markvoelker has quit IRC08:50
openstackgerritMerged openstack/openstack-ansible-openstack_hosts master: Updated from OpenStack Ansible Tests  https://review.openstack.org/63875508:54
*** shyamb has joined #openstack-ansible08:55
miloahere is the tempest report : http://paste.openstack.org/show/746476/08:55
miloaWhen I play only the tempest-install.yml playbook I have se same error. Any help would be appreciate, thanks :)08:58
*** rgogunskiy has joined #openstack-ansible09:00
*** rgogunskiy has quit IRC09:00
*** aedc has joined #openstack-ansible09:06
*** electrofelix has joined #openstack-ansible09:11
*** aedc has quit IRC09:14
*** sdake has joined #openstack-ansible09:16
*** olivierbourdon38 has joined #openstack-ansible09:20
*** phasespace has joined #openstack-ansible09:21
*** shyamb has quit IRC09:21
*** olivierbourdon38 has quit IRC09:22
*** rgogunskiy has joined #openstack-ansible09:23
*** shyamb has joined #openstack-ansible09:24
*** udesale has quit IRC09:28
*** udesale has joined #openstack-ansible09:29
chandankumarmiloa: It appears to be ssh timeout09:34
chandankumarmiloa: can you check whether 172.29.249.133 is pingable09:34
chandankumarmiloa: or look at the down of cirros image booting output09:35
*** shyamb has quit IRC09:36
*** shyamb has joined #openstack-ansible09:36
fnpanichi09:37
fnpanici switched from containers to bare metal for the elk stuff using the overlay-invetor09:38
fnpaniccleared the containers from dynamic inventroy with inventory manage09:38
fnpanicbut still the haproxy config points to the containers09:38
fnpanicany ideas?09:38
fnpanicinventory-manage .-G shows the correct groups09:39
fnpanicin the playbook run i see that the entries are skipped on haproxy config09:40
fnpanicwhy?09:41
*** shyamb has quit IRC09:41
openstackgerritDmitriy Rabotjagov (noonedeadpunk) proposed openstack/openstack-ansible-os_horizon master: Moved compilemessages to handlers  https://review.openstack.org/63996009:43
fnpanici tried with --tags haproxy-config and without09:43
fnpanic:-(09:43
*** sdake has quit IRC09:46
*** markvoelker has joined #openstack-ansible09:47
*** rgogunskiy has quit IRC09:49
miloachandankumar: yes, it is an ssh timeout. The problem to check if 172.29.249.133 is pingable, is that tempest erase the vm at end of tests.09:51
chandankumarmiloa: tempest cleans its stuff as a tear down process during the tests09:52
fnpanici found it09:54
fnpaniclooks like the groups ware lacking _all because i have not loaded the overlay inventory for the haproxy run09:54
miloachandankumar: yes, i know :) so i will try to do it manually.09:56
*** aedc has joined #openstack-ansible10:02
miloachandankumar: I create manualy a cirros instance with floating ip 172.29.249.111 from the aio host it is pinguable but not from the lxc utiliy container.10:06
chandankumarmiloa: it is allinone scenario na?10:08
chandankumarmiloa: sorry so tempest is running from lxc container?10:09
miloachandankumar: yes it is OSA aio stable/rocky, tag 18.14, install_method: distro, on a CentOS 7 VM.10:09
chandankumarjrosser: hello10:10
chandankumarjrosser: ^ please have a look at miloa issue when free! thanks!10:10
*** DanyC has joined #openstack-ansible10:11
miloachandankumar: at the end of aio deployment the playbook os-tempest-install.yml is played. And tempest run from utility container, the results are stored in : /openstack/log/aio1_utility_container-xxxxxx/utility/stestr_results.html10:12
*** aedc has quit IRC10:13
chandankumarmiloa: can you attach and see the tempest workspace there and ping from there10:13
odyssey4memiloa typically that means there's an issue in the network config somewhere, although I don't know enough to help you troubleshoot it10:14
odyssey4methe container/host running tempest will need a route to the public/external interface10:14
*** shyamb has joined #openstack-ansible10:14
chandankumarmiloa: there is a task also https://github.com/openstack/openstack-ansible-os_tempest/blob/master/tasks/tempest_resources.yml#L28210:14
chandankumarwhich does the same for you10:14
*** DanyC has quit IRC10:17
miloachandankumar: do you mean something like lxc-attach -n utility-container ping ... ?10:19
chandankumarmiloa: lxc-attach -n utility container10:20
chandankumarit will take you the container then try from there10:20
* chandankumar has not tried this scenario10:20
miloaalready done : "I create manualy a cirros instance with floating ip 172.29.249.111 from the aio host it is pinguable but not from the lxc utiliy container."10:21
miloa:)10:21
*** markvoelker has quit IRC10:21
*** sohny has joined #openstack-ansible10:21
chandankumarmiloa: as odyssey4me said that might be the issue10:22
jrosserwell - shouldnt eth0/lxcbr0 NAT the utility container to the instance IP?10:22
jrosserthis sounds like the failure to forward on bridges again10:23
*** udesale has quit IRC10:24
*** udesale has joined #openstack-ansible10:24
jrossernormally in an AIO the default route on eth0 should allow ping/ssh to tempest VM10:25
odyssey4mejrosser well, that would depend on the network setup - assuming this is not a standard AIO10:25
jrossermiloa> chandankumar: yes it is OSA aio stable/rocky, tag 18.14, install_method: distro, on a CentOS 7 VM.10:26
jrosser^ me wonders about centos not forwarding again10:26
*** DanyC has joined #openstack-ansible10:27
chandankumarjrosser: we have changed stuff on os_temepst side10:27
odyssey4meor perhaps whether that tag doesn't include the required fix10:27
jrosserodyssey4me: i want to re-organise the AIO networking a bit when i get time10:28
chandankumarafaik, openstack-ansible might have different network configuraiton10:28
chandankumar?10:28
jrossermake a "public" network that is for neutron and wired into the utility container for tempest10:28
jrosserand start to make the mgmt network more isolated rather than overload it for all these things10:28
odyssey4mebother, it looks like the opensuse 150 image is gone: http://logs.openstack.org/29/639729/1/check/openstack-ansible-deploy-aio_lxc-opensuse-150/4a8668c/logs/host/aria2c-image-prestage.log.txt.gz10:29
odyssey4meor the path or file name has changed10:29
chandankumaris it http://zuul.openstack.org/builds?job_name=openstack-ansible-deploy-aio_distro_lxc-centos-7 the same scenario which miloa has deployed?10:30
miloaIn the default user_variables.yml for aio there is : tempest_public_subnet_cidr: "172.29.248.0/22" and from the utility container I can ping 172.29.248.1 which is the br-vlan address. The problem would come from br-vlan not forwarding ?10:31
jrossermiloa: https://github.com/openstack/ansible-role-systemd_networkd/commit/242b3c3fb4a3f8ebeba5c24cf8c7a510cba1414310:32
noonedeadpunkfolks, let me know what you think about https://review.openstack.org/#/c/639960/ I'm not sure, that I've choose the correct way - probably it would be enough just to make code the same in 2 places instead of creating handler10:33
jrossermiloa: https://review.openstack.org/#/c/636162/10:33
odyssey4meaha, that's the problem: http://logs.openstack.org/29/639729/1/check/openstack-ansible-deploy-aio_lxc-opensuse-150/4a8668c/logs/ara-report/result/3a02f4ec-f992-4138-a146-b2118b9dc847/10:34
*** sohny has quit IRC10:34
*** sdake has joined #openstack-ansible10:37
*** shyamb has quit IRC10:37
*** mkuf has quit IRC10:37
openstackgerritDmitriy Rabotjagov (noonedeadpunk) proposed openstack/openstack-ansible-os_horizon master: Set appropriate branch key in _horizon_translations_pull  https://review.openstack.org/63997810:39
miloajrosser: thanks :) I checked on aio host : ipv4/ip_forward is set to 1 but conf/br-vlan/forwarding is set to 0, I will change it to 1 and replay the os-tempest-install.yml and let you know.10:40
miloajrosser: \o/ first test to ping manualy created instance from the utility container now works. I start replay the os-tempest-install.yml...10:43
openstackgerritJesse Pretorius (odyssey4me) proposed openstack/openstack-ansible-lxc_hosts master: Fix image download URL regex for opensuse-leap-15.0  https://review.openstack.org/63998110:50
odyssey4mejrosser evrardjp https://review.openstack.org/639981 should resolve the current opensuse 15 build failures10:50
*** aedc has joined #openstack-ansible10:59
*** mkuf has joined #openstack-ansible10:59
*** mkucia_ has quit IRC11:01
*** mkuf has quit IRC11:12
*** mkuf_ has joined #openstack-ansible11:12
*** markvoelker has joined #openstack-ansible11:18
*** sdake has quit IRC11:18
*** sdake has joined #openstack-ansible11:20
*** shyamb has joined #openstack-ansible11:20
openstackgerritErik Berg proposed openstack/openstack-ansible stable/queens: Remove heat_api_cloudfront from haproxy services  https://review.openstack.org/63914511:29
*** aedc has quit IRC11:30
openstackgerritChandan Kumar proposed openstack/openstack-ansible-os_tempest master: Use tempest_system_user_name/group for tempest directories  https://review.openstack.org/64000511:33
chandankumarodyssey4me: ^^ I am not sure it is correct, feel free to take a look when free, thanks!11:33
CeeMacanyone got a workaround for live migrations failing with host key verifications issues?11:35
CeeMacshould qemu+ssh work oob or do i need to tweak some settings11:35
CeeMacor revert to qemu+tcp?11:35
miloajrosser: I played the os-tempest-install.yml after setting manualy conf/br-vlan/forwarding to 1, but it failed, and the strange thing is that the conf/br-vlan/forwarding is back to 0, steps : http://paste.openstack.org/show/746482/11:48
*** markvoelker has quit IRC11:50
*** mkuf_ is now known as mkuf11:51
miloajrosser: replaying os-tempest-install.yml with -vvv and --diff to try to understand why conf/br-vlan/forwarding is set back to 0...11:52
CeeMacinteresting, compute1 has authorized_keys in /var/lib/nova/.ssh/ but compute 2 doesnt11:54
CeeMacshouldn't that be set up when running setup-openstack using the nova-key tag?11:54
*** sdake has quit IRC11:55
*** rgogunskiy has joined #openstack-ansible12:00
*** sm806 has quit IRC12:04
*** sm806 has joined #openstack-ansible12:05
*** shyamb has quit IRC12:14
*** ansmith_ has quit IRC12:38
*** fdz has joined #openstack-ansible12:43
CeeMacok, got around that by setting the variable to regenerate keys and not running the playbooks with the tag12:43
CeeMacI think this is the second time i've had issues with tags in nova playbook :(12:43
miloajrosser: some more element, I found that conf/br-vlan/forwarding is set back to 0 when tempest create a router. So I try it creating a router manually in horizon, and conf/br-vlan/forwarding is set back to 012:44
jrossermiloa: really?!12:44
CeeMachaving an issue with 'guest CPU doesn't match specification: missing features: vmx' now :/12:44
miloayes12:44
jrosserwow12:44
miloaI set a " watch -n 1 -d cat /proc/sys/net/ipv4/conf/br-vlan/forwarding" to check when exactly is go back to 012:45
miloaand when tempest create a router or when I create a router "manually from horizon dashboard" it is set back to 012:46
*** markvoelker has joined #openstack-ansible12:47
miloaso it is not a problem with os_tempest, I do not understand why when a router is created it change the value of /proc/sys/net/ipv4/conf/br-vlan/forwarding to 0.12:48
openstackgerritChandan Kumar proposed openstack/openstack-ansible-os_tempest master: Use tempest_system_user/group_name  for tempest directories  https://review.openstack.org/64000512:51
*** shyamb has joined #openstack-ansible12:53
*** shyamb has quit IRC12:58
*** shyamb has joined #openstack-ansible13:01
miloajrosser: when I do not select public network when creating the router it did not change br-vlan/forwarding. But if I select public network, it change the value, and in /var/log/messages of aio host I've got "Feb 28 12:56:33 localhost systemd: Starting Network Service..."13:02
miloajrosser: so there is a restart or reload of network service. And the forwarding value for the br-vlan is reset.13:03
*** sm806 has quit IRC13:11
*** sm806 has joined #openstack-ansible13:11
*** rgogunskiy has quit IRC13:16
*** strattao has joined #openstack-ansible13:18
*** markvoelker has quit IRC13:20
CeeMacanybody got any thoughts on why deploying nova to 2 identical hosts with the same variables would result in kvm nested virtualisation support being enabled on 1 host, but not the other?13:29
jrossermiloa: do you mean the ansible restarts the networking, or creating the neutron router with cli/horizon restarts the networking?13:29
miloajrosser: creating the neutron router with cli/horizon restarts the networking13:31
jrosserjamesdenton:  ^^^13:31
jamesdentonhi13:32
jrosserwe got some more non-forwarding funnies on centos by the look of things13:32
jamesdentonI cannot explain why networkd does the things it does13:32
*** shyamb has quit IRC13:32
jrosserhahaha13:32
jamesdentonum, i noticed stable/rocky. Were those patches backported?13:32
guilhermespmornin13:33
jamesdentonmornin13:33
jrossermiloa: did you apply the patches i linked, or just enable the forwarding by hand?13:33
miloaI enable forwarding by hand13:33
guilhermespI've seen since yesterday this failing accross two prs I submitted http://logs.openstack.org/11/639811/1/check/openstack-ansible-functional-centos-7/dcf4eaa/logs/ara-report/result/ae01869d-4840-4fa8-b47a-251e376e67a3/13:33
jrossermiloa: perhaps hand edit the networkd config file to add the forwarding option13:33
CeeMachi jamesdenton , guilhermesp13:33
guilhermespjamesdenton CeeMac o/13:34
jamesdentonmiloa i think you'll want to hand edit the file like jrosser mentioned13:34
jamesdentonguilhermesp I noticed many tempest failures yesterday just hadn't dug into it13:35
guilhermespyes, just started to notice now13:35
jamesdentonCeeMac Did you notice particular configurations existing on one but not the other? Can you elaborate?13:36
miloajrosser: yes, I was thinking of that, but I am trying to find if  i can set ipv4.conf.br-vlan.forwarding=1 to sysctl.conf.d and check if it works13:36
jamesdentonmiloa that does not work13:36
jamesdentonWell, i take that back. A '-w' may not be persistent. I don't know that we trying editing sysctl.conf13:37
*** ansmith has joined #openstack-ansible13:38
CeeMacjamesdenton, i was getting an error about 'vmx' capabilities13:39
fnpanichi13:39
CeeMacso i was asking the question in #openstack and someone said thats a module parameter for kvm on intel cpu and to check 3a13:40
fnpanici just wanted to do a run of os-horizon playbook13:40
CeeMacso I did, and 1 host is Y and the other is N13:40
fnpanici am getting this error13:40
fnpanichttp://paste.openstack.org/show/746498/13:40
CeeMaci got around it by setting the cpu arch to x86_64 in the flavor but was curious how it happened13:40
fnpanicpretty strange13:40
jamesdentonthat's a good question, i don't really know.13:41
CeeMacthat, and live migration breaks the novnc console13:42
CeeMacare today's fun13:42
CeeMacalso, tags dont work for me with os-nova-install, specifically nova-key13:44
CeeMacwhich caused me other issues earlier13:44
miloajrosser and jamesdenton: I edit /etc/systemd/network/13-general-br-vlan.network and add in section [Network] the value : IPForward=ipv4 and... it works \o/ the value do not change.13:45
jamesdentonfnpanic what version is this?13:45
jamesdentonyes, tags may be broken in spots13:46
miloajrosser and jamesdenton: replaying the os-tempest-install.yml  playbook13:46
CeeMacjamesdenton, it was just a little unfortunate as I'd followed the official documentation for adding a new host and with the tag not working the nova keys werent populated on the new host13:47
jamesdentonjrosser Thoughts on this? https://review.openstack.org/#/c/637026/. Curious to know if this would be a good approach to fixing some of the tag issues we're seeing13:47
jamesdentonCeeMac Understandable. See that patch ^^ on why (i think) it's happening13:48
* CeeMac is looking13:49
fnpanicjamesdenton: rocky on 18.0413:51
CeeMacthe includes' are getting deprecated in ansible right?  I'm sure i've seen purple messages about that when running playbooks13:51
jamesdentoninclude has been replaced by include_tasks and import_tasks, from what i understand.13:52
CeeMacah,yeah. it was 'import_playbooks' i've seen the message about, which is replacing the include for playbooks13:53
jamesdentonbut i can't speak authoritatively on that13:53
CeeMacso i guess its a common theme13:53
miloajrosser jamesdenton: \o/ os-tempest-install.yml :  aio1_utility_container-8d5d4380 : ok=41   changed=4    unreachable=0    failed=013:58
jamesdentonfnpanic i assume you ran thru an install of OSA and are later running the os-horizon playbook?13:58
jamesdentoncongrats miloa13:58
jrossermiloa: cool!14:00
openstackgerritMerged openstack/openstack-ansible-lxc_hosts master: Fix image download URL regex for opensuse-leap-15.0  https://review.openstack.org/63998114:04
openstackgerritJesse Pretorius (odyssey4me) proposed openstack/openstack-ansible-lxc_hosts stable/rocky: Fix image download URL regex for opensuse-leap-15.0  https://review.openstack.org/64003214:05
fnpanicjamesdenton: exactly14:05
fnpanicjamesdenton: was doing some theme install stuff14:05
fnpanichow can i debug this?14:05
fnpanicthe repo is reachable but the path it tries to access does not exist14:05
odyssey4mechandankumar could you indicate why it is you've chosen to do it that way rather than like http://eavesdrop.openstack.org/irclogs/%23openstack-ansible/%23openstack-ansible.2019-02-27.log.html#t2019-02-27T15:14:20 in https://review.openstack.org/#/c/640005/2/tasks/tempest_post_install.yml ?14:06
mnasero/14:13
*** markvoelker has joined #openstack-ansible14:18
jamesdentonyeah, for some reason it is not populating the version. that may be why it says undefined14:19
jamesdentonso the var may not be accessible. but i don't know how it's constructing the url14:19
fnpanichere is the -vvv output14:20
fnpanicthe end14:20
fnpanichttp://paste.openstack.org/show/746504/14:20
*** phasespace has quit IRC14:20
*** ianychoi has joined #openstack-ansible14:20
fnpanicmhh where is the version coming from?14:21
openstackgerritMartin Kopec proposed openstack/ansible-role-python_venv_build master: Fix link for the project documentation  https://review.openstack.org/64004014:24
chandankumarodyssey4me: first I was trying to create those dirs just by using ansible user itself whether it works or not, if not work then I would go through the os_nova way14:25
fnpanici think i know what maybe happend14:26
fnpanici did yesterday a git pull on the /opt/openstack-ansible14:26
odyssey4mechandankumar I think it would be better to just not apply a user/group and let the distro decide that based on the user doing the action.14:26
fnpanicmaybe something is broken upstream?14:26
fnpanici heard discussions about the repo containers recently14:27
chandankumarodyssey4me: sure, so we are providing more choices to users/distro let them decide how they want14:28
chandankumarodyssey4me: one more thing, since we are going to change the user, how we are going to make sure package installation works as it requires sudo privilages14:29
odyssey4mefnpanic you may need to apply https://review.openstack.org/63967214:29
chandankumarodyssey4me: at this part https://github.com/openstack/openstack-ansible-os_tempest/blob/master/tasks/tempest_install.yml#L1714:29
*** maxbab has quit IRC14:29
odyssey4mechandankumar the role is assuming that it is run as root - if you're using a different user, then you should apply 'become: yes' at the playbook14:30
chandankumarodyssey4me: is it not possible to apply at that task level as apart from that we donot need sudo privilages? as tempest is just an application14:31
openstackgerritGabriele Santomaggio proposed openstack/openstack-ansible-rabbitmq_server master: The pattern  '^(?!(amqp.)|(.*_fanout_)|(reply_)).*' excludes:  https://review.openstack.org/63929914:32
odyssey4mechandankumar well, there are several things done in the role which assume root... of course it could be changed, and that'd be a nice change14:33
chandankumarodyssey4me: let me handle the user part and then go with that part, thanks :-)14:34
*** marens has joined #openstack-ansible14:34
mnaserodyssey4me: yeah assuming our roles dont run as root would be _neat_14:35
*** maxbab has joined #openstack-ansible14:36
odyssey4mefnpanic also, do you have this patch in your git tree? https://github.com/openstack/openstack-ansible/commit/c75211306d41bd76babccd9273063eca49709b4714:39
odyssey4mefnpanic if so, when you updated git did you bootstrap-ansible again, like you always should after updating git?14:40
odyssey4meupdating the git repo for openstack-ansible I mean14:40
mnaserhistorically14:41
mnaserhave we assumed our roles are openstack specific?14:41
odyssey4memnaser sort-of, given that they're mostly only used within OSA14:41
mnaserodyssey4me: ah, i was hoping we can maybe slowly try to make our more infrastructure based roles less specific -- mostly around the context of https://review.openstack.org/#/c/639299/214:42
odyssey4meI see the discussion in https://review.openstack.org/639299 and it got me thinking that for the more generic roles, perhaps we should have a nice and easy flag to enable/disable the openstack assumptions.14:42
fnpanicodyssey4me: no did not upgrade14:42
fnpanicwill need to follow the upgrade procedure14:43
mnaseri think we can document it or something, we would need to have a hard decision on this14:43
odyssey4memnaser it'd be better to keep the stuff in the roles where possible, because it makes them easier to find and maintain - but have a flag to turn on the openstack use-case, then turn that flag on in group_vars... or another option is, in defaults, to check for the existance of certain groups which tells us it's an openstack environment14:44
mnaserodyssey4me: i like that, i think ceph-ansible does something similar?14:44
*** dxiri has joined #openstack-ansible14:45
openstackgerritMikael Loaec proposed openstack/openstack-ansible master: Forwarding for br-vlan in aio does not persist  https://review.openstack.org/64005314:45
mnaserhttps://github.com/ceph/ceph-ansible/blob/d7e77012ef4fe523762928c5d0081ff773be7068/roles/ceph-osd/tasks/main.yml#L89-L9514:46
odyssey4memnaser yeah, something like that - in this case we wouldn't need another task file though14:47
*** aedc has joined #openstack-ansible14:49
*** markvoelker has quit IRC14:50
miloa https://review.openstack.org/640053 if accepted, it would be cool if we can backport it to 18.1.4 because without it the users who follow the all in one quick install from doc will always have errors as 18.1.4 is the advised tag to use in the doc.14:51
*** maxbab has quit IRC14:52
mnasermiloa: the way it would work is that once we merge this and release a new version, the docs will update and advise to point at the newer version14:54
mnaserodyssey4me: how much do you feel like maybe posting about this to the ML for us to discuss it? :P14:54
odyssey4memnaser I just posted in review. I don't really feel like it needs discussion, given that there's a simple enough solution.14:55
evrardjpmnaser: I think odyssey4me's solution is elegant for the rabbitmq part15:06
evrardjpbut I think it should be done in a different commit :p15:07
evrardjpfirst adapt the bad default, then make it optional eventually15:07
miloamnaser: ok, but i am a bit worry :) because users who wants to use openstack-ansible for their production (like us) begin with the aio and apply what they read in the doc for testing. So if it ends with errors perhaps they will no search more and go away :). Because actually the aio install always fails. :)15:07
evrardjpisn't that a bad experience for a first contributor to have to go through those hoops?15:07
mnaserevrardjp: are you nominating yourself to make that follow up patch :D15:08
evrardjpanyway -- I will help him15:08
evrardjpfine for me15:08
jamesdentonmiloa You're saying the AIO fails with tempest, right?15:08
miloayes15:08
odyssey4meevrardjp mnaser I'm happy to do the follow up if the new contributor can't see how it can be done, but if you're also happy to do it then great.15:10
*** shyamb has joined #openstack-ansible15:10
miloabut if we apply :  https://review.openstack.org/640053 it must work. And the failure is not only tempest related.15:10
evrardjpI just want to help -- if it takes me an extra minute, it's not bad :)15:10
odyssey4me++15:10
evrardjpI also want to make sure the good practices are taught :)15:10
jamesdentonThere were patches implemented to fix this in master, IIRC. They may just a) need to be backported and/or b) aio network build needs to be tweaked15:12
jamesdentonmiloa you could always use ubuntu :D15:12
*** ivve has quit IRC15:13
miloajamesdenton: not only tempest, because if connect to aio horizon and create a router you will not be able to ping or connect to your instances through lxc utility container nor from the aio host.15:13
jamesdentonright - the patches weren't tempest specific, but they set a ipforward config in the networkd template15:14
miloajamesdenton: no, because it is our organisation choice to use CentOS.15:15
jamesdentonbut for an AIO, the bridge configuration needs to call out 'ipforward: true', which may not be happening15:15
jamesdentonmiloa that's fine, just teasing15:15
miloajamesdenton: :)15:15
chandankumarodyssey4me: just one stupid question15:32
chandankumarodyssey4me: https://github.com/openstack/openstack-ansible-os_nova/blob/master/tasks/nova_pre_install.yml#L85 -> we need tempest user/group needs to be created na if not exists?15:33
odyssey4mechandankumar For tempest, I would think that we shouldn't create or use any user account specifically - it should just deploy using whatever user ansible is accessing the machine as15:35
openstackgerritGabriele Santomaggio proposed openstack/openstack-ansible-rabbitmq_server master: Improve the Highly Available (Mirrored) Queues policy  https://review.openstack.org/63929915:38
*** hamzaachi has quit IRC15:39
*** shyamb has quit IRC15:42
*** markvoelker has joined #openstack-ansible15:47
*** sm806 has quit IRC15:56
*** sm806 has joined #openstack-ansible15:57
*** cmart has joined #openstack-ansible16:13
*** cshen has quit IRC16:13
*** udesale has quit IRC16:13
chandankumarodyssey4me: let me modify my patch!16:15
*** mathlin has quit IRC16:15
*** mathlin has joined #openstack-ansible16:16
*** markvoelker has quit IRC16:21
*** fdz has quit IRC16:25
*** strattao_ has joined #openstack-ansible16:28
*** strattao has quit IRC16:29
*** strattao_ is now known as strattao16:29
*** chandankumar is now known as raukadah16:34
*** devx has joined #openstack-ansible16:39
*** fdz has joined #openstack-ansible16:40
*** hamzy has quit IRC16:40
*** macza has joined #openstack-ansible16:41
*** fdz has quit IRC16:48
*** fdz has joined #openstack-ansible16:50
*** fdz has quit IRC16:56
*** sm806 has quit IRC16:56
*** sm806 has joined #openstack-ansible16:56
*** electrofelix has quit IRC16:59
*** sm806 has quit IRC17:03
*** sm806 has joined #openstack-ansible17:03
*** shananigans has joined #openstack-ansible17:05
*** gkadam has quit IRC17:08
*** strattao has quit IRC17:13
*** kopecmartin is now known as kopecmartin|off17:17
*** markvoelker has joined #openstack-ansible17:18
*** hogepodge has left #openstack-ansible17:20
*** cmart has quit IRC17:33
*** DanyC has quit IRC17:35
*** DanyC has joined #openstack-ansible17:35
*** hwoarang has quit IRC17:38
*** hwoarang has joined #openstack-ansible17:39
*** DanyC has quit IRC17:40
*** gyee has joined #openstack-ansible17:44
miloajamesdenton: are you using Ubuntu ? you do not have the br-vlan problem ?17:45
*** cmart has joined #openstack-ansible17:49
*** markvoelker has quit IRC17:51
miloajamesdenton: if you are using Ubuntu, which method to you use ? source or distro ? because I just finish to deploy on CentOS AIO with method 'source' and there is no tempest fail.17:51
*** strattao has joined #openstack-ansible17:51
* miloa wondering if the problem is related to RDO packages, 'cause with source install method there is no error. ref : https://review.openstack.org/64005317:52
raukadahmiloa: if it is a packaging problem on rdo side, I will get it fixed!17:53
miloaraukadah: thanks :)17:54
raukadahcloudnull: around?17:55
*** aludwar has quit IRC17:56
raukadahcloudnull: I have an idea I want to experiment in ci remove action plugins path we are setting in ansible.cfg to check whether defaults /usr/share picks or not from pip insyall17:56
raukadah*install17:56
*** fdz has joined #openstack-ansible17:59
jrossermiloa: generally most people are using the source installs, distro *should*18:02
jrossermiloa: interested to hear why you choose the distro install :)18:02
jrosseri mean, distro *should* work18:02
openstackgerritChandan Kumar proposed openstack/ansible-role-python_venv_build master: [WIP]Use share/ansible/roles for data files  https://review.openstack.org/64013318:04
raukadahthis whole ansible role path is a mess18:06
raukadahmay be mazer can solve it18:06
raukadahunless you have multiple ansible.cfg files on system18:07
miloajrosser: because distro install quickly :) and because I was thinking that distro (RDO package) are more stable in particular when you have to scale out and you are in production. Actually we are using AIO in order to learn.18:07
miloajrosser: but the final objective is to use it to our pre-production cloud and then to our production (if there is one :) )18:08
*** dxiri has quit IRC18:08
miloajrosser: so if it works, i think i will stick to this channel for a long time. lol18:11
openstackgerritChandan Kumar proposed openstack/openstack-ansible-os_tempest master: [wip] use requirments.txt for managing os_tempest dependency  https://review.openstack.org/64013818:12
raukadahmnaser: evrardjp as discussed on os_tempest deps management I have putted few patches https://review.openstack.org/#/q/topic:ostempest_packaging+(status:open+OR+status:merged) I will exercise it tomorrow18:13
raukadahfeel free to take a look18:13
*** hamzy has joined #openstack-ansible18:13
raukadahas we have one action plugins18:14
raukadahusing galaxy solves roles issue but not action plugin18:14
evrardjpraukadah: it does if you havne't configured things18:15
evrardjpbut I have to go -- we can talk about that later18:15
raukadahmazer may solve it https://github.com/ansible/mazer/wiki/Module-Install-Use-Cases18:18
jrossermiloa: the distro vs source install in the AIO gate is 1hr20 vs 1hr50.... but for an AIO. I think the reality is that you are not going to gain a massive amount.18:21
*** cmart has quit IRC18:22
jrosserJust be clear why you make the choice - it’s much harder to locally patch your stuff with distro packages, that may or may not be important to you18:22
miloajrosser: Thanks :), I am not the only one to decide :) and I think that using source for some people is something frightning. But I will propose it at our next team meeting. So if you have more point in favor of source vs distro. I will be happy to ear them :)18:26
jrosserThere are just pros and cons both ways :)18:26
raukadahmiloa: jrosser if you need stable go with distro, config is something distro nevers care about!18:27
raukadahmay be I am naive in that area just a small college passout kid trying to learn stuff!18:28
jrosserSo for the avoidance of doubt an OSA source install is stable. I don’t think that is a good basis to choose, on stability.18:29
miloajrosser: about the br-vlan problem, I test a distro deploy adding "ipforward: true" to user_variables.yml and... it works :) so perhaps we only need to add something in the doc and not change de code :)18:29
raukadahjrosser: this topic is kind of debatable, it depends on users what they want!18:30
raukadahand what works for them!18:30
jrosserIndeed like I say pros and cos both ways, I don’t want a flamewar18:30
raukadahjrosser: :-) I totally agree :-)18:30
miloaSorry, I have to go, many thanks to all for your help and advices :) see you tomorrow.18:32
*** miloa has quit IRC18:35
*** sdake has joined #openstack-ansible18:44
*** cshen has joined #openstack-ansible18:47
*** markvoelker has joined #openstack-ansible18:47
*** cshen has quit IRC18:51
*** palendae has joined #openstack-ansible18:52
*** DanyC has joined #openstack-ansible19:10
*** pcaruana has quit IRC19:14
*** markvoelker has quit IRC19:21
*** DanyC has quit IRC19:28
*** KenCrandall has joined #openstack-ansible19:32
openstackgerritNicolas Bock proposed openstack/openstack-ansible-openstack_hosts master: Add vendor override for openSUSE repositories  https://review.openstack.org/63974519:45
*** sdake has quit IRC20:07
*** goldenfri has joined #openstack-ansible20:23
*** DanyC has joined #openstack-ansible20:29
*** Leo_m has joined #openstack-ansible20:29
*** hamzy has quit IRC20:39
*** hamzy has joined #openstack-ansible20:43
*** strattao has quit IRC20:44
*** cshen has joined #openstack-ansible20:47
*** cshen has quit IRC20:52
*** bruvik has quit IRC20:56
openstackgerritMerged openstack/openstack-ansible master: Convert include to either import_playbook or include_tasks  https://review.openstack.org/63856220:59
openstackgerritMerged openstack/openstack-ansible master: Remove broken conditions when including playbooks  https://review.openstack.org/63857521:00
*** sdake has joined #openstack-ansible21:03
*** spatel has joined #openstack-ansible21:17
spateljamesdenton: what is that tweet about?21:17
*** sdake has quit IRC21:23
spotzbad day at Rackspace:(21:24
*** sdake has joined #openstack-ansible21:25
*** vnogin has joined #openstack-ansible21:36
*** ansmith has quit IRC21:38
*** ppetit has quit IRC21:40
d34dh0r53yep, rough day21:42
*** hamzy has quit IRC21:47
*** markvoelker has joined #openstack-ansible21:48
spotzyou ok d34dh0r53?21:49
d34dh0r53yeah :( really sad day though21:49
spotzI know:(21:49
*** sdake has quit IRC22:13
*** markvoelker has quit IRC22:22
* mnaser hands hugs22:24
*** cshen has joined #openstack-ansible22:47
*** cshen has quit IRC22:52
openstackgerritMerged openstack/openstack-ansible-os_nova stable/rocky: Drop SELinux support for CentOS 7  https://review.openstack.org/63961722:53
*** DanyC has quit IRC22:53
*** luksky has joined #openstack-ansible22:55
*** luksky has quit IRC23:03
*** eumel8 has quit IRC23:05
*** markvoelker has joined #openstack-ansible23:18
*** markvoelker has quit IRC23:24
*** markvoelker has joined #openstack-ansible23:24
*** shananigans has quit IRC23:34
*** ansmith has joined #openstack-ansible23:38
*** fdz has quit IRC23:44
*** Leo_m has quit IRC23:53
*** hwoarang has quit IRC23:57
*** hwoarang has joined #openstack-ansible23:59

Generated by irclog2html.py 2.15.3 by Marius Gedminas - find it at mg.pov.lt!