| *** spatel has quit IRC | 00:01 | |
| *** luksky has quit IRC | 00:04 | |
| *** MickyMan77 has quit IRC | 00:04 | |
| *** MickyMan77 has joined #openstack-ansible | 00:05 | |
| *** tosky has quit IRC | 00:20 | |
| *** jamesdenton has quit IRC | 00:48 | |
| *** jamesden_ has joined #openstack-ansible | 00:49 | |
| *** cp- has joined #openstack-ansible | 00:52 | |
| openstackgerrit | Merged openstack/openstack-ansible-os_keystone master: Add openstack-ansible-uw_apache focal job https://review.opendev.org/c/openstack/openstack-ansible-os_keystone/+/754123 | 01:05 |
|---|---|---|
| *** watersj has joined #openstack-ansible | 01:10 | |
| openstackgerrit | likui proposed openstack/openstack-ansible master: Update doc8 version https://review.opendev.org/c/openstack/openstack-ansible/+/765695 | 01:19 |
| *** watersj has quit IRC | 01:55 | |
| *** spatel has joined #openstack-ansible | 03:29 | |
| *** spatel has quit IRC | 03:32 | |
| *** vesper11 has quit IRC | 03:34 | |
| *** vesper11 has joined #openstack-ansible | 03:35 | |
| *** d34dh0r53 has joined #openstack-ansible | 03:59 | |
| *** jamesden_ has quit IRC | 04:34 | |
| *** jamesdenton has joined #openstack-ansible | 04:35 | |
| *** vesper has joined #openstack-ansible | 04:36 | |
| *** vesper11 has quit IRC | 04:37 | |
| *** raukadah is now known as chandankumar | 05:17 | |
| *** evrardjp has quit IRC | 05:33 | |
| *** evrardjp has joined #openstack-ansible | 05:33 | |
| *** chandankumar is now known as chkumar|ruck | 06:09 | |
| *** SiavashSardari has joined #openstack-ansible | 06:12 | |
| *** chkumar|ruck is now known as chkumar|rover | 06:14 | |
| *** chkumar|rover is now known as chkumar|ruck | 06:35 | |
| *** pto has joined #openstack-ansible | 07:01 | |
| *** miloa has joined #openstack-ansible | 07:21 | |
| *** luksky has joined #openstack-ansible | 07:40 | |
| *** sep has joined #openstack-ansible | 07:50 | |
| *** kleini has joined #openstack-ansible | 08:06 | |
| *** cshen has joined #openstack-ansible | 08:12 | |
| *** shyamb has joined #openstack-ansible | 08:14 | |
| *** pcaruana has joined #openstack-ansible | 08:15 | |
| openstackgerrit | Andrew Bonney proposed openstack/openstack-ansible master: Ensure kuryr repo is available within CI images https://review.opendev.org/c/openstack/openstack-ansible/+/765765 | 08:27 |
| *** andrewbonney has joined #openstack-ansible | 08:28 | |
| *** mensis has joined #openstack-ansible | 08:28 | |
| openstackgerrit | Andrew Bonney proposed openstack/openstack-ansible-os_zun master: Update zun role to match current requirements https://review.opendev.org/c/openstack/openstack-ansible-os_zun/+/763141 | 08:30 |
| *** stduolc1 has quit IRC | 08:36 | |
| *** rpittau|afk is now known as rpittau | 08:46 | |
| *** shyamb has quit IRC | 09:01 | |
| *** shyamb has joined #openstack-ansible | 09:05 | |
| noonedeadpunk | mornigns | 09:06 |
| noonedeadpunk | jrosser: how do you think - should I branch galera-client repo or should we retire it? | 09:19 |
| noonedeadpunk | I think worth branching as no time for retirement... | 09:20 |
| noonedeadpunk | ah, and we need deprecation of it not retitrement... | 09:22 |
| jrosser | yeah we shouldnt retire | 10:05 |
| noonedeadpunk | just forgot about difference between these 2 options at the moment of writing | 10:05 |
| noonedeadpunk | seems zun super close | 10:06 |
| *** mensis has quit IRC | 10:06 | |
| andrewbonney | Hopefully just a battle with the Docker rate limits remaining :) | 10:06 |
| jrosser | yes it is really close | 10:06 |
| jrosser | there seems to be some things could be better in the zun tempest tests but thats not an OSA problem | 10:07 |
| jrosser | i think andrewbonney is reasonably happy with the patches now and we will look at testing in the lab rather than AIO next | 10:08 |
| noonedeadpunk | and tempest passes | 10:10 |
| noonedeadpunk | agree, it's really great work | 10:10 |
| jrosser | strange fail on centos there | 10:12 |
| noonedeadpunk | it's currently run tempest | 10:12 |
| noonedeadpunk | ah well thats were it failed | 10:13 |
| noonedeadpunk | and failed just single test | 10:15 |
| jrosser | error while evaluating conditional ((groups['oslomsg_rpc_all'] | length) < 1): 'dict object' has no attribute 'oslomsg_rpc_all' | 10:15 |
| jrosser | which is kind of true i think becasue not sure where oslomsg_rpc_all really is defined | 10:16 |
| jrosser | https://zuul.opendev.org/t/openstack/build/584c607c6b534f8aaa74dc59f00f4f87/log/job-output.txt#13236 | 10:16 |
| noonedeadpunk | well it's upgrade jobs.... | 10:16 |
| noonedeadpunk | and we got things broken on U kind of... | 10:17 |
| jrosser | yes i think the role really is unusable on U | 10:18 |
| noonedeadpunk | so we probably want to packport things to U as well? | 10:19 |
| andrewbonney | I haven't tested with U yet, but hopefully it'll backport cleanly | 10:20 |
| jrosser | we'd need the kuryr fix backporting as well wouldnt we? | 10:21 |
| andrewbonney | They're all pulling things from master at the moment (admittedly non-ideal) as they're pinned within the role | 10:22 |
| jrosser | in order to keep the versions controlled properly for the victoria release i think we'll need another section here https://github.com/openstack/openstack-ansible/blob/master/playbooks/defaults/repo_packages/openstack_services.yml#L347-L351 | 10:27 |
| noonedeadpunk | oh, yes | 10:28 |
| andrewbonney | Ok, I'll add that | 10:28 |
| noonedeadpunk | Worth adjusting https://review.opendev.org/c/openstack/openstack-ansible/+/765765 - it's failing currently anyway | 10:29 |
| noonedeadpunk | let's jsut wait for 763141 to pass first | 10:29 |
| noonedeadpunk | nah, centos has failed again | 10:30 |
| noonedeadpunk | maybe we need smth like `tempest_run_concurrency: 0` in https://opendev.org/openstack/openstack-ansible/src/branch/master/tests/roles/bootstrap-host/templates/user_variables_zun.yml.j2 | 10:31 |
| jrosser | interesting ansible docs don't seem to mention playbook defaults in variable precedence | 10:31 |
| *** spatel has joined #openstack-ansible | 10:35 | |
| *** spatel has quit IRC | 10:40 | |
| openstackgerrit | Dmitriy Rabotyagov proposed openstack/openstack-ansible-galera_client master: Deprecate openstack-ansible-galera_client role https://review.opendev.org/c/openstack/openstack-ansible-galera_client/+/765779 | 10:41 |
| openstackgerrit | Merged openstack/openstack-ansible master: Update doc8 version https://review.opendev.org/c/openstack/openstack-ansible/+/765695 | 10:47 |
| *** yann-kaelig has joined #openstack-ansible | 11:04 | |
| *** tosky has joined #openstack-ansible | 11:22 | |
| openstackgerrit | Andrew Bonney proposed openstack/openstack-ansible-os_zun master: Update zun role to match current requirements https://review.opendev.org/c/openstack/openstack-ansible-os_zun/+/763141 | 11:26 |
| admin0 | morning \o | 11:40 |
| *** shyamb has quit IRC | 11:46 | |
| *** shyamb has joined #openstack-ansible | 11:48 | |
| *** mensis has joined #openstack-ansible | 11:50 | |
| *** masterpe has joined #openstack-ansible | 11:53 | |
| masterpe | afternoon | 11:53 |
| *** rfolco has joined #openstack-ansible | 11:54 | |
| *** yann-kaelig has quit IRC | 12:00 | |
| jrosser | o/ hello | 12:03 |
| admin0 | masterpe, thank you for the config .. i made some progress .. br-lbaas had to be in compute node also, else the agent complained and died: 2020-12-06 20:59:44.974 86303 ERROR neutron.plugins.ml2.drivers.linuxbridge.agent.linuxbridge_neutron_agent [-] Bridge br-lbaas for physical network lbaas does not exist. Agent terminated! | 12:07 |
| admin0 | so after creating the agent in all , amphora is up .. | 12:07 |
| admin0 | but its in its own bridge like an isolated island | 12:07 |
| admin0 | will be doing 1 more test to confirm | 12:07 |
| *** odyssey4me has joined #openstack-ansible | 12:15 | |
| masterpe | admin0: so you created the br-lbaas bridge youself via netplan? | 12:29 |
| noonedeadpunk | jrosser: should I wait for zun stuff to land before branching? I think it's pretty important to wait for it... | 12:35 |
| *** priteau has quit IRC | 12:40 | |
| *** shyamb has quit IRC | 12:47 | |
| *** spatel has joined #openstack-ansible | 12:52 | |
| *** cshen has quit IRC | 12:53 | |
| jrosser | noonedeadpunk: centos fails due to missing multipathd config and service not being enabled, just looking at that here | 12:56 |
| *** spatel has quit IRC | 12:57 | |
| *** shyamb has joined #openstack-ansible | 12:57 | |
| jrosser | so either that gets fixed today or we can make centos8 nv and backport a fix, should probably try not to wait too much longer | 12:57 |
| jrosser | it looks like there is a similar thing in the nova role where debuntu gets multipathd installed but centos does not | 12:58 |
| noonedeadpunk | Yeah was also thinking as making it nv just to merge | 13:06 |
| noonedeadpunk | btw, I've placed patches to revive monasca roles - mensis said he have fully working roles for U, and today pinged me about it :( | 13:07 |
| noonedeadpunk | but I think we should leave it for W or branch afterwards | 13:07 |
| noonedeadpunk | it's not smth I will be waiting for | 13:07 |
| admin0 | masterpe, this is what i have so far: https://gist.github.com/a1git/27e47c7ebcd6f65566663226ba4019eb | 13:09 |
| *** mgariepy has quit IRC | 13:14 | |
| *** mgariepy has joined #openstack-ansible | 13:19 | |
| *** priteau has joined #openstack-ansible | 13:27 | |
| *** shyamb has quit IRC | 13:34 | |
| *** shyamb has joined #openstack-ansible | 13:35 | |
| openstackgerrit | Andrew Bonney proposed openstack/openstack-ansible master: Ensure kuryr repo is available within CI images https://review.opendev.org/c/openstack/openstack-ansible/+/765765 | 13:40 |
| *** CeeMac has joined #openstack-ansible | 13:44 | |
| *** miloa has quit IRC | 14:10 | |
| *** shyamb has quit IRC | 14:18 | |
| *** owalsh has quit IRC | 14:19 | |
| mgariepy | jrosser, https://review.opendev.org/c/openstack/nova/+/758761 | 14:25 |
| admin0 | masterpe, yes .. the bridges are created self .. else neutron complained bridge not found even when the octavia playbooks were not run | 14:38 |
| admin0 | sorry for late reply .. was eating | 14:38 |
| *** owalsh has joined #openstack-ansible | 14:42 | |
| *** owalsh has quit IRC | 14:47 | |
| *** cshen has joined #openstack-ansible | 14:49 | |
| openstackgerrit | Dmitriy Rabotyagov proposed openstack/ansible-role-systemd_service master: Add possibility to configure systemd sockets https://review.opendev.org/c/openstack/ansible-role-systemd_service/+/763194 | 14:49 |
| openstackgerrit | Dmitriy Rabotyagov proposed openstack/ansible-role-systemd_service master: Add option to create systemd native service overrides https://review.opendev.org/c/openstack/ansible-role-systemd_service/+/765815 | 14:51 |
| *** owalsh has joined #openstack-ansible | 14:58 | |
| *** spatel has joined #openstack-ansible | 15:11 | |
| openstackgerrit | Dmitriy Rabotyagov proposed openstack/ansible-role-systemd_service master: Add option to create systemd native service overrides https://review.opendev.org/c/openstack/ansible-role-systemd_service/+/765815 | 15:14 |
| *** SiavashSardari has quit IRC | 15:44 | |
| *** ajg20 has joined #openstack-ansible | 15:44 | |
| *** ajg20 has quit IRC | 15:56 | |
| *** jamesdenton has quit IRC | 15:58 | |
| *** jamesdenton has joined #openstack-ansible | 15:59 | |
| admin0 | masterpe, can you show be a brctl output of a amphora instance you have, on how exactly the bridge connections have | 16:03 |
| admin0 | how exactly its supposed to work/connect | 16:04 |
| noonedeadpunk | jrosser: feels we have new resolver? https://b50ef5623c6036d09535-7575ab0330f7da34328cde6b7d597146.ssl.cf2.rackcdn.com/763194/5/check/openstack-ansible-linters/703ce8e/job-output.txt | 16:07 |
| noonedeadpunk | but can't reproduce locally... | 16:07 |
| jrosser | hrrm we fixed this in one place, which wasfor the neutron functional tests i remember | 16:10 |
| jrosser | https://github.com/openstack/openstack-ansible-tests/commit/aaecefaee91d468f6f96cac26bc8a119d5fd2dca | 16:10 |
| jrosser | oh | 16:10 |
| jrosser | right so because the systemd_service role won't be testing by using python_venv_build i expect we need a different fix | 16:11 |
| noonedeadpunk | I'm not sure linters does use it though | 16:11 |
| noonedeadpunk | but at this point I see no pip upgrade in log... | 16:12 |
| jrosser | this is a bionic node so will be running ancient pip 9.x | 16:13 |
| jrosser | but i guess the error comes from pip in the tox env and i'm not sure how that is getting there | 16:17 |
| noonedeadpunk | well on sandbox I'm getting 20.0.2 pip | 16:18 |
| noonedeadpunk | but it's the same as I had system-wise | 16:18 |
| jrosser | this looks kind of messed up https://github.com/openstack/ansible-role-systemd_service/blob/master/tox.ini | 16:19 |
| jrosser | seems really arbitrary which tox envs use upper constraints and which dont | 16:19 |
| jrosser | what i was going to suggest was putting another -c in here https://github.com/openstack/ansible-role-systemd_service/blob/master/tox.ini#L37 | 16:21 |
| jrosser | pointing to this https://github.com/openstack/openstack-ansible/blob/master/global-requirement-pins.txt | 16:22 |
| noonedeadpunk | well we would need to do this everywhere then | 16:22 |
| noonedeadpunk | as neutron also fails now | 16:23 |
| jrosser | that seems to indicate the CI image is different | 16:23 |
| jrosser | or whatever tox uses to build the venv is now different | 16:23 |
| jrosser | the/its | 16:23 |
| noonedeadpunk | well, error is different there.. | 16:23 |
| jrosser | so we have a very new version of virtualenv, only 5 hours old | 16:24 |
| jrosser | yes thats it "Bump pip to 20.3.1, setuptools to 51.0.0 and wheel to 0.36.1 - by @gaborbernat. (#2029)" | 16:25 |
| noonedeadpunk | uh... | 16:27 |
| noonedeadpunk | well, probably we should just ensure tox version somewhere in https://opendev.org/openstack/openstack-ansible-tests/src/branch/master/test-ansible-env-prep.sh | 16:28 |
| jrosser | well when i say "thats it" i can see how the new resolver is getting there | 16:28 |
| jrosser | perhaps not seeing why you cant reproduce it | 16:28 |
| *** macz_ has joined #openstack-ansible | 16:31 | |
| *** fresta has quit IRC | 16:36 | |
| *** fresta has joined #openstack-ansible | 16:37 | |
| *** macz_ has quit IRC | 16:41 | |
| admin0 | the amphora instance is up ..one lb is in pending_create, without any logs | 16:43 |
| *** macz_ has joined #openstack-ansible | 16:44 | |
| *** fresta has quit IRC | 16:45 | |
| *** fresta has joined #openstack-ansible | 16:47 | |
| *** fresta has quit IRC | 16:47 | |
| openstackgerrit | Jonathan Rosser proposed openstack/ansible-role-systemd_service master: Use upper-constraints for all tox environments https://review.opendev.org/c/openstack/ansible-role-systemd_service/+/765831 | 16:54 |
| guilhermesp | admin0: are you ocatvia containers able to reach the amphoras? you can see the status of amphoras by "openstack loadbalancer amphora list" | 16:54 |
| jrosser | noonedeadpunk: looks like integrated repo linters also fail https://storage.gra.cloud.ovh.net/v1/AUTH_dcaab5e32b234d56b626f72581e3644c/zuul_opendev_logs_f33/765765/2/check/openstack-ansible-linters/f336a98/job-output.txt | 16:57 |
| noonedeadpunk | uh | 16:57 |
| noonedeadpunk | but they failed differently | 16:58 |
| jrosser | yeah they did | 16:58 |
| *** rpittau is now known as rpittau|afk | 16:58 | |
| admin0 | guilhermesp, i see 1 in booting phase | 16:59 |
| admin0 | for hours | 16:59 |
| guilhermesp | and the instance is active and running? ports allocated and etc? | 17:00 |
| guilhermesp | this sounds to me that octavia health manger is not able to contact the amphoras | 17:00 |
| jrosser | admin0: did you wire br-lbass to the underlying interface? | 17:01 |
| admin0 | i had them via netplan | 17:01 |
| jrosser | and you see that with brctl? becasue your paste earlier only showed on member on the bridge i think? | 17:02 |
| admin0 | you mean for every amphora crated, i have to manually add a connection between that bridge and br-lbaas ? | 17:02 |
| jrosser | neutron will plug each amphora onto the bridge | 17:02 |
| jrosser | but if the bridge does not connect off the host then they won't be able to talk to the controllers | 17:03 |
| *** fresta has joined #openstack-ansible | 17:03 | |
| guilhermesp | on computes, you should have this configuration `physical_interface_mappings = <provider_network>:<bond>` | 17:05 |
| admin0 | let me go step by step .. just like we create br-mgmt br-storage etc, i have br-lbaas in every node ( controllers, compute,network) | 17:05 |
| guilhermesp | if the bonding is correctly wired, neutron should create the bridges and you should be able yo reach then through octavia containers | 17:05 |
| admin0 | guilhermesp, there is where i am lost i think .. in the 1st step .. who creates br-lbaas ? | 17:07 |
| admin0 | i do right ? | 17:07 |
| admin0 | on all controllers, compute and network node | 17:07 |
| guilhermesp | yes, but i dont think you need on computes. The only reason to create the br-lbaas on controllers is to make lxc able to create the bridges for the containers. So octavia container are able to connect to port on computes/network nodes through your bonding | 17:09 |
| guilhermesp | i suggest you to test your link first between controllers and computes | 17:10 |
| admin0 | guilhermesp, when i did not create in computes, it happend as soon as i ran neutron playbook: ERROR neutron.plugins.ml2.drivers.linuxbridge.agent.linuxbridge_neutron_agent [-] Bridge br-lbaas for physical network lbaas does not exist. Agent terminated! | 17:10 |
| guilhermesp | probably your linux bridge agent on computes are mapping to br-lbaas. which doestn exists afaik | 17:11 |
| *** klamath_atx has quit IRC | 17:11 | |
| jrosser | br-lbaas needs to exist on the computes for a flat network type | 17:11 |
| jrosser | not for vlan | 17:11 |
| admin0 | do we need vlan or do we need flat ? | 17:12 |
| jrosser | :) I said many times it is up to you | 17:12 |
| jrosser | you can choose, then you make config to suit which one you want | 17:12 |
| admin0 | yes, but i need just one that works .. and i have not been able to make any work | 17:12 |
| admin0 | say we chose flat | 17:12 |
| admin0 | because its easier for operators like me to have one extra br-lbaas just like br-mgmt etc | 17:12 |
| admin0 | that makes it easy | 17:13 |
| admin0 | so if we do flat, the IP on br-lbaas is also unrouted right ? like br-mgmt ? | 17:13 |
| admin0 | so far, so good on that .. .. create br-lbaas on all, just treat it like br-mgmt, give it an unrouted ip range on /22 .. that i did | 17:13 |
| jrosser | there is no IP on br-lbaas | 17:14 |
| * jrosser not sure thats what you mean? | 17:14 | |
| admin0 | the docs have 172.29.232.0/22 as an ip range for br-lbaas if i recall | 17:15 |
| admin0 | so we don't need ips on br-lbaas ? it is just a bridge like br-vlan ? | 17:18 |
| jrosser | correct, on a compute node it is a L2 bridge that allows the amphora to talk to the controller | 17:19 |
| jrosser | for a flat network | 17:19 |
| jrosser | if you make it vlan none of this is needed at all and it's just a regular neutron provider network | 17:19 |
| admin0 | https://developer.rackspace.com/docs/private-cloud/rpc/master/rpc-octavia-internal/octavia-install-guide/ | 17:19 |
| admin0 | for a vlan based one, can you give me a few pointers ? like how it would look . gist/pastebin if possible | 17:20 |
| admin0 | i have br-vlan and lets assume we can use vlan200 for this | 17:20 |
| jrosser | here is an example https://satishdotpatel.github.io//openstack-ansible-octavia/ | 17:24 |
| jrosser | there is no 'right way' to do this, you just need to have connectivity from the amphora to the containers on the controllers | 17:25 |
| jrosser | everyone has a slightly different approach in some way because all deployments are a little different | 17:25 |
| admin0 | does the br-lbaas have an ip ? | 17:27 |
| admin0 | in vlan example ? | 17:27 |
| jrosser | no, the IP are on eth14 in the octavia containers | 17:28 |
| jrosser | the bridge is just for LXC to plug those onto | 17:28 |
| admin0 | ok | 17:28 |
| spatel | admin0: In my case i do have IP on br-lbaas but that bridge isn't attached to any physical interface | 17:31 |
| spatel | reason i put ip just to do ping test, its not required to have IP | 17:32 |
| spatel | My octavia working great without any issue the way i have deployed. | 17:33 |
| admin0 | so suppose my vlan is on eth4 .. and i want to use vlan20, in that case, in controllers, i create br-lbaas which is on top of eth4.20 ? | 17:33 |
| admin0 | or is it just an empty bridge with nothing underlying it ? | 17:33 |
| admin0 | ok | 17:35 |
| admin0 | empty bridge .. got it | 17:35 |
| admin0 | i will try :) | 17:35 |
| admin0 | rebuilding again | 17:35 |
| spatel | admin0: check out this thread also - http://lists.openstack.org/pipermail/openstack-discuss/2020-September/017598.html | 17:35 |
| spatel | just create br-lbaas bridge file but don't attach it and when you run my veth script it will map it with br-vlan and your lbaas VLAN which you going to specify in veth script. | 17:37 |
| admin0 | ok | 17:37 |
| admin0 | give me a few hrs to rebuild and come back :) | 17:37 |
| spatel | cool! | 17:38 |
| jrosser | noonedeadpunk: perhaps we pin virtualenv back for now just here https://github.com/openstack/openstack-ansible-tests/blob/bb630f047f4807451fa4ecb4ab2539458f659b95/run_tests_common.sh#L79 | 17:40 |
| * jrosser heads out | 17:43 | |
| *** gyee has joined #openstack-ansible | 17:54 | |
| noonedeadpunk | yeah, good idea | 17:58 |
| openstackgerrit | Dmitriy Rabotyagov proposed openstack/openstack-ansible-tests master: Bump virtualenv to version prior to 20.2.2 https://review.opendev.org/c/openstack/openstack-ansible-tests/+/765839 | 18:01 |
| jrosser | I also could not reproduce the linters error here in a vm | 18:07 |
| *** andrewbonney has quit IRC | 18:17 | |
| admin0 | spatel, so the net_name and ip_from_q has no issues if its named lbaas or octavia ? | 18:35 |
| admin0 | you have named it octavia, diff from all other examples | 18:35 |
| admin0 | but i could not find how it will be used later on | 18:35 |
| *** mmercer has joined #openstack-ansible | 18:36 | |
| spatel | truly speaking i have no idea what ip_from_q: "octavia" does, i just get that directly from OSA official example | 18:36 |
| spatel | jrosser: or noonedeadpunk may know about that | 18:37 |
| admin0 | i think i can replicate your setup .. will use vlan 27 as well :D | 18:38 |
| admin0 | just to make sure i am not making any mistake | 18:38 |
| admin0 | spatel, so if the br-lbaas does not need any ips, then lbaas: 172.27.40.0/24 -- is not also required right ? | 18:42 |
| admin0 | sorry .. what is this for : octavia_container_network_name: lbaas_address | 18:44 |
| admin0 | will copy for now and worry about it later :) | 18:46 |
| *** cshen has quit IRC | 18:46 | |
| spatel | octavia_container_network_name: lbaas_address its related to lbaas: 172.27.40.0/24 | 18:48 |
| admin0 | your exact copy failed because you have octavia in net-name and i_from_q while you mentioned lbaas on top | 18:51 |
| admin0 | matching those 3 made it worked | 18:51 |
| admin0 | i also found that not having horizon_enable_neutron_lbaas: False will break horizon | 18:52 |
| jrosser | admin0: would be great if you can patch/fix that horizon issue | 18:55 |
| admin0 | let me make this work first .. been on it for 2+ weeks fulltime now | 18:56 |
| admin0 | but most of the time was spent rebuilding platform when it broke | 18:56 |
| jrosser | imho go slower and debug | 18:56 |
| jrosser | this does not just magically work by running with osa defaults, kind of similar to magnum | 18:57 |
| spatel | admin0: you are right about i_from_q, let me fix my documentation | 19:00 |
| spatel | i do have lbaas in production | 19:00 |
| spatel | jrosser: what is ip_from_q ? | 19:02 |
| spatel | admin0: i have fixed my blog | 19:04 |
| admin0 | been with osa for 8 years now .. i kind of expect things to run smoothly when its included .. | 19:17 |
| admin0 | the defaults just work flawlessly .. so the expectation is high | 19:17 |
| *** cshen has joined #openstack-ansible | 19:21 | |
| jrosser | spatel: it tells the dynamic inventory which ip range to take addresses from when assigning them to container interfaces “ip from queue” | 19:25 |
| jrosser | so for octavia his will typically only be needing 3, and the rest of the cidr should be covered by “used_ips” | 19:26 |
| jrosser | that used_ips range can be given to neutron as the set it’s allowed to use for the amphora | 19:26 |
| spatel | does it matter if i change ip_from_q: foo ? | 19:27 |
| jrosser | for an existing deployment maybe? | 19:27 |
| jrosser | not sure | 19:27 |
| admin0 | what i did was created the br-lbaas via netplan, then used the script from spatel to connect it to vlan27 | 19:31 |
| admin0 | now setting up the rest of the infra/openstack so that i have a clean system to test this | 19:31 |
| jrosser | if you did put an ip on the bridge your first validation could be to check the bridges all can ping each other | 19:33 |
| jrosser | if not, no matter, but check eth14<>eth14 in your octavia containers after setup_hosts.yml | 19:34 |
| admin0 | i only am using 1 controller for this | 19:35 |
| admin0 | i mean this test setup is c1 + n1 + h1 ( controller1, network node1, hypervisor1) | 19:35 |
| jrosser | ok, no problem. you can always add ip temporarily on n1/h1 to check the underlying connectivity from the container | 19:36 |
| admin0 | so far, i have only created the bridge in the controller | 19:39 |
| admin0 | do i need to create the bridge in the network also ? | 19:39 |
| jrosser | oh err no sorry | 19:41 |
| jrosser | not for vlan | 19:41 |
| admin0 | ok | 19:43 |
| *** luksky has quit IRC | 19:54 | |
| *** luksky has joined #openstack-ansible | 19:54 | |
| spatel | jrosser: thanks for explanation | 19:56 |
| spotz | jrosser: We've released Victoria no? | 20:02 |
| jrosser | not yet :/ | 20:02 |
| spotz | jrosser: hehe good thing I asked and didn't go patch docs:) | 20:02 |
| jrosser | branching real soon for rc | 20:02 |
| spotz | Ok mI'll keep an eye out and patch on release:) | 20:04 |
| spatel | jrosser: +1 for Victoria release. (hope tomorrow) | 20:09 |
| *** jamesdenton has quit IRC | 20:15 | |
| *** jamesdenton has joined #openstack-ansible | 20:15 | |
| jrosser | spatel: it will be a release candidate first | 20:26 |
| jrosser | everyone testing ++ | 20:26 |
| spatel | jrosser: but we can keep rolling out minor release until reach stable right ? | 20:26 |
| jrosser | rc1, rc2..... | 20:27 |
| jrosser | the release proper when happy | 20:27 |
| jrosser | seems to be confusion here | 20:27 |
| jrosser | currently there is no stable/victoria branch for osa, that will be done very soon | 20:27 |
| jrosser | the tag will be a release candidate, not a 'proper release' | 20:28 |
| spatel | I am ok to deploy rc1 and then when stable come out i can do quick minor upgrade to stable | 20:28 |
| spatel | jrosser: that is doable right? | 20:29 |
| jrosser | sure | 20:29 |
| jrosser | it's just a tag on the same branch either way | 20:29 |
| spatel | perfect! | 20:29 |
| *** nurdie__ has quit IRC | 20:58 | |
| *** watersj has joined #openstack-ansible | 21:04 | |
| watersj | what network switches are you using and how do you manage various vLANs (and or vxlans) you create? | 21:13 |
| *** luksky has quit IRC | 21:14 | |
| admin0 | watersj, what i do is tag a block of vlans in the switch .. like 3000-3025 etc .. and that would be enough for all current/future needs | 21:17 |
| *** jamesdenton has quit IRC | 21:18 | |
| admin0 | mgmt, vxlan, storage, replication = 4 already .. | 21:18 |
| admin0 | and some for the external networks | 21:18 |
| watersj | admin0, trunk access port for you infra and nova nodes | 21:18 |
| *** jamesdenton has joined #openstack-ansible | 21:18 | |
| watersj | makes sense, what about ironic provisioning | 21:19 |
| admin0 | not using ironic (yet) | 21:22 |
| admin0 | would love a good howto on it | 21:22 |
| admin0 | some switches allow tagged vlans with ports on some default vlan | 21:22 |
| admin0 | so ssh goes on the untagged on, rest all are tagged | 21:22 |
| watersj | we are going to have multiple customer that will have dynamically allocated baremetal for. Can't have all ports tag for those nodes, but nova nodes and infra sure | 21:25 |
| *** luksky has joined #openstack-ansible | 21:33 | |
| *** watersj has quit IRC | 21:42 | |
| spatel | ironic required some special PXE vlan etc.. to make it work | 21:47 |
| *** sshnaidm has quit IRC | 21:49 | |
| *** csmart has quit IRC | 22:02 | |
| *** odyssey4me has quit IRC | 22:02 | |
| *** johnsom has quit IRC | 22:02 | |
| *** jhesketh has quit IRC | 22:06 | |
| *** jhesketh has joined #openstack-ansible | 22:08 | |
| *** sshnaidm has joined #openstack-ansible | 22:08 | |
| *** mensis has quit IRC | 22:08 | |
| *** tobberydberg has quit IRC | 22:14 | |
| *** tobberydberg has joined #openstack-ansible | 22:19 | |
| *** pcaruana has quit IRC | 22:21 | |
| *** cshen has quit IRC | 22:28 | |
| *** johnsom has joined #openstack-ansible | 22:30 | |
| *** csmart has joined #openstack-ansible | 22:31 | |
| *** rfolco has quit IRC | 22:34 | |
| *** spatel has quit IRC | 22:40 | |
| *** rfolco has joined #openstack-ansible | 22:43 | |
| *** rfolco has quit IRC | 22:48 | |
| *** ThiagoCMC has joined #openstack-ansible | 23:09 | |
| *** ThiagoCMC has quit IRC | 23:34 | |
| *** ThiagoCMC has joined #openstack-ansible | 23:36 | |
| *** cshen has joined #openstack-ansible | 23:45 | |
| *** cshen has quit IRC | 23:49 | |
| *** luksky has quit IRC | 23:51 | |
| *** spatel has joined #openstack-ansible | 23:57 | |
Generated by irclog2html.py 2.17.2 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!