| *** yifei has quit IRC | 00:01 | |
| *** PramodJ has joined #openstack-ansible | 00:05 | |
| *** chyka_ has quit IRC | 00:07 | |
| *** markvoelker_ has quit IRC | 00:07 | |
| *** markvoelker has joined #openstack-ansible | 00:08 | |
| *** pramodrj07 has quit IRC | 00:08 | |
| *** yifei has joined #openstack-ansible | 00:11 | |
| *** markvoelker has quit IRC | 00:12 | |
| *** weezS has quit IRC | 00:13 | |
| *** markvoelker has joined #openstack-ansible | 00:18 | |
| *** dxiri has quit IRC | 00:42 | |
| *** dxiri has joined #openstack-ansible | 00:42 | |
| *** dave-mccowan has joined #openstack-ansible | 00:47 | |
| *** jiaopengju has joined #openstack-ansible | 00:49 | |
| *** xingchao has joined #openstack-ansible | 00:54 | |
| *** xingchao has quit IRC | 00:54 | |
| *** xingchao has joined #openstack-ansible | 00:58 | |
| openstackgerrit | Jimmy McCrory proposed openstack/openstack-ansible master: [DNM] Inventory migrate CLI https://review.openstack.org/504508 | 00:59 |
|---|---|---|
| openstackgerrit | Jimmy McCrory proposed openstack/openstack-ansible master: [DNM] Inventory migrate CLI https://review.openstack.org/504508 | 01:01 |
| *** woodard_ has joined #openstack-ansible | 01:03 | |
| *** woodard has quit IRC | 01:03 | |
| *** markvoelker has quit IRC | 01:03 | |
| *** markvoelker has joined #openstack-ansible | 01:03 | |
| *** PramodJ has quit IRC | 01:04 | |
| *** dxiri has quit IRC | 01:08 | |
| *** markvoelker has quit IRC | 01:08 | |
| *** hw_wutianwei has joined #openstack-ansible | 01:10 | |
| *** dxiri has joined #openstack-ansible | 01:15 | |
| *** snowman48391 has quit IRC | 01:18 | |
| openstackgerrit | Jimmy McCrory proposed openstack/openstack-ansible master: [DNM] Inventory migrate CLI https://review.openstack.org/504508 | 01:19 |
| *** dxiri has quit IRC | 01:20 | |
| *** snowman4839 has joined #openstack-ansible | 01:20 | |
| *** woodard_ has quit IRC | 01:46 | |
| *** vnogin has quit IRC | 01:53 | |
| *** dxiri has joined #openstack-ansible | 01:59 | |
| *** dxiri has quit IRC | 01:59 | |
| *** dxiri has joined #openstack-ansible | 01:59 | |
| *** sxc731 has joined #openstack-ansible | 01:59 | |
| *** kukacz has quit IRC | 02:00 | |
| *** galstrom_zzz is now known as galstrom | 02:00 | |
| *** vnogin has joined #openstack-ansible | 02:01 | |
| *** kukacz has joined #openstack-ansible | 02:01 | |
| *** vnogin has quit IRC | 02:03 | |
| *** vnogin has joined #openstack-ansible | 02:13 | |
| *** xingchao has quit IRC | 02:14 | |
| *** xingchao has joined #openstack-ansible | 02:17 | |
| captaindave | Hey all. I've installed successfully and things seem operational, but I'm getting what appear to be certificate errors in the nova-api-os-compute-container logs for nova-api-wsgi. It looks like it's unhappy about hitting the externally-facing keystone with its OSA-created self-signed cert. Interestingly, adding that cert to /etc/pki/ca-trust/source/anchors and running 'update-ca-trust extract' | 02:23 |
| captaindave | doesn't seem to change things. Is this expected behavior or am I doing something wrong? | 02:23 |
| *** xingchao has quit IRC | 02:24 | |
| captaindave | Example output: http://sprunge.us/SJDZ | 02:24 |
| *** xingchao has joined #openstack-ansible | 02:27 | |
| *** sxc731 has quit IRC | 02:37 | |
| *** chhavi has joined #openstack-ansible | 02:37 | |
| *** galstrom is now known as galstrom_zzz | 02:54 | |
| *** xingchao has quit IRC | 02:54 | |
| *** xingchao has joined #openstack-ansible | 02:56 | |
| *** xingchao has quit IRC | 02:58 | |
| *** thorst has joined #openstack-ansible | 03:00 | |
| *** xingchao has joined #openstack-ansible | 03:03 | |
| *** dave-mccowan has quit IRC | 03:05 | |
| *** udesale has joined #openstack-ansible | 03:16 | |
| *** xingchao has quit IRC | 03:19 | |
| *** xingchao has joined #openstack-ansible | 03:20 | |
| *** xingchao has quit IRC | 03:22 | |
| *** threestrands has quit IRC | 03:29 | |
| *** xingchao has joined #openstack-ansible | 03:30 | |
| *** vnogin has quit IRC | 03:33 | |
| *** vnogin has joined #openstack-ansible | 03:34 | |
| *** xingchao has quit IRC | 03:35 | |
| *** xingchao has joined #openstack-ansible | 03:37 | |
| *** vnogin has quit IRC | 03:38 | |
| *** xingchao has quit IRC | 03:40 | |
| *** thorst has quit IRC | 04:03 | |
| *** snowman48391 has joined #openstack-ansible | 04:10 | |
| *** snowman4839 has quit IRC | 04:11 | |
| *** albertcard1 has quit IRC | 04:18 | |
| *** gkadam_ has joined #openstack-ansible | 04:21 | |
| *** gkadam_ has quit IRC | 04:22 | |
| *** gkadam_ has joined #openstack-ansible | 04:22 | |
| *** sxc731 has joined #openstack-ansible | 04:22 | |
| *** gkadam_ is now known as gkadam | 04:23 | |
| *** agrebennikov has quit IRC | 04:37 | |
| *** dxiri has quit IRC | 05:08 | |
| *** dxiri has joined #openstack-ansible | 05:16 | |
| *** armaan has quit IRC | 05:17 | |
| *** armaan has joined #openstack-ansible | 05:18 | |
| *** armaan has quit IRC | 05:23 | |
| *** armaan has joined #openstack-ansible | 05:24 | |
| *** armaan has quit IRC | 05:30 | |
| *** armaan has joined #openstack-ansible | 05:30 | |
| captaindave | Also, I'm getting the following error on attempted reconfiguration of nova: http://sprunge.us/BWGd | 05:40 |
| captaindave | If I remove my inventory, it will fail but not kill the playbook. | 05:40 |
| *** taseer2 is now known as Taseer | 05:42 | |
| captaindave | adding a static: no seems to remedy it | 05:43 |
| captaindave | I take that back. Sorry. | 05:44 |
| captaindave | Wrong terminal. | 05:44 |
| captaindave | (the static part) | 05:44 |
| *** jbadiapa has joined #openstack-ansible | 05:52 | |
| *** pcaruana has joined #openstack-ansible | 05:58 | |
| *** armaan has quit IRC | 06:28 | |
| *** armaan has joined #openstack-ansible | 06:29 | |
| *** armaan has quit IRC | 06:30 | |
| *** armaan has joined #openstack-ansible | 06:30 | |
| *** PaulusNL has joined #openstack-ansible | 06:35 | |
| *** vnogin has joined #openstack-ansible | 06:37 | |
| *** vnogin has quit IRC | 06:41 | |
| *** sxc731 has quit IRC | 06:56 | |
| *** vnogin has joined #openstack-ansible | 07:01 | |
| *** bruvik_ has joined #openstack-ansible | 07:02 | |
| openstackgerrit | Hengqing Hu proposed openstack/openstack-ansible-os_neutron master: Set enable_dvr when router_distributed is enabled https://review.openstack.org/517187 | 07:05 |
| *** vnogin has quit IRC | 07:05 | |
| *** hachi has joined #openstack-ansible | 07:07 | |
| *** Oku_OS-away is now known as Oku_OS | 07:09 | |
| *** hachi__ has joined #openstack-ansible | 07:11 | |
| *** fxpester has quit IRC | 07:11 | |
| *** hachi has quit IRC | 07:13 | |
| *** chhavi__ has joined #openstack-ansible | 07:16 | |
| *** a0x3e8 has joined #openstack-ansible | 07:18 | |
| a0x3e8 | hello. I have a problem with new a fresh install of openstack test environment using ansible. after making a local git and pip repository, install finished without any errors. all the services are up and running. but when I try to start an instance, I get the error NOT ENOUGH HOSTS. I checked block storage, nova hypervisor, and essentially anything | 07:22 |
| a0x3e8 | . there is no error in nova log about hypervisor. how can I debug this. | 07:22 |
| *** bruvik_ has quit IRC | 07:25 | |
| *** vnogin has joined #openstack-ansible | 07:31 | |
| *** huxinhui__ has joined #openstack-ansible | 07:32 | |
| *** bruvik_ has joined #openstack-ansible | 07:33 | |
| *** huxinhui__ is now known as huxinhui | 07:34 | |
| *** jvidal has joined #openstack-ansible | 07:34 | |
| *** bruvik_ has quit IRC | 07:38 | |
| *** vnogin has quit IRC | 07:38 | |
| berendt | mhayden: regarding https://review.openstack.org/#/c/515634/, to you prefer to disable debsum by default? will add docs/release notes for the new parameter now | 07:46 |
| hwoarang | good morning | 07:47 |
| admin0 | \o | 07:47 |
| *** sxc731 has joined #openstack-ansible | 07:48 | |
| *** vnogin has joined #openstack-ansible | 07:49 | |
| *** sxc731 has quit IRC | 07:53 | |
| *** sxc731 has joined #openstack-ansible | 07:54 | |
| *** armaan has quit IRC | 07:55 | |
| *** armaan has joined #openstack-ansible | 07:55 | |
| *** vnogin has quit IRC | 07:55 | |
| openstackgerrit | Christian Berendt proposed openstack/ansible-hardening master: Make debsums check configurable https://review.openstack.org/515634 | 07:56 |
| evrardjp | Taseer: https://github.com/openstack/openstack-ansible-os_congress | 08:00 |
| evrardjp | :) | 08:00 |
| evrardjp | good morning hwoarang admin0 a0x3e8 | 08:01 |
| openstackgerrit | Christian Berendt proposed openstack/ansible-hardening master: Make check of package checksums configurable https://review.openstack.org/515634 | 08:01 |
| evrardjp | a0x3e8: you should check in your nova scheduler I guess | 08:02 |
| *** vnogin has joined #openstack-ansible | 08:04 | |
| a0x3e8 | evrardjp: I found the cause of the error. it was the neutron config file in compute node. after removing flat:eth12 from /etc/neutron/plugins/ml2/linuxbridge_agent.ini, I can deploy instances. what type should I use for the external network | 08:05 |
| Taseer | evrardjp: That was quick :) | 08:05 |
| evrardjp | Taseer: it's not official yet | 08:05 |
| Taseer | atleast it is progressing | 08:06 |
| evrardjp | a0x3e8: I don't know your network configuration, but I guess it all depends on what nic are available to your neutron agent instances :) | 08:06 |
| evrardjp | a0x3e8: it's all about wiring. Check if you have host overrides in your network configuration, and if they match what you need (it's inside the openstack_user_config) | 08:07 |
| a0x3e8 | evrardjp: used exactly the test environment in openstack-ansible ocata docs. | 08:07 |
| evrardjp | Taseer: yes I followed that to make sure it happens! | 08:08 |
| evrardjp | :D | 08:08 |
| openstackgerrit | Merged openstack/ansible-hardening master: Update to RHEL 7 STIG V1R3 https://review.openstack.org/516660 | 08:08 |
| *** vnogin has quit IRC | 08:08 | |
| evrardjp | a0x3e8: please allow me to read the ocata docs then :) | 08:08 |
| evrardjp | I guess there are things in comments we can easily miss too | 08:09 |
| a0x3e8 | evrardjp: the only change I've made so far is instead of using vlan on single nic, I used 3 physical ports. | 08:09 |
| evrardjp | a0x3e8: may I ask why not deploying pike? | 08:09 |
| evrardjp | a0x3e8: go on? | 08:09 |
| evrardjp | you have three networks then? | 08:10 |
| a0x3e8 | evrardjp: well, I have a really weak internet connection. I mirrored all ocata repositories. it is not so simple for me to get all repos again! :) | 08:11 |
| evrardjp | we should ship you a usb key :p | 08:11 |
| a0x3e8 | evrardjp: I have 3 networks instead of 3 vlans, yes. I use vlans in my switch! | 08:12 |
| evrardjp | ok | 08:12 |
| evrardjp | a0x3e8: and you give only flat networks then? | 08:12 |
| a0x3e8 | evrardjp: the first problem that I had for a long time was git repos fail with a weak connection. | 08:12 |
| evrardjp | I don't understand | 08:12 |
| evrardjp | the idea of vlan network type is that you can sinply give a nic, and traffic will get tagged, which could then be handled by your switching and routing infra | 08:13 |
| a0x3e8 | evrardjp: instead of eth0.10, I use eth0... instead of eth0.20, I use eth1,..... and so on | 08:13 |
| evrardjp | but for what part? | 08:13 |
| evrardjp | sorry if I don't understand your setup, hard to understand without seeing switch configuration/user configuration :D | 08:14 |
| a0x3e8 | evrardjp: let me send it then... | 08:14 |
| evrardjp | :D | 08:15 |
| evrardjp | a0x3e8: did you read what was commented here: https://docs.openstack.org/project-deploy-guide/openstack-ansible/ocata/app-config-test.html#host-network-configuration ? | 08:15 |
| evrardjp | on the compute part | 08:15 |
| openstackgerrit | Markos Chandras (hwoarang) proposed openstack/openstack-ansible-os_tacker master: Initial OSA zuul v3 role jobs https://review.openstack.org/517204 | 08:15 |
| *** shardy has joined #openstack-ansible | 08:16 | |
| openstackgerrit | Jean-Philippe Evrard proposed openstack/openstack-ansible-os_swift master: Add OpenStack-Ansible metadata https://review.openstack.org/517206 | 08:20 |
| openstackgerrit | Markos Chandras (hwoarang) proposed openstack/openstack-ansible-os_tacker stable/pike: Initial OSA zuul v3 role jobs https://review.openstack.org/516306 | 08:20 |
| yifei | Hi guys, I have encountered an ERROR while deploying rabbitmq with OSA Pike. It always failed in the task "Read rabbit cookie". The log is here: http://paste.openstack.org/show/625316/ | 08:20 |
| yifei | However when I replaced "rabbitmq_set_cookie.yml" by the code in OSA Ocata, no ERROR occurred. My operating system is 16.04.3. | 08:20 |
| yifei | Is there anyone can help me with this? Thanks | 08:20 |
| a0x3e8 | evrardjp: yes. my setup is exactly like this. but I used eth0 for eth0.10, eth1 for eth0.20, and eth2 for eth0.30 | 08:20 |
| a0x3e8 | and user_config.yml is exactly the same. | 08:21 |
| evrardjp | yifei: don't need to replace. It's not an issue | 08:24 |
| evrardjp | yifei: the task is marked as an error, but it's immediately rescued by another task | 08:24 |
| evrardjp | yifei: ansible currently has no way to show "rescued" task as green. But this will be changed in ansible 2.5 | 08:25 |
| evrardjp | yifei: if you have a hard failure, that must be after that task | 08:25 |
| yifei | evrardjp: you mean I need to upgrade my ansible to 2.5? | 08:25 |
| evrardjp | yifei: ansible 2.5 doesn't exist yet | 08:25 |
| evrardjp | yifei: stay with what you have, and ignore the error | 08:26 |
| *** pbandark has quit IRC | 08:26 | |
| *** vnogin has joined #openstack-ansible | 08:26 | |
| evrardjp | as I told you, the task should get rescued and the playbook should continue | 08:26 |
| evrardjp | it should have no impact on you | 08:26 |
| evrardjp | afk | 08:26 |
| yifei | evrardjp: but after running setup-infrastructure.yml, it indeed marks it as failed like this: 2017-11-02 05:53:52,086 p=34344 u=root | host3_rabbit_mq_container-b7879c1e : ok=90 changed=50 unreachable=0 failed=1 | 08:27 |
| openstackgerrit | lu.li proposed openstack/openstack-ansible master: Update docs config for linters https://review.openstack.org/517207 | 08:27 |
| yifei | evrardjp: i unerstand that the ansible block module uses rescue | 08:28 |
| yifei | evrardjp: but in my env, ansible marked running setup-infrastructure.yml ad failed and the deployment stopped | 08:30 |
| *** oanson has quit IRC | 08:30 | |
| *** vnogin has quit IRC | 08:30 | |
| yifei | evrardjp: or is there any way to avoid this? Anyway, thanks alot | 08:31 |
| *** oanson has joined #openstack-ansible | 08:31 | |
| *** pbandark has joined #openstack-ansible | 08:34 | |
| *** gkadam has quit IRC | 08:34 | |
| *** gkadam has joined #openstack-ansible | 08:35 | |
| a0x3e8 | evrardjp: https://thepasteb.in/p/98hRqkXL0VyUk | 08:37 |
| a0x3e8 | evrardjp: this one is for compute node: https://thepasteb.in/p/nZhlD72v8JBiY | 08:38 |
| *** ivveh- has joined #openstack-ansible | 08:42 | |
| *** ivveh has quit IRC | 08:44 | |
| andymccr | gunix: what are you trying to do/where are you at currently with the swift deploy? | 08:44 |
| *** oanson has quit IRC | 08:45 | |
| *** oanson has joined #openstack-ansible | 08:46 | |
| hwoarang | any cores around for a quick review of https://review.openstack.org/#/c/516544/ ? :) | 08:49 |
| *** armaan has quit IRC | 08:49 | |
| *** armaan has joined #openstack-ansible | 08:50 | |
| hwoarang | andymccr: i am trying your rpc-ceph repo now but i get "ERROR! Unexpected Exception, this is probably a bug: 'InventoryManager' object has no attribute 'get_group_dict'" | 08:51 |
| hwoarang | ./run_tests.sh alone didn't work immediately for me. i need to run it as root, and also install ansible myself | 08:52 |
| openstackgerrit | Merged openstack/openstack-ansible-os_nova stable/newton: Force the installation of qemu packages https://review.openstack.org/517091 | 08:52 |
| hwoarang | oh hmm | 08:52 |
| hwoarang | should have used sudo run_tests.sh functional i suppose | 08:52 |
| hwoarang | yep that seems to restore sanity :) | 08:54 |
| evrardjp | yifei: do you have another issue in setup-infra ? | 08:56 |
| evrardjp | if not, maybe it didn't fail | 08:56 |
| evrardjp | check with echo $? | 08:57 |
| yifei | evrardjp: totally two issues, both with using block module, i have checked the code | 08:57 |
| evrardjp | yifei: were both rescued? | 08:58 |
| yifei | evrardjp: yes | 08:58 |
| evrardjp | then you're fine | 08:58 |
| yifei | both in rabbitmq | 08:58 |
| yifei | ok | 08:58 |
| evrardjp | go on with the other playbooks | 08:58 |
| yifei | thanks | 08:58 |
| evrardjp | re-try with the same playbook | 08:58 |
| evrardjp | just in case | 08:58 |
| evrardjp | it should show everything alright | 08:58 |
| yifei | thank you | 08:59 |
| a0x3e8 | evrardjp: ok I sent the config files for compute and infra node. | 08:59 |
| a0x3e8 | I tried using br-vxlan as external network which does not work... | 09:00 |
| evrardjp | that is not a good idea a0x3e8 | 09:01 |
| jrosser | a0x3e8: this might be useful https://docs.openstack.org/ocata/networking-guide/deploy-lb-selfservice.html | 09:01 |
| jrosser | it is up to you to set up the external networks as you wish after having done the deploy, it's not done for you | 09:02 |
| evrardjp | jrosser: thanks for the link that's exactly that | 09:03 |
| evrardjp | jrosser: can you take it from there? | 09:03 |
| evrardjp | a0x3e8: check on what's in your infra neutron agent container | 09:04 |
| evrardjp | the nics that are defined there | 09:04 |
| evrardjp | if there is a nic that's not linked to your network infra or not properly configured that's your issue with your flat or your vlan network | 09:05 |
| evrardjp | a0x3e8: I prefer generally flat network for external network, and vxlan or vlan for tenant networks | 09:06 |
| evrardjp | that's the easiest to debug | 09:06 |
| jrosser | this does keep coming up and peoples attention is focussed on the compute nodes, where actually the external networks need to land at the network nodes | 09:06 |
| evrardjp | yes.. | 09:07 |
| evrardjp | What can we do? | 09:07 |
| evrardjp | We already have network nodes, that's already explained | 09:07 |
| evrardjp | maybe we should clarify this | 09:07 |
| evrardjp | in the docs | 09:07 |
| evrardjp | in the examples | 09:07 |
| jrosser | i think the docs might currently make you thinkn the job is done when ASO completes the deploy | 09:07 |
| jrosser | which to an extent it is | 09:07 |
| evrardjp | which part? | 09:07 |
| evrardjp | the network configuration of the tenant/provider networks? | 09:08 |
| openstackgerrit | lu.li proposed openstack/openstack-ansible-tests master: Update docs config for linters https://review.openstack.org/517218 | 09:08 |
| evrardjp | jrosser: I've added something in the operations guide | 09:08 |
| evrardjp | https://docs.openstack.org/openstack-ansible/latest/admin/index.html | 09:08 |
| evrardjp | that, for me, is good enough of a story. | 09:08 |
| evrardjp | but we can discuss on how to improve the docs, I'm all ears. | 09:09 |
| evrardjp | On top of that I have a few ideas of my own | 09:09 |
| evrardjp | jrosser: for now it's kinda bare bones: https://docs.openstack.org/openstack-ansible/latest/admin/openstack-operations/network-service.html | 09:09 |
| evrardjp | maybe we can improve there... | 09:09 |
| *** fxpester has joined #openstack-ansible | 09:10 | |
| evrardjp | and we have that too: https://docs.openstack.org/openstack-ansible/latest/admin/troubleshooting.html#troubleshooting-instance-connectivity-issues | 09:11 |
| *** kiennt26 has joined #openstack-ansible | 09:13 | |
| *** sxc731 has quit IRC | 09:15 | |
| *** sxc731 has joined #openstack-ansible | 09:15 | |
| *** jafeha has joined #openstack-ansible | 09:16 | |
| gunix | andymccr: i was trying to creat an environment with osa-ctrl, osa-compute-1, osa-compute-2, osa-cinder and osa-swift. the swift host didn't deploy properly (was listening only on 20 and 25 and not on any ports related to object storage) | 09:18 |
| jrosser | evrardjp: i think that a lot of people land here https://docs.openstack.org/project-deploy-guide/openstack-ansible/newton/app-config-test.html | 09:18 |
| jrosser | well, ignore "netwton" you know what i mean | 09:19 |
| evrardjp | yeah | 09:19 |
| jrosser | and that is a bit sparse on describing the scope of what that example does | 09:19 |
| evrardjp | jrosser: I have an idea on how to improve that, if you're willing to step up/in (I don't know what the proper term), we can work on some patch together | 09:20 |
| jrosser | sure, no problem | 09:20 |
| jafeha | good morning | 09:21 |
| jrosser | the other info is all out there it just needs pointing to i think | 09:21 |
| *** vnogin has joined #openstack-ansible | 09:22 | |
| *** armaan has quit IRC | 09:22 | |
| *** armaan has joined #openstack-ansible | 09:22 | |
| gunix | is 17 now the stable? | 09:23 |
| gunix | and not 16 ? | 09:23 |
| *** gouthamr has quit IRC | 09:26 | |
| *** vnogin has quit IRC | 09:26 | |
| evrardjp | 16 is pike | 09:26 |
| evrardjp | 17 is master | 09:27 |
| evrardjp | if that's what you're talking about | 09:27 |
| evrardjp | odyssey4me: andymccr is there a reason why some repos where not removed from governance while retired? Thinking about pip_lock_down and openstack-ansible-security | 09:27 |
| evrardjp | for releases maybe? | 09:28 |
| *** electrofelix has joined #openstack-ansible | 09:30 | |
| jiaopengju | hi, whether someone is developing ansible role for karbor? | 09:33 |
| andymccr | gunix: can i see your conf.d settings for swift? and also does the swift run fail? | 09:39 |
| andymccr | evrardjp: because they are still relevant for older releases, pip_lock_down may be done now | 09:39 |
| hwoarang | evrardjp: i think that when you move roles to unmaintained status you will also have to blacklist them from the propose bot | 09:39 |
| evrardjp | andymccr: let me rephrase that, we used to have a problem with the acl in old releases | 09:39 |
| hwoarang | evrardjp: https://github.com/openstack/openstack-ansible-tests/blob/master/gen-projects-list.sh#L41 | 09:39 |
| evrardjp | but that should be fixed. But now maybe we'll have additional issues | 09:40 |
| hwoarang | i dont know if you moved them already just wanted to point that out since i just remembered | 09:40 |
| evrardjp | like the jobs maybe still check what's in governance | 09:40 |
| evrardjp | maybe I should talk with releases | 09:40 |
| evrardjp | hwoarang: done | 09:40 |
| evrardjp | hwoarang: I just need to backport on some branches | 09:41 |
| hwoarang | hmm i can't seem them blacklisted | 09:41 |
| hwoarang | or maybe the patch is not merged yet | 09:41 |
| evrardjp | https://review.openstack.org/#/c/517062/1/gen-projects-list.sh | 09:41 |
| *** sxc731 has quit IRC | 09:42 | |
| hwoarang | oh the github copy is just outdated then | 09:42 |
| hwoarang | ok great :) | 09:42 |
| hwoarang | goot catch :D | 09:42 |
| andymccr | when are you flying out evrardjp | 09:43 |
| evrardjp | saturday | 09:44 |
| evrardjp | why? | 09:44 |
| andymccr | just wondering! dont let the jet lag get you | 09:44 |
| evrardjp | I have yet to prepare the slides :) | 09:44 |
| evrardjp | andymccr: you have a good tip? | 09:44 |
| evrardjp | hwoarang: told me about something IIRC | 09:44 |
| andymccr | haha no you will get jet lag, its like 2 long flights in a row with a layover. i think there is nothing you can do to avoid that :P | 09:44 |
| evrardjp | oh no it was someone else | 09:45 |
| evrardjp | melatonin | 09:45 |
| andymccr | ahh yeah ive not tried that - some ppl swear by it | 09:45 |
| *** sxc731 has joined #openstack-ansible | 09:45 | |
| openstackgerrit | Merged openstack/ansible-hardening master: rhel7stig: sshd.yml: Respect the STIG sshd configuration https://review.openstack.org/516544 | 09:46 |
| *** vnogin has joined #openstack-ansible | 09:46 | |
| evrardjp | andymccr: I don't know it seems risky to play with your body. | 09:47 |
| *** chas has joined #openstack-ansible | 09:49 | |
| *** yifei has quit IRC | 09:51 | |
| *** hw_wutianwei has quit IRC | 09:52 | |
| openstackgerrit | Markos Chandras (hwoarang) proposed openstack/ansible-hardening stable/pike: rhel7stig: sshd.yml: Respect the STIG sshd configuration https://review.openstack.org/517233 | 09:53 |
| openstackgerrit | Markos Chandras (hwoarang) proposed openstack/ansible-hardening stable/ocata: rhel7stig: sshd.yml: Respect the STIG sshd configuration https://review.openstack.org/517234 | 09:54 |
| *** drifterza has joined #openstack-ansible | 09:57 | |
| drifterza | Hi | 09:57 |
| evrardjp | hey drifterza! | 10:02 |
| evrardjp | how are you? | 10:02 |
| drifterza | I'm good thanks man, you / | 10:02 |
| drifterza | ? | 10:02 |
| odyssey4me | evrardjp because they were still part of releases at the time (mitaka was still being released) and the release validation checked governance | 10:04 |
| evrardjp | ok so we check actively governance. Let's keep them that way, we can still remove them later | 10:04 |
| evrardjp | drifterza: good good :) | 10:04 |
| odyssey4me | evrardjp does https://review.openstack.org/#/c/516929/ need to be ported back to pike too? it seems centos is failing and might be for the same reason | 10:05 |
| evrardjp | I can't confirm to 100% | 10:05 |
| evrardjp | but yes maybe. | 10:05 |
| Adri2000 | hello | 10:06 |
| evrardjp | or you can actively recheck. | 10:06 |
| evrardjp | :p | 10:06 |
| evrardjp | let me propose it and we'll see | 10:06 |
| evrardjp | darn it doesn't cleanly apply. | 10:06 |
| evrardjp | Ok I will have a look later today. | 10:06 |
| * evrardjp writes one more thing on his todo list | 10:06 | |
| Adri2000 | has anyone already thought about an elegant way to remove the oldest venvs (in openstack/venvs) in the various hosts and containers? and like keep the N most recent ones | 10:07 |
| evrardjp | does it takes a large room? | 10:07 |
| evrardjp | I am pretty sure we can add some instruction to remove it using ansible ad hoc command that's just a venv anyway :p | 10:08 |
| Adri2000 | evrardjp: each might take a few hundreds MB, so with the number of containers on some hosts, and when your deployment starts getting a bit old, yes it does take a lot of disk space :) | 10:09 |
| odyssey4me | evrardjp I think https://review.openstack.org/517062 might need to be ported back to all current branches, otherwise proposal bot will keep trying with those repositories. | 10:10 |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible stable/newton: Update os_nova role SHA for qemu package fix https://review.openstack.org/517238 | 10:11 |
| Adri2000 | evrardjp: how would such a command look like? I found there is a fact containing the current version, so it might be easy to remove all directories but that one. not sure how to do in order to keep the most recent N versions though :/ is it even useful to do that? (can an older venv be rebuilt? in case I want to "roll back") | 10:11 |
| *** jiaopengju has quit IRC | 10:12 | |
| SmearedBeard | evrardjp on this page you posted (https://docs.openstack.org/project-deploy-guide/openstack-ansible/ocata/app-config-test.html#host-network-configuration), in the openstack_user_config.yml there are two blocks for br-vlan, one flat and with VLAN - but in the networking config, it's just attached to eth0.. are both used? | 10:12 |
| evrardjp | odyssey4me: I know that part. I already have that on my list | 10:14 |
| evrardjp | :) | 10:14 |
| evrardjp | Adri2000: there is no roll back, except if you roll back your db and everything | 10:15 |
| evrardjp | SmearedBeard: you have to select what you need :) | 10:15 |
| evrardjp | Adri2000: it's up to the administrator to know what to do. He probably knows what to delete. | 10:15 |
| evrardjp | then you can remove only that folder | 10:16 |
| SmearedBeard | evrardjp OK so that makes sense.. so if br-vlan is attached to eth0.3 then :type:vlan range 3:3' would be want I need? | 10:16 |
| evrardjp | SmearedBeard: nope I'd say if you just want that, you should use a flat network instead using eth0.3 | 10:17 |
| evrardjp | if eth0 is used for infra, you should probably have a separate nic | 10:17 |
| SmearedBeard | OK so range, is not then VLAN that interfaces used but the range Neutron can used? | 10:17 |
| evrardjp | SmearedBeard: yes. | 10:17 |
| openstackgerrit | Jean-Philippe Evrard proposed openstack/openstack-ansible-tests stable/pike: Remove new retired projects from update bot https://review.openstack.org/517240 | 10:18 |
| openstackgerrit | Jean-Philippe Evrard proposed openstack/openstack-ansible-tests stable/ocata: Remove new retired projects from update bot https://review.openstack.org/517241 | 10:18 |
| openstackgerrit | Jean-Philippe Evrard proposed openstack/openstack-ansible-tests stable/newton: Remove new retired projects from update bot https://review.openstack.org/517242 | 10:18 |
| SmearedBeard | right OK, I think Im getting it.. and br-vlan is used to create networks either private 'per tenant' or external, or shared networks right? | 10:20 |
| openstackgerrit | Jean-Philippe Evrard proposed openstack/openstack-ansible-os_swift stable/pike: Temporary make CentOS non voting https://review.openstack.org/517243 | 10:20 |
| evrardjp | yeah, basically the networks that are defined there are for neutron configuration too. You can use what's defined there for creating your networks afterwards | 10:21 |
| evrardjp | so external networks for example | 10:21 |
| evrardjp | or tenant networks | 10:21 |
| evrardjp | in your case you want just one vlan "content", so you should probably have a flat thing | 10:22 |
| SmearedBeard | 'want' as far as I understand things now, might be what I actually want... | 10:23 |
| SmearedBeard | so br-vxlan is optional, but you could use this to separate external network, and tenant networks? | 10:23 |
| SmearedBeard | which for me, the segregration might be beneficial, rather than the number of potential networks | 10:26 |
| evrardjp | SmearedBeard: use vxlan and/or vlan for tenant networks | 10:28 |
| evrardjp | it's simple and it works | 10:28 |
| evrardjp | use your flat thingy for the external network | 10:28 |
| evrardjp | :) | 10:29 |
| evrardjp | simple | 10:29 |
| evrardjp | odyssey4me: I jinxed the other swift patch by saying "it will pass gates, I feel it". | 10:41 |
| evrardjp | :( | 10:41 |
| openstackgerrit | Jean-Philippe Evrard proposed openstack/openstack-ansible-os_swift stable/pike: Initial OSA zuul v3 role jobs https://review.openstack.org/513813 | 10:42 |
| evrardjp | odyssey4me: could you therefore vote on https://review.openstack.org/#/c/513813/4 ? | 10:42 |
| openstackgerrit | Dimitrios Markou proposed openstack/openstack-ansible-os_neutron master: [WIP] Add OvS-NSH support https://review.openstack.org/517259 | 10:46 |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible master: Zuul: add file extension to playbook path https://review.openstack.org/516098 | 10:47 |
| odyssey4me | evrardjp whitespace | 10:49 |
| *** mardim has joined #openstack-ansible | 10:50 | |
| odyssey4me | evrardjp you fix or me fix ? | 10:50 |
| openstackgerrit | Jean-Philippe Evrard proposed openstack/openstack-ansible-os_swift stable/pike: Initial OSA zuul v3 role jobs https://review.openstack.org/513813 | 10:54 |
| evrardjp | odyssey4me: this gerrit online editor should show whitespaces! | 10:55 |
| evrardjp | done | 10:55 |
| openstackgerrit | Markos Chandras (hwoarang) proposed openstack/openstack-ansible master: Vagrantfile: Remove openSUSE Leap 42.2 https://review.openstack.org/517268 | 11:02 |
| *** David_Orange has joined #openstack-ansible | 11:03 | |
| openstackgerrit | Dimitrios Markou proposed openstack/openstack-ansible master: Add openvswitch_nsh group to inventory https://review.openstack.org/517269 | 11:03 |
| openstackgerrit | Dimitrios Markou proposed openstack/openstack-ansible-os_neutron master: Add OvS-NSH support https://review.openstack.org/517259 | 11:05 |
| openstackgerrit | Dimitrios Markou proposed openstack/openstack-ansible master: Add openvswitch_nsh group to inventory https://review.openstack.org/517269 | 11:12 |
| hwoarang | odyssey4me: evrardjp: have you seen any other opensuse failures? i can't find any. maybe my last patch fixed it .... | 11:14 |
| *** smatzek has joined #openstack-ansible | 11:15 | |
| gunix | andymccr: i am trying to run an aio now and just not use this environment. whatever i failed. :D if you want i can boot the osa-swift-1 vm and show you conf.d. the playbook is successful. | 11:16 |
| mardim | hey guys I have a patch that depends to another patch but zull gives me -1 and this message 'This change depends on a change that failed to merge' do you know what is going wrong? | 11:17 |
| andymccr | gunix: if you can get the conf that'd be good - it should fail if it isnt successfully setup - | 11:17 |
| andymccr | if you re-run the swift playbook does anything happen? | 11:17 |
| gunix | andymccr: yea, it's successful again :) | 11:18 |
| andymccr | gunix: but the services arent running and you cant use the swift cli? | 11:18 |
| andymccr | gunix: can you give me the output of a os-swift-install.yml run? | 11:18 |
| hwoarang | mardim: you probably need to rebase the one that failed to merge | 11:18 |
| odyssey4me | hwoarang I haven't seen any - and the suse builds appear faster than the centos builds now | 11:18 |
| odyssey4me | hwoarang I ported those changes back to pike too | 11:18 |
| mardim | hwoarang, ok let me try | 11:19 |
| odyssey4me | and that's also helped get things moving again | 11:19 |
| openstackgerrit | Dimitrios Markou proposed openstack/openstack-ansible-os_neutron master: Add OvS-NSH support https://review.openstack.org/517259 | 11:19 |
| gunix | andymccr: swift cli is on utility container. first there was a problem with swift-proxy container on ctrl. i destroyed the swift-proxy container and rebuild it, and it started to listen on 8080. but after that it "swift list" because the osa-swift-1 node failedf | 11:20 |
| gunix | because osa-swift-1 was not listening on a port ... if i remember correctly it was 6002 | 11:20 |
| hwoarang | odyssey4me: ok that's good | 11:20 |
| gunix | andymccr: right now i am thinking about killing this entire setup (made of 5 VMs) and just run osa-aio on one node so i have something at home ... | 11:20 |
| andymccr | gunix: it sounds like none of the services are starting, 6002 is the account service - so the storage services should have port 6000, 6001 and 6002 all listening | 11:21 |
| gunix | andymccr: i am also thinking about building ceph for prod so that means i won't use swift any more, i'll just use ceph for glance | 11:21 |
| gunix | andymccr: osa-swift-1 had only 22 and 25 listening :) | 11:21 |
| *** David_Orange has quit IRC | 11:22 | |
| andymccr | gunix: yeah it sounds like something is wrong with either the run or hte inventory/conf setup | 11:22 |
| openstackgerrit | Dimitrios Markou proposed openstack/openstack-ansible master: Add openvswitch_nsh group to inventory https://review.openstack.org/517269 | 11:22 |
| gunix | andymccr: i can show you the openstack_user_config if you want | 11:22 |
| andymccr | gunix: that'd be useful, hopefully we can figure it out :) it should work pretty reliably so im imagining something initial isnt setup | 11:22 |
| mardim | hwoarang, I think I did that but no luck https://review.openstack.org/#/c/517259/ https://review.openstack.org/#/c/517269/ those are the patches | 11:23 |
| odyssey4me | evrardjp I see in that newton integrated patch the functional tests did not run... I've seen this in another repo too - I'll raise it with infra | 11:24 |
| *** David_Orange has joined #openstack-ansible | 11:24 | |
| gunix | andymccr: https://bpaste.net/show/dcc397dc37fe | 11:24 |
| *** armaan has quit IRC | 11:26 | |
| *** armaan has joined #openstack-ansible | 11:27 | |
| *** huxinhui has quit IRC | 11:31 | |
| andymccr | gunix: ok a few things - firstly are the devices format/mounted at /srv/node/vdb and /srv/node/vdc (https://docs.openstack.org/openstack-ansible-os_swift/latest/configure-swift-devices.html) | 11:32 |
| andymccr | gunix: also a couple of those settings arent required - e.g. storage_ip on the swift_host | 11:33 |
| andymccr | and the limit_container_types var too | 11:33 |
| andymccr | i would remove that | 11:33 |
| andymccr | gunix: the next thing i'd do is look at the logs on the hosts, if the run is succeeding but the services arent running they must have tried to start and failed | 11:34 |
| andymccr | you could check systemctl debug logs, and try to manually start the swift-object-server for example, or swift-account-server (which is the one that should've been listening on port 6002) | 11:34 |
| openstackgerrit | Merged openstack/openstack-ansible-os_swift master: Add OpenStack-Ansible metadata https://review.openstack.org/517206 | 11:35 |
| *** David_Orange has quit IRC | 11:37 | |
| gunix | andymccr: the devices are mounting as they should be. if the storage_ip is critical, that must be the reason of the fail. it should be 172.29.244.41 | 11:37 |
| *** armaan has quit IRC | 11:39 | |
| *** armaan has joined #openstack-ansible | 11:39 | |
| andymccr | gunix: so as long as that ip is on br-storage you don't need to specify storage_ip, it will figure it out for you | 11:40 |
| *** PaulusNL has quit IRC | 11:44 | |
| gunix | andymccr: yea, it's on br-storage :) | 11:53 |
| spotz | evrardjp: Might want to send that email to the ops list to get more ops folks at least to the forum session. Anyaways last ride before I head to the airport this afternoon. Back in a bit | 11:53 |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible-os_tacker stable/pike: Initial OSA zuul v3 role jobs https://review.openstack.org/517284 | 11:55 |
| mhayden | morning | 11:56 |
| andymccr | well hello there mhayden | 11:58 |
| mhayden | ohai | 11:58 |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible-os_tacker stable/pike: Initial OSA zuul v3 role jobs https://review.openstack.org/516306 | 11:59 |
| *** shardy is now known as shardy_lunch | 12:01 | |
| *** stuartgr has joined #openstack-ansible | 12:02 | |
| *** armaan has quit IRC | 12:05 | |
| gunix | andymccr: if my aio runs without problem, i will just use that for tests. i got it 20GB ram, 6 cores and 115GB storage with 5gb swap | 12:09 |
| andymccr | gunix: sounds good. if you do want to get it working on your other deploy id recommend looking at why the services arent starting as a first step :) | 12:11 |
| gunix | andymccr: yea, problem was i am kind of a noob on swift and i didn't deploy it ever without OSA yet ... the other services i can troubleshoot, but not this one. | 12:12 |
| evrardjp[m] | it's already sent to ops... | 12:12 |
| openstackgerrit | Dimitrios Markou proposed openstack/openstack-ansible master: Add openvswitch_nsh group to inventory https://review.openstack.org/517269 | 12:12 |
| evrardjp[m] | or it should be at least | 12:13 |
| gunix | andymccr: odyssey4me: cloudnull: you guys really did most of the development for osa? you have on github like 13k lines of code together | 12:13 |
| andymccr | gunix: that's ok - it's not that complicated, it does have a lot of services though - but im imagining you will see why it failed in the logs :) | 12:13 |
| odyssey4me | gunix ;) | 12:13 |
| gunix | andymccr: yea. how do you find ceph compared to swift, for object storage? for glance i mean, and maybe giving tokens to specific web apps for integration with app layer of the customer? | 12:13 |
| andymccr | gunix: ceph is pretty cool because you can use block storage too for like cinder/instance integration | 12:14 |
| openstackgerrit | Dimitrios Markou proposed openstack/openstack-ansible-os_neutron master: Add OvS-NSH support https://review.openstack.org/517259 | 12:14 |
| andymccr | it works ok with glance so i dont think thats an issue | 12:14 |
| gunix | andymccr: my thinking exactly. if i use ceph, i get HA block&object storage ! | 12:17 |
| gunix | andymccr: only problem is, i am a noob at ceph too. i should start my testing env for ceph | 12:17 |
| andymccr | gunix: gotta start somewhere :) | 12:17 |
| gunix | yea | 12:18 |
| evrardjp[m] | wow thats a different set of skills | 12:18 |
| gunix | evrardjp[m]: what? | 12:18 |
| evrardjp[m] | scaling storage systems i mean | 12:18 |
| gunix | evrardjp[m]: that's an interesting remark. do you have like one department for ceph and another one for core services like nova/neutron/glance/keystone? and somehow people know how to merge the two together? | 12:19 |
| xdfil | Hello, how can I get osa to /not/ configure the IP on my vxlan bridge? | 12:19 |
| evrardjp[m] | it's good to have in one's skillset | 12:19 |
| xdfil | oh wait ip_from_q or whatever | 12:19 |
| gunix | xdfil: each node requires to have an IP in the VXLAN network for the service to work. it's part of how vxlan works. it encapsulates layer 2 frames within layer 3 packages | 12:20 |
| xdfil | The problem I have is I am trying to do OVS | 12:20 |
| xdfil | and I need to put the IP on the OVS bridge internal interface | 12:21 |
| xdfil | and remove it from the veth | 12:21 |
| xdfil | at least, thats what I think I need to be doing right now.... | 12:21 |
| evrardjp[m] | gunix in my previous company storage was indeed handled by different ppl than compute | 12:21 |
| gunix | i am not sure OVS is supported by osa | 12:21 |
| xdfil | gunix: I'm working on it | 12:21 |
| Bico_Fino | xdfil https://docs.openstack.org/openstack-ansible-os_neutron/latest/app-openvswitch.html | 12:21 |
| xdfil | lol jk so many others have put effort into OVS before me | 12:22 |
| gunix | evrardjp[m]: all our storage is done by emc... and ... it's bad ... | 12:22 |
| *** armaan has joined #openstack-ansible | 12:22 | |
| openstackgerrit | Major Hayden proposed openstack/openstack-ansible-os_almanach stable/pike: Updated from OpenStack Ansible Tests https://review.openstack.org/517083 | 12:22 |
| gunix | xdfil: did you go through this: https://docs.openstack.org/openstack-ansible-os_neutron/latest/app-openvswitch.html | 12:22 |
| openstackgerrit | Major Hayden proposed openstack/openstack-ansible-os_almanach master: Updated from OpenStack Ansible Tests https://review.openstack.org/514692 | 12:22 |
| gunix | ? | 12:22 |
| xdfil | numerous times | 12:22 |
| xdfil | I intend to fix those docs soon | 12:23 |
| xdfil | I found all the logic bombs | 12:23 |
| Bico_Fino | What issues are you facing xdfil ? | 12:23 |
| Bico_Fino | I'm trying to deploy OSA with ovs right now. :D | 12:23 |
| xdfil | I just need to remove the IP... I already found my answer hold on :) | 12:23 |
| jafeha | xdfil: i hope to finalize my first deployment with ovs today, but i expect similar issues. could you share your fixes / solutions? | 12:25 |
| Bico_Fino | I got some issues on setup-hosts.yml with neutron_agent containers that don't start. | 12:25 |
| Bico_Fino | would be cool xdfil to share your fixes. | 12:25 |
| xdfil | jafeha: yes for sure, as soon as I iron out the last details | 12:26 |
| jafeha | xdfil: cool, thank you | 12:26 |
| gunix | xdfil: why is ovs important for your build? which advantages do you have by using ovs? | 12:26 |
| xdfil | Bico_Fino: The containers wont start at all? | 12:26 |
| openstackgerrit | Merged openstack/openstack-ansible-tests master: Update docs config for linters https://review.openstack.org/517218 | 12:26 |
| evrardjp[m] | jafeha xdfil, maybe worth sharing on etherpad, to list the work to be done ? | 12:26 |
| Bico_Fino | xdfil nope. Got some veth lxc issues | 12:27 |
| jafeha | evrardjp[m]: definitely | 12:27 |
| xdfil | Bico_Fino: sometimes that container restart / start tasks take a long long time are you sure its failing? | 12:27 |
| xdfil | evrardjp[m]: yeah you read my mind | 12:27 |
| Bico_Fino | xdfil I think I fixed it. Container "infra1_neutron_agents_container-9fbda74e" is in "RUNNING" state | 12:27 |
| jafeha | we're also switching to ceph-ansible deployed ceph, so we're squashing numerous bugs in our deployment ;) | 12:27 |
| evrardjp[m] | great. | 12:28 |
| xdfil | evrardjp[m] Bico_Fino jafeha : I just want to make sure I am doing it right before I go and derail others | 12:28 |
| evrardjp[m] | jafeha: that's what we are using in osa... great to hear! | 12:28 |
| Bico_Fino | xdfil We can share the configs | 12:28 |
| xdfil | https://etherpad.openstack.org/p/OSA-OVS-throwdown | 12:29 |
| xdfil | I'll post my fixes there once I verify them | 12:29 |
| Bico_Fino | xdfil nope. Failed. | 12:30 |
| jafeha | xdfil: thank you :) | 12:30 |
| *** gkadam has quit IRC | 12:30 | |
| xdfil | Bico_Fino: that's got to be some kind of missing pre-req on the LXC host | 12:30 |
| Bico_Fino | xdfil posted the logs there. | 12:31 |
| xdfil | Bico_Fino: I've been using OSA for a few months and I've never had that exact problem | 12:31 |
| Bico_Fino | xdfil the strange is that only with neutron_agent containers. | 12:32 |
| openstackgerrit | Merged openstack/openstack-ansible master: Update docs config for linters https://review.openstack.org/517207 | 12:32 |
| openstackgerrit | Merged openstack/openstack-ansible stable/newton: Update os_nova role SHA for qemu package fix https://review.openstack.org/517238 | 12:32 |
| *** udesale has quit IRC | 12:33 | |
| *** Miouge has joined #openstack-ansible | 12:33 | |
| xdfil | Bico_Fino: hmm then I would look at your openstack_user_config.yml networks | 12:33 |
| openstackgerrit | OpenStack Proposal Bot proposed openstack/openstack-ansible-os_almanach master: Updated from OpenStack Ansible Tests https://review.openstack.org/514692 | 12:33 |
| xdfil | maybe you have some issue with_groups | 12:33 |
| xdfil | did you mess with the env.d files at all? | 12:34 |
| *** smatzek has quit IRC | 12:34 | |
| Bico_Fino | xdfil nope. | 12:34 |
| Bico_Fino | xdfil posted there the networks that I have configured | 12:35 |
| openstackgerrit | OpenStack Proposal Bot proposed openstack/openstack-ansible-os_monasca master: Updated from OpenStack Ansible Tests https://review.openstack.org/514695 | 12:35 |
| openstackgerrit | OpenStack Proposal Bot proposed openstack/openstack-ansible-os_monasca-agent master: Updated from OpenStack Ansible Tests https://review.openstack.org/514696 | 12:36 |
| openstackgerrit | Merged openstack/openstack-ansible-lxc_hosts master: Use handlers to restart services and move dnsmasq to a unit file https://review.openstack.org/516822 | 12:36 |
| openstackgerrit | OpenStack Proposal Bot proposed openstack/openstack-ansible-os_monasca-ui master: Updated from OpenStack Ansible Tests https://review.openstack.org/514697 | 12:36 |
| *** dave-mccowan has joined #openstack-ansible | 12:36 | |
| Bico_Fino | grabbing coffee. | 12:36 |
| xdfil | Bico_Fino: brctl show on your container host? | 12:38 |
| *** thorst has joined #openstack-ansible | 12:41 | |
| *** a0x3e8 has quit IRC | 12:42 | |
| xdfil | lxc-ls -f | 12:43 |
| xdfil | ^^ that too | 12:43 |
| xdfil | I think maybe you have zombie containers... idk | 12:43 |
| *** wlmbasson has joined #openstack-ansible | 12:45 | |
| wlmbasson | Does anyone know how to use an external DHCP server with provider networks? I would like to integrate my AIO build with our existing DNS/DHCP | 12:46 |
| jafeha | evrardjp[m]: yeah, we're using the osa ceph-ansible integration now. we had built it on our own in earlier deployments because we were not sure how to do it exactly ;) | 12:46 |
| Bico_Fino | xdfil posted brctl show too | 12:47 |
| xdfil | Bico_Fino: the br-vxlan bridge doesn't exist | 12:48 |
| xdfil | and that relates to the agents containers | 12:49 |
| xdfil | they are the only ones that get an interface on that bridge | 12:49 |
| *** markvoelker has joined #openstack-ansible | 12:49 | |
| xdfil | try: lxc-system-manage system-rebuild | 12:49 |
| xdfil | i think that's the command :) | 12:49 |
| xdfil | see if it makes a br-vxlan | 12:50 |
| xdfil | or not wait | 12:50 |
| *** markvoelker_ has joined #openstack-ansible | 12:50 | |
| xdfil | what am i saying | 12:50 |
| Bico_Fino | xdfil inside /etc/network/interfaces br-vxlan is misconfigured. :( | 12:50 |
| xdfil | the OS makes that bridge | 12:50 |
| Bico_Fino | xdfil yeah | 12:50 |
| Bico_Fino | let me fix that | 12:50 |
| *** jiaopengju has joined #openstack-ansible | 12:53 | |
| Bico_Fino | Do I need an IP on br-vxlan xdfil ? | 12:53 |
| *** woodard has joined #openstack-ansible | 12:53 | |
| *** markvoelker has quit IRC | 12:53 | |
| *** smatzek has joined #openstack-ansible | 12:54 | |
| Miouge | I can’t find documentation about how to setup OSA with Ceph RadosGW (instead of Swift). Anyone who can point me in the right direction? | 12:56 |
| *** weezS has joined #openstack-ansible | 12:56 | |
| odyssey4me | hmm, evrardjp cloudnull it would appears that the recent merges to lxc_hosts have broken the integrated build: http://logs.openstack.org/98/516098/2/check/openstack-ansible-deploy-aio-ubuntu-xenial/b1fb879/job-output.txt.gz#_2017-11-02_11_03_04_303504 | 12:58 |
| openstackgerrit | Merged openstack/openstack-ansible-os_swift stable/pike: Initial OSA zuul v3 role jobs https://review.openstack.org/513813 | 12:58 |
| *** smatzek has quit IRC | 12:59 | |
| evrardjp | odyssey4me: yes I have seen | 12:59 |
| evrardjp | I still wonder how, because sha bump didn't happen | 12:59 |
| evrardjp | oh you mean master? | 13:00 |
| evrardjp | yeah | 13:00 |
| evrardjp | I need to fix that real quick | 13:00 |
| evrardjp | it's something done yesterday i think | 13:00 |
| odyssey4me | well, cloudnull should be around shortly and should be able to fix it - unless you specifically need it fixed right now | 13:01 |
| evrardjp | finishing my dessert first | 13:01 |
| evrardjp | I'm still at food stage right now. You know me. | 13:01 |
| odyssey4me | haha | 13:01 |
| odyssey4me | second breakfast | 13:01 |
| evrardjp | I had only one. So it's 2 lunches instead | 13:02 |
| evrardjp | :) | 13:02 |
| *** smatzek has joined #openstack-ansible | 13:03 | |
| evrardjp | I have no McCoy crisps anymore, that makes me really sad too. | 13:03 |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible-os_tacker stable/pike: Initial OSA zuul v3 role jobs https://review.openstack.org/516306 | 13:06 |
| *** smatzek has quit IRC | 13:07 | |
| odyssey4me | evrardjp actually I think that issue was resolved with https://review.openstack.org/516822 | 13:08 |
| xdfil | Well, removing the IP from the neutron_agents br-vxlan seems to break other peices | 13:13 |
| xdfil | The OVS "Manager" config seems to be dependant on the br-vxlan IP | 13:14 |
| xdfil | so if you remove it that breaks | 13:14 |
| xdfil | We just need a little bit of flexibility with br-vxlan and I'm pretty sure it will work | 13:15 |
| Bico_Fino | xdfil running setup-hosts now | 13:16 |
| evrardjp | odyssey4me: let me do a quick recheck then | 13:18 |
| evrardjp | but I have seen that in the past | 13:18 |
| evrardjp | in zuul only | 13:18 |
| evrardjp | when doing some include when with a {{ }} | 13:19 |
| Bico_Fino | xdfil bingo. worked! | 13:19 |
| evrardjp | could someone take care of the bug triage next week? | 13:22 |
| xdfil | (crickets) | 13:22 |
| *** smatzek has joined #openstack-ansible | 13:23 | |
| *** smatzek has quit IRC | 13:24 | |
| *** shardy_lunch is now known as shardy | 13:25 | |
| *** sxc731 has quit IRC | 13:28 | |
| openstackgerrit | Major Hayden proposed openstack/openstack-ansible-os_almanach master: Updated from OpenStack Ansible Tests https://review.openstack.org/514692 | 13:30 |
| *** thorst has quit IRC | 13:30 | |
| mhayden | evrardjp: enjoy the trip to sydney ;) | 13:32 |
| mhayden | you too, cloudnull | 13:32 |
| openstackgerrit | Merged openstack/openstack-ansible-tests stable/ocata: Remove new retired projects from update bot https://review.openstack.org/517241 | 13:33 |
| openstackgerrit | Merged openstack/openstack-ansible-tests stable/newton: Remove new retired projects from update bot https://review.openstack.org/517242 | 13:33 |
| openstackgerrit | Merged openstack/openstack-ansible-tests stable/pike: Remove new retired projects from update bot https://review.openstack.org/517240 | 13:33 |
| *** jwitko has joined #openstack-ansible | 13:33 | |
| *** lbragstad has joined #openstack-ansible | 13:34 | |
| odyssey4me | oh mhayden don't bother rebasing those proposal bot changes on top of the zuul job changes... proposal bot will reset them anyway | 13:35 |
| mhayden | odyssey4me: ah, i was starting to notice it un-doing my work :| | 13:35 |
| odyssey4me | we either need to drop our patch and move the changes into the proposal bot changes, or we need to focus on getting our patches merged | 13:35 |
| odyssey4me | evrardjp thoughts - should we just move our job changes into that patch? | 13:36 |
| mhayden | i keep scratching my head on those that don't seem to run functional tests | 13:36 |
| mhayden | the project.yaml looks reasonable | 13:36 |
| odyssey4me | mhayden yeah, I've raised it in infra - it's been spotted in the wild, but not consistently | 13:37 |
| mhayden | thanks for that | 13:37 |
| odyssey4me | we have one consistent one though, so hopefully we can help debug it | 13:37 |
| odyssey4me | I'm going to convert my & evrardjp's patches for jobs into the proposal bot proposals where they make sense to cut some of the noise | 13:38 |
| *** flaviosr has quit IRC | 13:38 | |
| openstackgerrit | OpenStack Proposal Bot proposed openstack/openstack-ansible-os_almanach stable/pike: Updated from OpenStack Ansible Tests https://review.openstack.org/517083 | 13:39 |
| openstackgerrit | OpenStack Proposal Bot proposed openstack/openstack-ansible-os_aodh stable/pike: Updated from OpenStack Ansible Tests https://review.openstack.org/516285 | 13:39 |
| openstackgerrit | OpenStack Proposal Bot proposed openstack/openstack-ansible-os_monasca stable/pike: Updated from OpenStack Ansible Tests https://review.openstack.org/516299 | 13:40 |
| openstackgerrit | OpenStack Proposal Bot proposed openstack/openstack-ansible-os_nova stable/pike: Updated from OpenStack Ansible Tests https://review.openstack.org/516301 | 13:40 |
| openstackgerrit | OpenStack Proposal Bot proposed openstack/openstack-ansible-os_sahara stable/pike: Updated from OpenStack Ansible Tests https://review.openstack.org/516304 | 13:40 |
| *** sxc731 has joined #openstack-ansible | 13:40 | |
| openstackgerrit | OpenStack Proposal Bot proposed openstack/openstack-ansible-os_swift stable/pike: Updated from OpenStack Ansible Tests https://review.openstack.org/516305 | 13:40 |
| openstackgerrit | OpenStack Proposal Bot proposed openstack/openstack-ansible-os_tacker stable/pike: Updated from OpenStack Ansible Tests https://review.openstack.org/517310 | 13:40 |
| openstackgerrit | Major Hayden proposed openstack/ansible-hardening master: [Docs] Specify STIG version in branches https://review.openstack.org/517311 | 13:41 |
| evrardjp | mhayden: thanks | 13:41 |
| *** markvoel_ has joined #openstack-ansible | 13:41 | |
| openstackgerrit | OpenStack Proposal Bot proposed openstack/openstack-ansible-os_aodh stable/ocata: Updated from OpenStack Ansible Tests https://review.openstack.org/514748 | 13:41 |
| evrardjp | odyssey4me: experience tell me that we should ignore proposal bot things, get our stuff merged, and be done with it | 13:41 |
| *** flaviosr has joined #openstack-ansible | 13:42 | |
| *** markvoelker_ has quit IRC | 13:42 | |
| evrardjp | but fine for me too to update the proposal bot things | 13:42 |
| evrardjp | as long as we know what we are doing | 13:42 |
| odyssey4me | evrardjp yeah, let me take a closer look | 13:42 |
| openstackgerrit | OpenStack Proposal Bot proposed openstack/openstack-ansible-os_ceilometer stable/newton: Updated from OpenStack Ansible Tests https://review.openstack.org/516259 | 13:43 |
| openstackgerrit | OpenStack Proposal Bot proposed openstack/openstack-ansible-os_gnocchi stable/newton: Updated from OpenStack Ansible Tests https://review.openstack.org/516260 | 13:43 |
| openstackgerrit | OpenStack Proposal Bot proposed openstack/openstack-ansible-os_sahara stable/newton: Updated from OpenStack Ansible Tests https://review.openstack.org/516262 | 13:43 |
| odyssey4me | any voters for https://review.openstack.org/516306 to get that one out of the way? | 13:46 |
| *** markvoel_ has quit IRC | 13:47 | |
| odyssey4me | evrardjp you're right - the proposal bot uses the topic to find its previous patch, so if we change it then it'll just make another patch | 13:47 |
| odyssey4me | so better to just use our patches and get them merged | 13:48 |
| openstackgerrit | Jean-Philippe Evrard proposed openstack/openstack-ansible-plugins stable/newton: Make possible to use ansible 2.4 https://review.openstack.org/517314 | 13:54 |
| evrardjp | shitballz. | 13:54 |
| odyssey4me | hahaha | 13:54 |
| odyssey4me | that's a stretch | 13:54 |
| openstackgerrit | Jean-Philippe Evrard proposed openstack/openstack-ansible-plugins master: Make possible to use ansible 2.4 https://review.openstack.org/517315 | 13:54 |
| evrardjp | darn. It's wrong. | 13:55 |
| evrardjp | that is why you should always commit from your environment, not from a different host! | 13:55 |
| odyssey4me | or just verify your checkout before you start working on things ;) | 13:56 |
| evrardjp | well I have no prompt like I do on my dev machine. Darn. | 13:56 |
| evrardjp | I am very sorry | 13:57 |
| openstackgerrit | Jean-Philippe Evrard proposed openstack/openstack-ansible-plugins master: Make possible to use ansible 2.4 https://review.openstack.org/517315 | 13:58 |
| *** smatzek has joined #openstack-ansible | 13:59 | |
| openstackgerrit | Jean-Philippe Evrard proposed openstack/openstack-ansible-plugins master: Make possible to use ansible 2.4 https://review.openstack.org/517315 | 13:59 |
| evrardjp | jmccrory: ^ this way we can continue with 2.4 without caring about the vars plugins | 13:59 |
| cloudnull | mornings | 14:00 |
| *** smatzek_ has joined #openstack-ansible | 14:02 | |
| *** smatzek has quit IRC | 14:03 | |
| *** sudhamsd has joined #openstack-ansible | 14:05 | |
| openstackgerrit | Dimitrios Markou proposed openstack/openstack-ansible-os_neutron master: Add OvS-NSH support https://review.openstack.org/517259 | 14:06 |
| *** esberglu has joined #openstack-ansible | 14:07 | |
| *** galstrom_zzz is now known as galstrom | 14:08 | |
| *** sudhamsd has quit IRC | 14:10 | |
| *** mardim has quit IRC | 14:10 | |
| odyssey4me | cloudnull the functional test in https://review.openstack.org/513999 is failing - any chance you could take a look into it and update the patch to resolve it? | 14:10 |
| gunix | guys, regarding osa aio: how is the br-flat network connected to the internet? i found nothing in iptables and nothing on the host ... i found it has two IPs, 248.1 and 248.100, and there is not in the iface config that .1 is for connectivity, but i swear i can't see at linux level how this is connected to ens3 which is the physical device | 14:11 |
| cloudnull | sure thing | 14:11 |
| cloudnull | ^ odyssey4me | 14:11 |
| odyssey4me | thanks cloudnull - apologies for handing you ye olde issue | 14:11 |
| cloudnull | gunix: on the AIO none of the bridges are connected to the physical NICs. its all IPTables that makes it happen . | 14:12 |
| cloudnull | all of the bridges should have a bridge-port of none | 14:12 |
| evrardjp | cloudnull: I know you were interested by the EOL of N, nova's EOL not gonna happen before around the 20 november. | 14:13 |
| *** thorst has joined #openstack-ansible | 14:13 | |
| evrardjp | so this way we know. | 14:13 |
| cloudnull | cool | 14:13 |
| gunix | cloudnull: they do have a bridge port of none. so iptables connects the networks? because i did iptables -t nat -L and didn't find it | 14:13 |
| gunix | cloudnull: where there is a masquarade ALL, is that it? i'll paste config, wait | 14:14 |
| *** thorst_ has joined #openstack-ansible | 14:14 | |
| gunix | https://bpaste.net/show/4b120b059de1 | 14:14 |
| cloudnull | Yup | 14:14 |
| cloudnull | and then we checksum all of the traffic too | 14:14 |
| cloudnull | that way services feel like the traffic is passing through a real nic | 14:15 |
| *** agrebennikov has joined #openstack-ansible | 14:16 | |
| cloudnull | for the AIO we cant control any of the deployment locations and dont want to break networking by making assumptions that the bridges have to be connected to the gateway interface. | 14:16 |
| cloudnull | so we IPTables our way to success. | 14:16 |
| openstackgerrit | Kevin Carter (cloudnull) proposed openstack/openstack-ansible-lxc_container_create stable/pike: Fix container mac generation and remove extra tasks https://review.openstack.org/516344 | 14:17 |
| openstackgerrit | Kevin Carter (cloudnull) proposed openstack/openstack-ansible-os_swift stable/pike: Revert "Don't run sync in upgrades" https://review.openstack.org/516333 | 14:17 |
| *** vnogin has quit IRC | 14:17 | |
| *** thorst has quit IRC | 14:18 | |
| *** vnogin has joined #openstack-ansible | 14:18 | |
| *** vnogin has quit IRC | 14:19 | |
| *** vnogin has joined #openstack-ansible | 14:20 | |
| cloudnull | odyssey4me: interesting that its only failing on trusty | 14:20 |
| odyssey4me | cloudnull yeah, it's probably a service not starting properly or something - I haven't had much time to look into it | 14:21 |
| cloudnull | I've got a test vm spun up now, hopefully more soon. | 14:21 |
| *** drifterza has quit IRC | 14:22 | |
| odyssey4me | ta | 14:22 |
| *** armaan has quit IRC | 14:23 | |
| *** mardim has joined #openstack-ansible | 14:24 | |
| *** vnogin has quit IRC | 14:25 | |
| evrardjp | cloudnull: odyssey4me thanks on that! | 14:26 |
| *** andressanchez has joined #openstack-ansible | 14:26 | |
| *** vnogin has joined #openstack-ansible | 14:29 | |
| *** sxc731 has quit IRC | 14:31 | |
| cloudnull | qq- I'd like to know if folks are open to backporting the lxc_{hosts,container_create} storage bits to pike? | 14:42 |
| cloudnull | logan-: evrardjp: mgariepy: ^ | 14:42 |
| cloudnull | idea being that we would enable the option to deploy an infra into zfs, machinectl, btrfs, etc within our stable release. | 14:43 |
| *** David_Orange has joined #openstack-ansible | 14:43 | |
| *** jiaopengju has quit IRC | 14:46 | |
| *** armaan has joined #openstack-ansible | 14:48 | |
| mgariepy | yes i would like very much to have zfs support on pike, | 14:48 |
| evrardjp | I'd like to abstain from this vote. I have not enough view of the risk/impact. | 14:50 |
| evrardjp | and on top of that it was failing yesterday on master, so let's really think about it... :) | 14:51 |
| evrardjp | I'd prefer waiting for next release on all branches too. Because we still have to recover. | 14:52 |
| evrardjp | but after that, I'd say it's a community decision, and you have already 2 members of the community okay for it :) | 14:53 |
| openstackgerrit | Kevin Carter (cloudnull) proposed openstack/openstack-ansible-lxc_hosts stable/pike: Combined backport for LXC container create storage options https://review.openstack.org/517338 | 14:53 |
| cloudnull | ^ folks can vote :D | 14:53 |
| *** qcrivera has joined #openstack-ansible | 14:54 | |
| evrardjp | that works | 14:58 |
| openstackgerrit | Jimmy McCrory proposed openstack/openstack-ansible master: [TEST] Update Ansible to 2.4.2.0-0.1.beta1 https://review.openstack.org/501814 | 14:59 |
| openstackgerrit | Jimmy McCrory proposed openstack/openstack-ansible master: [TEST] Update Ansible to 2.4.2.0-0.1.beta1 https://review.openstack.org/501814 | 15:00 |
| openstackgerrit | Kevin Carter (cloudnull) proposed openstack/openstack-ansible-lxc_hosts stable/pike: Combined backport to fix issues and enhance efficiency https://review.openstack.org/517341 | 15:01 |
| Neptu_ | evrardjp: hej, Ive been testing the gate deployment and i get funny certificate messages from monasca | 15:01 |
| Neptu_ | https://pastebin.com/KtFBmV93 | 15:02 |
| Neptu_ | im not an ace on pki but i guess is because the certificates are self signed | 15:03 |
| Neptu_ | you do not have a certbot or something to generate certificates | 15:04 |
| evrardjp | Neptu_: you can do whatever you want with your certificates. We'll use them. | 15:05 |
| Neptu_ | evrardjp: im not pki expert but maybe is something to be added to the project... seems monasca python dies when the sertificates are insecure... and will be nice to have actually EFF certs so adding suport for cerbot is a + | 15:07 |
| Neptu_ | i will look into that if i have time | 15:07 |
| evrardjp | sure thing | 15:08 |
| evrardjp | you can use the external_lb_vip_address as a variable for that | 15:08 |
| Neptu_ | https://www.eff.org/ | 15:08 |
| evrardjp | that should hold the external domain name, or at least openstack's fqdn | 15:08 |
| Neptu_ | yep | 15:08 |
| Neptu_ | we will see I just installed as the manual said a monasca system on a k8s cluster next to the gate openstack | 15:09 |
| Neptu_ | trying to connect the beast | 15:09 |
| openstackgerrit | Kevin Carter (cloudnull) proposed openstack/openstack-ansible-lxc_container_create stable/pike: Combined backport to support LXC container storage options https://review.openstack.org/517345 | 15:09 |
| *** qcrivera has quit IRC | 15:10 | |
| openstackgerrit | Merged openstack/openstack-ansible-os_nova stable/pike: Initial OSA zuul v3 role jobs https://review.openstack.org/516596 | 15:10 |
| Neptu_ | been reading the developer document today ;) | 15:10 |
| odyssey4me | cloudnull can you please also port back the tests to make sure that they're exercised on pike | 15:11 |
| cloudnull | ++ | 15:12 |
| Neptu_ | evrardjp: just for the sake of this tests is it possible to dissable keystone on https and run it as http only? | 15:14 |
| *** Brew has joined #openstack-ansible | 15:16 | |
| *** chyka has joined #openstack-ansible | 15:17 | |
| mhayden | xgerman_: have you reviewed this one yet ? https://review.openstack.org/#/c/515698/ | 15:22 |
| *** gkadam has joined #openstack-ansible | 15:23 | |
| openstackgerrit | Major Hayden proposed openstack/openstack-ansible-os_octavia stable/pike: Add zuul v3 support/jobs https://review.openstack.org/517348 | 15:27 |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible-os_almanach master: Initial OSA zuul v3 role jobs https://review.openstack.org/512729 | 15:30 |
| *** hachi__ has quit IRC | 15:30 | |
| openstackgerrit | Major Hayden proposed openstack/openstack-ansible-os_octavia stable/pike: Enable V2 Octavia API (Experimental) https://review.openstack.org/515698 | 15:32 |
| openstackgerrit | Major Hayden proposed openstack/openstack-ansible-os_octavia stable/pike: Updated from OpenStack Ansible Tests https://review.openstack.org/516302 | 15:32 |
| openstackgerrit | Major Hayden proposed openstack/openstack-ansible-os_octavia master: Fix for keystone_authtoken key names in octavia.conf https://review.openstack.org/457331 | 15:32 |
| openstackgerrit | Major Hayden proposed openstack/openstack-ansible-os_octavia master: add segmentation id option to octavia mgmt network creation task https://review.openstack.org/515035 | 15:33 |
| openstackgerrit | Major Hayden proposed openstack/openstack-ansible-os_octavia master: Add allocation pool for octavia mgmt subnet https://review.openstack.org/515378 | 15:33 |
| mhayden | trying to see if we can get some octavia tests to vote | 15:33 |
| *** KMT has joined #openstack-ansible | 15:37 | |
| xgerman_ | mhayden I guess we only have zuul in the master branch | 15:38 |
| *** KMT has quit IRC | 15:38 | |
| odyssey4me | xgerman_ nope, you should already have pike jobs too https://review.openstack.org/516598 | 15:40 |
| openstackgerrit | Merged openstack/openstack-ansible-plugins master: Make possible to use ansible 2.4 https://review.openstack.org/517315 | 15:40 |
| mhayden | xgerman_: i rebased those patches in the hopes of getting some zuul voting going | 15:41 |
| xgerman_ | ok, so let’s see if that happens if not I will try to figure out what’s going on | 15:41 |
| armaan | folks, anyone running Magnum with Pike here? | 15:41 |
| odyssey4me | mhayden you should never need to rebase unless you have two patches in review that depend on each other | 15:42 |
| odyssey4me | zuul rebases before it tests | 15:42 |
| odyssey4me | evrardjp about the comment in https://review.openstack.org/517338 - this won't affect the current release unless there's a sha bump in the integrated repo | 15:46 |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible-os_gnocchi stable/newton: Initial OSA zuul v3 role jobs https://review.openstack.org/514005 | 15:46 |
| evrardjp | odyssey4me: yes indeed | 15:47 |
| evrardjp | so basically I am allowing myself to have more room to breathe for the next two releases | 15:48 |
| odyssey4me | ah, I see | 15:48 |
| odyssey4me | okie dokey | 15:48 |
| odyssey4me | fair enough, given all the ruckus that's been caused by the zuul switchover | 15:49 |
| evrardjp | putting the head of the water | 15:49 |
| evrardjp | out of* | 15:49 |
| *** weezS has quit IRC | 15:50 | |
| *** chyka_ has joined #openstack-ansible | 15:51 | |
| *** chyka has quit IRC | 15:54 | |
| *** weezS has joined #openstack-ansible | 15:56 | |
| mhayden | hey so horizon's newton branch is missing | 15:56 |
| mhayden | is that normal? | 15:56 |
| idlemind | disk space recommendation for the lxc containers on a shared-infra host? | 15:56 |
| robcresswell | mhayden: Missing? | 15:56 |
| mhayden | https://github.com/openstack/horizon/tree/stable/newton | 15:56 |
| mhayden | unless i'm totally missing something | 15:56 |
| odyssey4me | mhayden the github sync from git.o.o is broken | 15:56 |
| robcresswell | newton-eol ? | 15:56 |
| odyssey4me | so use git.o.o as your reliable source | 15:56 |
| robcresswell | https://github.com/openstack/horizon/tree/newton-eol | 15:57 |
| odyssey4me | oh, and there's that | 15:57 |
| robcresswell | stable-newton is gone from everywhere | 15:57 |
| robcresswell | Or, it should be. idk if requirements has branched yet. | 15:57 |
| mhayden | oh, i see, that caught me off guard | 15:57 |
| mhayden | i see the newton-eol tag | 15:57 |
| *** Smeared_Beard has joined #openstack-ansible | 15:58 | |
| *** shardy has quit IRC | 15:59 | |
| *** SmearedBeard has quit IRC | 16:00 | |
| robcresswell | mhayden: Last I checked, it was only really nova and requirements that weren't tagged eol | 16:00 |
| *** a0x3e8 has joined #openstack-ansible | 16:00 | |
| robcresswell | I havent been following the official status though | 16:00 |
| mhayden | i haven't kept up with this very well lately ;) | 16:00 |
| errr | if there is an ansible module with a module.log() statement how do I run my osa playbook so that the log message will show? | 16:01 |
| errr | I tried adding -vvv and that isnt doing the job | 16:01 |
| *** Oku_OS is now known as Oku_OS-away | 16:06 | |
| cloudnull | errr: are you writing a custom module ? | 16:10 |
| cloudnull | https://github.com/openstack/openstack-ansible-plugins/blob/master/connection/ssh.py#L128 | 16:10 |
| errr | no Im adding logging to ones yall made because Im getting an annoying failure and I cant figure out why | 16:10 |
| cloudnull | so you should be able to import display | 16:11 |
| cloudnull | the v method will output information | 16:11 |
| errr | so the ansible folks said I need to be doing module.log() | 16:11 |
| errr | also as a side note i find it highly annoying standard python logging doesnt work from ansible modules. :D | 16:12 |
| cloudnull | it does | 16:12 |
| cloudnull | https://github.com/openstack/openstack-ansible-plugins/blob/master/filter/osa-filters.py#L93 | 16:12 |
| cloudnull | you just have to access the log handler | 16:12 |
| errr | are you sure? https://github.com/ansible/proposals/issues/33 | 16:12 |
| cloudnull | yes | 16:13 |
| cloudnull | https://github.com/openstack/openstack-ansible-plugins/blob/master/filter/osa-filters.py#L92-L120 | 16:13 |
| cloudnull | I'm 100% sure | 16:13 |
| errr | ok Ill try adding that. I am doing it *slightly* different than you are in that file and its not working so let me change to what you have here | 16:14 |
| odyssey4me | logan- hmm, https://github.com/openstack/openstack-ansible-galera_client/commit/65350268347c58d31d2c82ccd7485f29d83f6a3a and https://github.com/openstack/openstack-ansible-galera_server/commit/0b4c0f4fd7b83659656ea76b11101ae090e5eeab appear to be causing the py_pkgs lookup to fail as it doesn't interpret the jinja | 16:16 |
| evrardjp | robcresswell: nova not tagged | 16:16 |
| evrardjp | around 20/11 iirc | 16:17 |
| errr | actually what I was doing *was* working. I didnt think to look in the container I was looking on the deploy host. I can be such an idiot sometimes | 16:18 |
| logan- | odyssey4me: :( | 16:18 |
| logan- | odyssey4me: what if you try defining galera_client_major_version on the repo group | 16:19 |
| logan- | i think the repo_build resolves the jinja if the var exists | 16:19 |
| logan- | we have a similar situation in the nova role with barbican | 16:19 |
| Tahvok | So I've tried bringing up aio instance | 16:20 |
| odyssey4me | logan- hmm, this is because I'm using the lookup in a different way | 16:20 |
| Tahvok | Did not make any custom confiuguration, and was using HEAD | 16:20 |
| Tahvok | This is what I got: | 16:20 |
| Tahvok | https://gist.github.com/Tahvok/7a7b8b3e22c636cb2f700f011bcb3cb8 | 16:20 |
| Tahvok | Any suggestions? | 16:20 |
| *** dave-mccowan has quit IRC | 16:22 | |
| odyssey4me | logan- it's weird, because it is in defaults so it should be found | 16:23 |
| cloudnull | Tahvok: HEAD from when? | 16:23 |
| Tahvok | 2 hours ago? | 16:24 |
| cloudnull | can you look to see if the lxc-container-create role has this commit | 16:24 |
| cloudnull | https://github.com/openstack/openstack-ansible-lxc_container_create/commit/82c1dfa68fcc0fac58f8df7d09b8de955c8f8bbd#diff-4f5b08a7d7f80042ede9d9fff47cc54e | 16:24 |
| Tahvok | tried git pulling | 16:24 |
| Tahvok | The only change was this file: releasenotes/source/conf.py | 16:24 |
| Tahvok | cloudnull: yes | 16:25 |
| Tahvok | I have that commit | 16:25 |
| cloudnull | did the rescue not run ? | 16:25 |
| Tahvok | what's rescue? | 16:26 |
| cloudnull | the task Set define static mac address from an existing interface | 16:26 |
| cloudnull | is in a block | 16:26 |
| cloudnull | https://github.com/openstack/openstack-ansible-lxc_container_create/blob/master/tasks/lxc_container_config.yml#L51-L92 | 16:26 |
| cloudnull | so if the first task fails the second will pick up and continue | 16:26 |
| Tahvok | The task after the fail was: Gather hardware addresses to be used as facts | 16:28 |
| odyssey4me | and boom - gnocchi/newton is working :) https://review.openstack.org/514005 | 16:28 |
| cloudnull | nice | 16:28 |
| cloudnull | Tahvok: you should see the task "Set unique interface mac address (when no facts exist)" run | 16:29 |
| xdfil | you know it felt good when the stdout was 'boom' | 16:29 |
| cloudnull | if you don't then I've failed to trap something there and need to fix it | 16:29 |
| *** armaan has quit IRC | 16:30 | |
| cloudnull | Tahvok: was this a new aio | 16:30 |
| Tahvok | cloudnull: well, it didn't | 16:30 |
| Tahvok | cloudnull: yes, it's a new aio | 16:30 |
| *** armaan has joined #openstack-ansible | 16:30 | |
| cloudnull | were there multiple runs of the lxc-containerc-create playbook ? | 16:30 |
| xdfil | Hey, I am pretty sure that in order to get vxlan working with OVS a couple things need to happen and they seem fundamentaly opposed to how OSA does stuff | 16:30 |
| xdfil | anyone have time to discuss? | 16:30 |
| Tahvok | cloudnull: no, it's a first run of setup-hosts.yml playbook | 16:31 |
| cloudnull | ok. | 16:31 |
| * cloudnull trying to reproduce it | 16:31 | |
| Tahvok | xdfil: I have an ovs working installation. | 16:31 |
| Tahvok | What's the problem you're facing? | 16:32 |
| xdfil | with VXLAN? | 16:32 |
| Tahvok | yes | 16:32 |
| xdfil | ok let me start with this: Where is the IP address configured for the interface that corresponds to br-vxlan in your deployment | 16:32 |
| Tahvok | xdfil: let me show you our configuration. Give me a couple of minutes | 16:33 |
| xdfil | talking about neutron agents container | 16:33 |
| xdfil | I'm making the assumption it will need to be on the OVS br-tun bridge interface, and that the veth that connects to br-vxlan needs to be added to br-tun | 16:35 |
| xdfil | paste it here if you like: https://etherpad.openstack.org/p/OSA-OVS-throwdown | 16:36 |
| openstackgerrit | Merged openstack/openstack-ansible-os_swift stable/pike: Revert "Don't run sync in upgrades" https://review.openstack.org/516333 | 16:37 |
| odyssey4me | cloudnull ^ its probably best to sha bump the role in the integrated repo for that one? | 16:43 |
| cloudnull | ++ | 16:43 |
| xdfil | I curious what your thoughts are on the ODL / Repo port conflict (8181) | 16:43 |
| xdfil | who moves ? lol | 16:44 |
| Tahvok | xdfil: https://gist.github.com/Tahvok/8d522fa308041529b8248b2dfc2e6e4c | 16:44 |
| *** weezS has quit IRC | 16:44 | |
| Tahvok | Inside each neutron agent container you have to run the following: | 16:45 |
| xdfil | Tahvok: can you show me your ovs-vsctl show output? | 16:46 |
| Tahvok | ovs-vsctl add-br br-provider && ovs-vsctl add-port br-provider eth12 | 16:46 |
| Tahvok | Also, the following on the compute nodes: | 16:46 |
| *** lbragstad has quit IRC | 16:46 | |
| Tahvok | ovs-vsctl add-br br-provider && ovs-vsctl add-port br-provider bond0.2 | 16:47 |
| xdfil | So eth12 is your br-vxlan veth? | 16:47 |
| Tahvok | yes | 16:47 |
| Tahvok | wait, no | 16:47 |
| Tahvok | yes, it's eth12 | 16:48 |
| odyssey4me | Tahvok any chance you could update that gist with a complete set of host setup instructions and config to make it all go? | 16:48 |
| xdfil | Tahvok: sorry, your eth10 is vxlan and your eth12 is vlan | 16:49 |
| openstackgerrit | Antony Messerli proposed openstack/openstack-ansible-ops master: Abort run if VMs aren't up before timeout https://review.openstack.org/517371 | 16:49 |
| Tahvok | odyssey4me: added instuctions to the gist | 16:52 |
| xdfil | Tahvok: the parts you are showing me arent the parts I'm having problems with exactly | 16:52 |
| Tahvok | xdfil: yes, that's because your network inside openstack will use the vxlan interface | 16:52 |
| xdfil | I need to see what's inside the agents container | 16:52 |
| xdfil | and I need to see how OVS is configured | 16:53 |
| odyssey4me | Tahvok fantastic, thanks | 16:53 |
| Tahvok | xdfil: https://gist.github.com/Tahvok/a0a9be64bc11bb0209a36d9d12c9e685 | 16:53 |
| odyssey4me | Tahvok if you can keep that updated with whatever other questions arise, we can perhaps help to make that a documented config | 16:53 |
| Tahvok | br-int and br-tun are created automatically | 16:54 |
| openstackgerrit | Kevin Carter (cloudnull) proposed openstack/openstack-ansible-lxc_container_create master: Correct mac generation block https://review.openstack.org/517372 | 16:54 |
| cloudnull | Tahvok: ^ | 16:54 |
| jafeha | Tahvok: aaaaw how cool, i'll diff my setup tomorrow against yours. at least my veth conf seems not to differ that much | 16:54 |
| Tahvok | You only need to add br-provider and the appropriate port on it | 16:54 |
| cloudnull | mind testing that in your env ? | 16:54 |
| jafeha | Tahvok: thank you! | 16:54 |
| openstackgerrit | Merged openstack/openstack-ansible-os_octavia master: Improve event streaming with a better Octavia verison https://review.openstack.org/514852 | 16:54 |
| jafeha | odyssey4me: as soon as our setup is up and running i could also assist improving the ovs conf documentation | 16:55 |
| *** David_Orange has quit IRC | 16:55 | |
| Tahvok | Would be nice if ovs parts would be made automaticaly with ansible.. But I think that's too much to ask, as people will have too different configurations | 16:56 |
| odyssey4me | jafeha Tahvok fantastic, it'd be great to see another example appendix to show how | 16:56 |
| xdfil | Tahvok: wow, so no "physical" ports need to be added to the br-tun | 16:56 |
| *** jvidal has quit IRC | 16:56 | |
| xdfil | I've been obsessing on that detail for months | 16:56 |
| Tahvok | xdfil: you don't do anything with br-tun | 16:56 |
| Tahvok | You only add br-provider | 16:56 |
| jafeha | Tahvok: we've written some ansible playbooks to simply drop our ovs configuration before running osa | 16:57 |
| Tahvok | Everything else is managed by openstack | 16:57 |
| Bico_Fino | You have to add br-provider Tahvok ? | 16:57 |
| Tahvok | Bico_Fino: yes | 16:57 |
| * odyssey4me really ensjoys watching the magic of a community at work together. | 16:57 | |
| Tahvok | Everything I said here, is according to the networking guide docs | 16:58 |
| Bico_Fino | network_mappings: "physnet1:br-provider" | 16:58 |
| Bico_Fino | this ^? | 16:58 |
| Tahvok | This is what I used: https://docs.openstack.org/ocata/networking-guide/deploy-ovs-selfservice.html | 16:59 |
| cloudnull | odyssey4me: ++ | 16:59 |
| jafeha | we've adapted our lb setup to ovs and might still have some dead code, so it is really great to see that there are others out there trying the same stuff right now. | 16:59 |
| Tahvok | Bico_Fino: yeah, I forgot | 16:59 |
| jafeha | this will be really useful for bugsquashing | 17:00 |
| Tahvok | You also need to add the following in user_variables.yml: | 17:00 |
| jafeha | even though my collegue will have to take over those conversations from next week on, i'll be on vacation for a while | 17:00 |
| Tahvok | https://gist.github.com/Tahvok/5068e5ecb96417bacae224a44e2b3fb5 | 17:01 |
| Tahvok | This is needed as well ^ | 17:01 |
| Bico_Fino | Tahvok my config is almost the same, let me post on that gist | 17:01 |
| odyssey4me | Tahvok can you rather add them all into a single gist? | 17:01 |
| odyssey4me | It'll be easier to track. | 17:02 |
| Tahvok | odyssey4me: yeah, sec | 17:02 |
| Bico_Fino | I'm using neutron_plugin_type: ml2.ovs.dvr | 17:02 |
| Bico_Fino | I'm almost done with setup-openstack.yml | 17:03 |
| Tahvok | odyssey4me: updated. Please understand that this is for ocata release. We have not moved to pike yet | 17:03 |
| Bico_Fino | oh | 17:03 |
| Bico_Fino | I'm deploying Pike. | 17:03 |
| *** armaan has quit IRC | 17:03 | |
| *** armaan has joined #openstack-ansible | 17:03 | |
| odyssey4me | Tahvok if you change the title of the gist to something like 'OVS with OSA for Ocata' then it'll be more obvious | 17:03 |
| Bico_Fino | Tahvok You are consolidating all the info into this one -> https://gist.github.com/Tahvok/5068e5ecb96417bacae224a44e2b3fb5 ? | 17:04 |
| odyssey4me | nope, this one: https://gist.github.com/Tahvok/8d522fa308041529b8248b2dfc2e6e4c | 17:04 |
| Tahvok | Bico_Fino: I didn't see much difference with pike. Just ensure that you follow the pike guide, and check whether roles already not default to what you need: https://docs.openstack.org/neutron/latest/admin/deploy-ovs.html | 17:05 |
| Bico_Fino | Going test a playbook to add the ovs interface | 17:06 |
| Bico_Fino | br-provider | 17:06 |
| odyssey4me | it'd be super-great if that interface bit could be added to the neutron role or wherever it's best suited so that it can be automated | 17:07 |
| *** markvoelker_ has joined #openstack-ansible | 17:07 | |
| openstackgerrit | Merged openstack/openstack-ansible-os_tacker stable/pike: Initial OSA zuul v3 role jobs https://review.openstack.org/516306 | 17:07 |
| xdfil | I can't decide if I'm happy or mad | 17:08 |
| xdfil | I'm both, thats possible heh | 17:08 |
| Tahvok | odyssey4me: on sunday, if I won't forget, I could add the selfservice and provider network configurations as well | 17:09 |
| openstackgerrit | Merged openstack/openstack-ansible-os_nova master: Update test image for nova-lxd https://review.openstack.org/515020 | 17:10 |
| Tahvok | odyssey4me: the best I could do, was adding the description to the gist | 17:10 |
| odyssey4me | Tahvok tyvm, that will help a lot of people | 17:11 |
| Tahvok | This is the gist: https://gist.github.com/Tahvok/8d522fa308041529b8248b2dfc2e6e4c | 17:11 |
| *** markvoelker_ has quit IRC | 17:11 | |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible-os_nova stable/pike: Update test image for nova-lxd https://review.openstack.org/517376 | 17:12 |
| *** markvoelker has joined #openstack-ansible | 17:12 | |
| *** markvoelker has quit IRC | 17:13 | |
| xdfil | Tahvok: odyssey4me I think the most important thing that needs to be made clear in the docs is how OVS needs to be configured. The docs until now say nothing about vxlan details on OVS. The configuration of that interface has been a mystery to me for months. | 17:14 |
| xdfil | Especially when you come accross other non-osa examples where people are putting their vxlan endpoint address on the OVS brdige interface | 17:15 |
| Bico_Fino | Yeap, you don't find much information about OVS with osa | 17:15 |
| Bico_Fino | linuxbridge is easier, but we got HA issues with neutron routers. | 17:15 |
| odyssey4me | that's due to most consumers of OSA not using it - so it takes a few consumers to figure it out, then own the changes that need to be made to the docs and code to make it simpler and better | 17:16 |
| xdfil | So I for some reason was assuming that in order for OVS to do vxlan stuff the endpoint interface actually had to be on an OVS bridge. But it doesn't which makes sense because its layer-3 at that point | 17:16 |
| *** vnogin has quit IRC | 17:16 | |
| Tahvok | We worked for 2 months on the osa project to get a working osa with ovs configuration | 17:16 |
| Tahvok | But it was worth it | 17:16 |
| odyssey4me | so, thank you guys for taking what was a lab experiment back in newton, and starting to figure out how to make it work for real | 17:17 |
| Bico_Fino | Tahvok you changed playbooks or anything else ? | 17:17 |
| Tahvok | odyssey4me: good to know that it was an experiment :O | 17:17 |
| Bico_Fino | :D | 17:18 |
| *** markvoelker has joined #openstack-ansible | 17:18 | |
| Tahvok | If we knew it was only an experiment, we wouldn't go with ovs XD | 17:18 |
| Tahvok | Bico_Fino: didn't touch anything. Only the user_variables as I provided in the gist | 17:18 |
| odyssey4me | it was done by automagically, who has sadly moved on to other work - he put together http://trumant.github.io/openstack-ansible-openvswitch.html and http://trumant.github.io/openstack-ansible-dvr-with-openvswitch.html | 17:18 |
| Tahvok | Yes, this are the guides we based our work on | 17:19 |
| Bico_Fino | Tahvok thanks for the info. :) | 17:19 |
| odyssey4me | and the work to make that possible was put together by him and another guys from jmccrory's employer IIRC... | 17:20 |
| xdfil | Tahvok: yes thanks | 17:20 |
| openstackgerrit | Antony Messerli proposed openstack/openstack-ansible-ops master: Specify containers to destroy https://review.openstack.org/517380 | 17:20 |
| odyssey4me | so unfortunately, when they had to move on to other work the integration lost momentum and it remained something we knew possible, but that some bits were missing to close the loop | 17:20 |
| Bico_Fino | I think my setup-openstack.yml is almost done | 17:21 |
| Bico_Fino | Going add the br-provider and make some tests | 17:21 |
| Bico_Fino | if it works I will put a gist with the configs | 17:21 |
| Tahvok | What's setup-openstack.yml? | 17:21 |
| Bico_Fino | the playbook | 17:21 |
| Tahvok | You edit the playbook? | 17:21 |
| Bico_Fino | no | 17:21 |
| *** markvoelker_ has joined #openstack-ansible | 17:21 | |
| Bico_Fino | I mean openstack-ansible setup-openstack.yml | 17:21 |
| Bico_Fino | the command. :P | 17:21 |
| Tahvok | cloudnull: you reverted here? https://review.openstack.org/#/c/517372/1/tasks/lxc_container_config.yml | 17:22 |
| *** markvoelker has quit IRC | 17:23 | |
| cloudnull | kinda, i combined the two tasks | 17:23 |
| cloudnull | it was something like 4, then 2, now 1 :D | 17:23 |
| openstackgerrit | Merged openstack/openstack-ansible-os_gnocchi stable/newton: Initial OSA zuul v3 role jobs https://review.openstack.org/514005 | 17:23 |
| Tahvok | Hope it works | 17:24 |
| *** markvoelker has joined #openstack-ansible | 17:24 | |
| Tahvok | I wanted to test aio for first time today | 17:24 |
| cloudnull | on a new aio it seems to be ok. | 17:24 |
| Tahvok | Guess I'll do it next week | 17:24 |
| Tahvok | Anyway, bye for now | 17:24 |
| cloudnull | take care Tahvok | 17:24 |
| Bico_Fino | bye | 17:25 |
| *** markvoelker_ has quit IRC | 17:26 | |
| *** markvoelker_ has joined #openstack-ansible | 17:28 | |
| *** mwynne has quit IRC | 17:28 | |
| *** markvoelker has quit IRC | 17:29 | |
| *** lbragstad has joined #openstack-ansible | 17:29 | |
| *** markvoelker_ has quit IRC | 17:34 | |
| *** germs has joined #openstack-ansible | 17:38 | |
| *** electrofelix has quit IRC | 17:44 | |
| *** chhavi has quit IRC | 17:45 | |
| *** chhavi__ has quit IRC | 17:46 | |
| *** weezS has joined #openstack-ansible | 17:49 | |
| *** bruvik_ has joined #openstack-ansible | 17:49 | |
| openstackgerrit | Kevin Carter (cloudnull) proposed openstack/openstack-ansible-os_ceilometer stable/newton: Initial OSA zuul v3 role jobs https://review.openstack.org/513999 | 17:55 |
| *** vnogin has joined #openstack-ansible | 17:55 | |
| *** vnogin has quit IRC | 17:55 | |
| odyssey4me | was that it cloudnull ? (whitespace trailing btw) | 17:56 |
| cloudnull | odyssey4me: evrardjp: if you're still around - https://review.openstack.org/#/c/517372 - this fixes a bug i created | 17:56 |
| odyssey4me | I'm thinking of pushing up a horde of patches to set the eol tag instead of the branch all over the place | 17:56 |
| cloudnull | i'm still testing | 17:56 |
| odyssey4me | cloudnull hey, that's mhayden's job | 17:56 |
| cloudnull | however I needed that to go | 17:56 |
| evrardjp | odyssey4me: you should wait a little that everything is using EOL | 17:56 |
| evrardjp | but yes why not | 17:56 |
| evrardjp | odyssey4me: haha | 17:56 |
| odyssey4me | evrardjp it's breaking too many jobs, and not difficult to do | 17:57 |
| evrardjp | (on the mhayden job) | 17:57 |
| cloudnull | figured i'd throw it that in place and continue beating on it | 17:57 |
| odyssey4me | it's also breaking some RPC jobs now | 17:57 |
| evrardjp | odyssey4me: fine for me | 17:57 |
| *** poopcat has joined #openstack-ansible | 17:57 | |
| odyssey4me | cloudnull LGTM | 17:58 |
| odyssey4me | nice to have brought those tasks from 4 to 1 | 17:58 |
| cloudnull | indeed | 17:59 |
| evrardjp | yeah | 17:59 |
| cloudnull | I was using blocks wrong | 17:59 |
| *** albertcard1 has joined #openstack-ansible | 17:59 | |
| cloudnull | i think we could make that better still, but another day | 18:00 |
| odyssey4me | agreed | 18:00 |
| * odyssey4me preps up his repo changes script for an branch->eol tag bonanza | 18:00 | |
| *** gouthamr has joined #openstack-ansible | 18:00 | |
| cloudnull | woot! | 18:00 |
| * cloudnull loves a good pr bonanza | 18:01 | |
| *** poopcat has quit IRC | 18:02 | |
| *** vnogin has joined #openstack-ansible | 18:04 | |
| *** pbandark has quit IRC | 18:07 | |
| *** pbandark has joined #openstack-ansible | 18:07 | |
| *** gkadam has quit IRC | 18:08 | |
| *** gkadam has joined #openstack-ansible | 18:09 | |
| Bico_Fino | hum. openstack-ansible setup-openstack.yml failed, http://paste.openstack.org/show/625375/ | 18:10 |
| *** pcaruana has quit IRC | 18:10 | |
| *** esberglu has quit IRC | 18:10 | |
| Bico_Fino | ansible_br_vxlan | 18:10 |
| *** esberglu has joined #openstack-ansible | 18:10 | |
| cloudnull | Bico_Fino: if br-vxlan defined in the openstack_user_config.yml | 18:12 |
| cloudnull | and is that bridge available on your hosts? | 18:12 |
| Bico_Fino | yes, it is. | 18:12 |
| Bico_Fino | hum | 18:12 |
| *** pbandark has quit IRC | 18:13 | |
| Bico_Fino | I had to fix it | 18:13 |
| Bico_Fino | after setup-hosts | 18:13 |
| Bico_Fino | should I run setup-hosts again? | 18:13 |
| cloudnull | I'd just run lxc-container-create again. however you might try purging the cached facts in /etc/openstack_deploy/ansible_facts | 18:14 |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible-os_cinder stable/newton: Change developer_mode branch to newton-eol https://review.openstack.org/517391 | 18:14 |
| cloudnull | then rerun the failed playbook | 18:14 |
| Bico_Fino | sounds good | 18:15 |
| Bico_Fino | Going try that, thanks! | 18:15 |
| cloudnull | let me know how it goes. | 18:15 |
| *** esberglu has quit IRC | 18:15 | |
| odyssey4me | evrardjp cloudnull does https://review.openstack.org/517391 look like a satisfactory template? | 18:15 |
| Bico_Fino | sure! | 18:15 |
| cloudnull | odyssey4me: LGTM | 18:16 |
| evrardjp | odyssey4me: I think you need to change things in tests | 18:16 |
| evrardjp | don't we? | 18:16 |
| odyssey4me | evrardjp yes, for the actual tests | 18:17 |
| odyssey4me | I'll do that too, but seperately | 18:17 |
| odyssey4me | tests only affects the commonly used stuff anyway - here I'll be touching everything we have | 18:17 |
| cloudnull | the tests repo in newton is pinned anyway | 18:17 |
| odyssey4me | yup | 18:17 |
| evrardjp | yeah I remember the tests repo pin | 18:18 |
| evrardjp | Weirdly I thought we had something in repo too | 18:18 |
| evrardjp | but it doesn't seem so | 18:18 |
| evrardjp | so go ahead I'd say | 18:18 |
| odyssey4me | this will make it simple to know which upstream services are eol already ;) | 18:18 |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible-os_aodh stable/newton: Change developer_mode branch to newton-eol https://review.openstack.org/517396 | 18:19 |
| evrardjp | well, all of them except two | 18:19 |
| evrardjp | nova and ... | 18:19 |
| evrardjp | I can't remember | 18:19 |
| evrardjp | (tempest being still special) | 18:20 |
| evrardjp | would the second be heat? | 18:20 |
| evrardjp | just a sec | 18:20 |
| odyssey4me | evrardjp don't bother - my script will tell us | 18:20 |
| odyssey4me | the reviews would come in faster if I wasn't vpn'd :/ | 18:20 |
| evrardjp | ironic | 18:21 |
| evrardjp | it's just opening a ML thread | 18:21 |
| evrardjp | not that hard :p | 18:21 |
| *** esberglu has joined #openstack-ansible | 18:21 | |
| odyssey4me | oh... | 18:21 |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible-os_ceilometer stable/newton: Change developer_mode branch to newton-eol https://review.openstack.org/517397 | 18:21 |
| * odyssey4me is behind on ML thread things | 18:21 | |
| evrardjp | so ironic should be alright apparently | 18:22 |
| evrardjp | which means they will have a EOL tag soon | 18:23 |
| evrardjp | so if you avoid EOL on nova and ironic you're good | 18:23 |
| openstackgerrit | Kevin Carter (cloudnull) proposed openstack/openstack-ansible master: Run gate playbooks in parallel https://review.openstack.org/497742 | 18:24 |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible-os_glance stable/newton: Change developer_mode branch to newton-eol https://review.openstack.org/517398 | 18:24 |
| *** wlmbasson has quit IRC | 18:24 | |
| *** chyka has joined #openstack-ansible | 18:26 | |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible-os_heat stable/newton: Change developer_mode branch to newton-eol https://review.openstack.org/517399 | 18:27 |
| *** bruvik_ has quit IRC | 18:28 | |
| *** chyka_ has quit IRC | 18:29 | |
| openstackgerrit | Kevin Carter (cloudnull) proposed openstack/openstack-ansible master: Add infra healthchecks https://review.openstack.org/495492 | 18:29 |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible-os_keystone stable/newton: Change developer_mode branch to newton-eol https://review.openstack.org/517400 | 18:30 |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible-os_sahara stable/newton: Change developer_mode branch to newton-eol https://review.openstack.org/517401 | 18:33 |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible-os_swift stable/newton: Change developer_mode branch to newton-eol https://review.openstack.org/517402 | 18:35 |
| openstackgerrit | Merged openstack/openstack-ansible-os_sahara stable/newton: Updated from OpenStack Ansible Tests https://review.openstack.org/516262 | 18:38 |
| *** armaan has quit IRC | 18:42 | |
| *** armaan has joined #openstack-ansible | 18:43 | |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible-os_neutron stable/newton: Change developer_mode branch to newton-eol https://review.openstack.org/517405 | 18:44 |
| openstackgerrit | Merged openstack/openstack-ansible master: Vagrantfile: Remove openSUSE Leap 42.2 https://review.openstack.org/517268 | 18:45 |
| odyssey4me | right, that's all right now: https://review.openstack.org/#/q/topic:newton-eol-dev-mode-branch-change+(status:open+OR+status:merged) | 18:46 |
| odyssey4me | hmm, it seems a few got missed for some reason | 18:47 |
| openstackgerrit | Kevin Carter (cloudnull) proposed openstack/openstack-ansible-os_ceilometer stable/newton: Initial OSA zuul v3 role jobs https://review.openstack.org/513999 | 18:48 |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible-os_magnum stable/newton: Change developer_mode branch to newton-eol https://review.openstack.org/517409 | 18:55 |
| openstackgerrit | Antony Messerli proposed openstack/openstack-ansible-ops master: Specify containers to destroy https://review.openstack.org/517380 | 18:56 |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible-os_horizon stable/newton: Change developer_mode branch to newton-eol https://review.openstack.org/517410 | 18:57 |
| *** hachi has joined #openstack-ansible | 18:58 | |
| *** bruvik_ has joined #openstack-ansible | 18:58 | |
| *** stuartgr has quit IRC | 19:01 | |
| *** bruvik_ has quit IRC | 19:02 | |
| *** albertcard1 has quit IRC | 19:07 | |
| *** poopcat has joined #openstack-ansible | 19:08 | |
| *** albertcard1 has joined #openstack-ansible | 19:13 | |
| *** poopcat has quit IRC | 19:15 | |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible-os_aodh stable/newton: Change developer_mode branch to newton-eol https://review.openstack.org/517396 | 19:21 |
| openstackgerrit | Kevin Carter (cloudnull) proposed openstack/openstack-ansible-os_ceilometer stable/newton: Initial OSA zuul v3 role jobs https://review.openstack.org/513999 | 19:23 |
| *** vishwanathj has quit IRC | 19:23 | |
| Bico_Fino | cloudnull hum. failed for gnocchi now. | 19:25 |
| Bico_Fino | No file was found when using with_first_found. Use the 'skip: true' option to allow this task to be skipped if no files are found | 19:28 |
| *** smatzek_ is now known as smatzek | 19:30 | |
| Bico_Fino | strange, I ran the playbook alone and it worked openstack-ansible os-gnocchi-install.yml | 19:33 |
| *** dave-mccowan has joined #openstack-ansible | 19:33 | |
| *** vnogin has quit IRC | 19:34 | |
| *** vnogin has joined #openstack-ansible | 19:35 | |
| *** vnogin has quit IRC | 19:36 | |
| *** vnogin has joined #openstack-ansible | 19:37 | |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible-os_horizon stable/newton: Change developer_mode branch to newton-eol https://review.openstack.org/517410 | 19:39 |
| openstackgerrit | Merged openstack/openstack-ansible-os_sahara stable/pike: Initial OSA zuul v3 role jobs https://review.openstack.org/516600 | 19:40 |
| *** albertcard1 is now known as poopcat | 19:48 | |
| *** cathrichardson has joined #openstack-ansible | 19:51 | |
| *** gkadam has quit IRC | 19:54 | |
| *** dave-mccowan has quit IRC | 19:56 | |
| *** vnogin has quit IRC | 20:00 | |
| *** hachi has quit IRC | 20:13 | |
| *** dave-mccowan has joined #openstack-ansible | 20:15 | |
| *** thorst_ has quit IRC | 20:17 | |
| openstackgerrit | Merged openstack/openstack-ansible-os_cinder stable/newton: Change developer_mode branch to newton-eol https://review.openstack.org/517391 | 20:18 |
| Smeared_Beard | hey, after running OSA, I thought neutron agent on the compute host, would be in a container, but it's not.. is this expected and is there an option to have it in a container? I looking now | 20:18 |
| *** pbandark has joined #openstack-ansible | 20:18 | |
| Smeared_Beard | at the docs to see if I can find an option | 20:18 |
| *** vnogin has joined #openstack-ansible | 20:24 | |
| *** vnogin has quit IRC | 20:24 | |
| *** smatzek has quit IRC | 20:25 | |
| *** vnogin has joined #openstack-ansible | 20:35 | |
| *** dave-mccowan has quit IRC | 20:35 | |
| *** cathrichardson has quit IRC | 20:36 | |
| *** mwynne has joined #openstack-ansible | 20:38 | |
| *** vnogin has quit IRC | 20:40 | |
| *** xingchao has joined #openstack-ansible | 20:41 | |
| *** vnogin has joined #openstack-ansible | 20:43 | |
| cloudnull | Bico_Fino: theres | 20:43 |
| cloudnull | **there's a lookup that checks the os | 20:43 |
| cloudnull | if the facts are stale it can cause that issue | 20:43 |
| Bico_Fino | After I ran it alone | 20:43 |
| Bico_Fino | it worked | 20:44 |
| cloudnull | success ! :D | 20:44 |
| Bico_Fino | but got the same issue on utility one. | 20:44 |
| Bico_Fino | I ran it alone, worked it again | 20:44 |
| Bico_Fino | I'm using ubuntu 16.04 | 20:44 |
| cloudnull | when I've ran into this issue i've purged the facts under /etc/openstack_deploy/ansible_facts | 20:44 |
| cloudnull | however it may be something worth looking into more broadly | 20:44 |
| *** xingchao has quit IRC | 20:44 | |
| Bico_Fino | maybe some race condition? | 20:45 |
| Bico_Fino | Because if you run the playbook alone, it works. | 20:45 |
| cloudnull | could be | 20:45 |
| cloudnull | I've also thought maybe its facts that are on the cusp of expiring | 20:45 |
| cloudnull | but hard to tell/track down | 20:45 |
| cloudnull | Smeared_Beard: neutron agents are in containers on your network nodes. | 20:45 |
| cloudnull | however they're running on baremetal on compute nodes. | 20:46 |
| Bico_Fino | running setup-openstack again | 20:46 |
| Bico_Fino | lets see how that goes | 20:46 |
| cloudnull | things like neutron-linuxbridge-agent needs to run along side nova compute | 20:46 |
| cloudnull | Bico_Fino: ++ | 20:46 |
| cloudnull | I'll be around, let me know what you find out. | 20:46 |
| Bico_Fino | thanks. | 20:47 |
| *** vnogin has quit IRC | 20:51 | |
| Smeared_Beard | cloudnull cool, that is what I am seeing.. I guess I didn't really think about it so was unexpected when there was no container for L3 agent | 20:51 |
| openstackgerrit | Merged openstack/openstack-ansible-ops master: add example basic resource generation scripts https://review.openstack.org/515743 | 20:57 |
| *** galstrom is now known as galstrom_zzz | 21:10 | |
| *** jwitko_ has joined #openstack-ansible | 21:10 | |
| *** vnogin has joined #openstack-ansible | 21:11 | |
| *** jwitko has quit IRC | 21:14 | |
| *** vnogin has quit IRC | 21:16 | |
| *** jwitko_ has quit IRC | 21:16 | |
| idlemind | does a vagrant file exist for quickly standing up a deployment host? | 21:19 |
| cloudnull | we do have a vagrant file, but ive never used it . | 21:21 |
| idlemind | where would that gem be nestled by chance? | 21:22 |
| cloudnull | https://github.com/openstack/openstack-ansible/blob/master/Vagrantfile | 21:23 |
| idlemind | ah how'd i miss that one :) | 21:23 |
| idlemind | thx | 21:23 |
| cloudnull | a lot of the roles have a Vagrantfile too | 21:23 |
| cloudnull | https://github.com/openstack/openstack-ansible-os_nova/blob/master/Vagrantfile | 21:23 |
| cloudnull | which may be useful for testing a specific role | 21:23 |
| *** vnogin has joined #openstack-ansible | 21:26 | |
| *** xingchao has joined #openstack-ansible | 21:26 | |
| *** vnogin has quit IRC | 21:30 | |
| *** xingchao has quit IRC | 21:31 | |
| *** Brew has quit IRC | 21:36 | |
| *** Smeared_Beard has quit IRC | 21:38 | |
| *** jwitko has joined #openstack-ansible | 21:44 | |
| *** SmearedBeard has joined #openstack-ansible | 21:44 | |
| Bico_Fino | cloudnull worked! | 21:45 |
| cloudnull | nice! | 21:45 |
| *** vnogin has joined #openstack-ansible | 21:49 | |
| Bico_Fino | cloudnull doing the ovs stuff now | 21:50 |
| *** vnogin has quit IRC | 21:50 | |
| *** SmearedBeard has quit IRC | 21:51 | |
| *** xingchao has joined #openstack-ansible | 22:01 | |
| cloudnull | that sounds like a BLAST! | 22:04 |
| * cloudnull loves OvS | 22:04 | |
| jrosser | cloudnull: you tried anything like this? https://community.mellanox.com/docs/DOC-2964 | 22:05 |
| cloudnull | nope | 22:05 |
| cloudnull | 1. Linux Kernel >= 4.13-rc5 -- too new for my blood | 22:06 |
| cloudnull | :D | 22:06 |
| * cloudnull uses openSUSE tumbleweed w/ 4.13.8 for a workstation | 22:07 | |
| cloudnull | that draft looks cool though | 22:08 |
| cloudnull | jrosser: do you all run w/ mellanox gear? | 22:08 |
| jrosser | nearly all the nic yes | 22:09 |
| jrosser | but not switches | 22:10 |
| *** threestrands has joined #openstack-ansible | 22:10 | |
| *** threestrands has quit IRC | 22:10 | |
| *** threestrands has joined #openstack-ansible | 22:10 | |
| cloudnull | I've had good experiences with their connect x3-pros | 22:10 |
| jrosser | i am interested in the recent ovs chatter here becasue if i can make that work i can slip ASAP^2 in behind it without much difficulty | 22:10 |
| jrosser | i have connectx4-lx pretty much everywhere | 22:12 |
| cloudnull | OVS is something we're definitely compatible with. | 22:12 |
| cloudnull | however IDK what you'd have to do to make ASAP2 go | 22:13 |
| jrosser | that mellanox doc is the first i've seen from them claiming sensible openstack interop | 22:13 |
| cloudnull | but I can't imagine it would require much hackery once your deployment was up and running with ovs | 22:13 |
| jrosser | there was always a pile of caveats up to now | 22:13 |
| cloudnull | based on the doc they're just using openvswitch with the ml2 plugin | 22:16 |
| *** vnogin has joined #openstack-ansible | 22:16 | |
| *** xingchao has quit IRC | 22:16 | |
| cloudnull | and whitelisting, which can be added into config using config_template. | 22:16 |
| jrosser | it references os-vif, is that part of a normal deployment | 22:16 |
| *** esberglu has quit IRC | 22:18 | |
| cloudnull | ah yes. that would have to be added in as a plugin. | 22:18 |
| cloudnull | however https://github.com/openstack/requirements/blob/master/upper-constraints.txt#L465 | 22:18 |
| cloudnull | its part of requirements | 22:18 |
| jrosser | they do cunningly give an example which is east-west traffic i note | 22:18 |
| cloudnull | so we should have a wheel for it | 22:18 |
| jrosser | north-south is another matter | 22:18 |
| cloudnull | https://github.com/openstack/requirements/blob/stable/ocata/upper-constraints.txt#L231 | 22:19 |
| cloudnull | its even in ocata | 22:19 |
| cloudnull | so all the parts needed should be there. | 22:19 |
| cloudnull | https://docs.openstack.org/os-vif/latest/user/plugins/linux-bridge.html#supported-vif-types | 22:20 |
| cloudnull | it even has linux bridge support | 22:20 |
| cloudnull | so maybe that thing is already in service | 22:20 |
| * cloudnull checking | 22:20 | |
| cloudnull | its not used by default in pike | 22:21 |
| cloudnull | however it is built as a wheel | 22:21 |
| cloudnull | http://paste.openstack.org/show/625394/ | 22:21 |
| jrosser | thats cool, like you say its mostly there | 22:22 |
| jrosser | i would love to see similar things applied in the network nodes | 22:22 |
| cloudnull | it looks like we could just add it as a dep | 22:23 |
| jrosser | but the current nics are missing the nat parts in hardware | 22:23 |
| jrosser | next generation will do that but i expect the drivers will take some time to catch up | 22:23 |
| cloudnull | and all of the things, except the pci pass through, would be there. | 22:23 |
| *** lbragstad has quit IRC | 22:24 | |
| cloudnull | ah ha! | 22:26 |
| cloudnull | http://paste.openstack.org/show/625395/ | 22:26 |
| cloudnull | its actually used by nova directly | 22:26 |
| cloudnull | and installed by default | 22:26 |
| spotz | cloudnull when you arriving in syd? | 22:26 |
| cloudnull | so its there | 22:26 |
| cloudnull | spotz: sunday | 22:26 |
| spotz | cloudnull: Cool:) | 22:26 |
| cloudnull | spotz: you ? | 22:27 |
| spotz | cloudnull: I'm at the airport now, arrive 6am Sat. Doing the Upstream Institute | 22:27 |
| cloudnull | jrosser: so in fact, if you deployed w/ OVS, had kernel 4.13.x, and a compatible connectx nic you'd only need to map out what the pass though devices for your compute hosts and you'd be good to go | 22:28 |
| cloudnull | spotz: ah. cool! | 22:28 |
| cloudnull | well then I'll see you in 48 hours :D | 22:28 |
| jrosser | yeah, and all th stuff in their doc would be done as part of prepping the compute node before deploy | 22:28 |
| spotz | cloudnull: Sweet:) Safe travels | 22:29 |
| cloudnull | jrosser: indeed. | 22:29 |
| cloudnull | spotz: you too. | 22:29 |
| cloudnull | have fun at the upstream institute ! | 22:29 |
| jrosser | while on the networking topic, have you seen networking-vpp? | 22:29 |
| jrosser | becasue thats super interesting for L3 functions | 22:30 |
| cloudnull | nope. | 22:31 |
| * cloudnull reading https://github.com/openstack/networking-vpp/blob/master/README.rst | 22:31 | |
| jrosser | cloudnull: https://www.youtube.com/watch?v=for1CLcQ9Q4 | 22:32 |
| cloudnull | looks like they're setting expected state and communicating through etcd ? | 22:32 |
| *** vnogin has quit IRC | 22:32 | |
| cloudnull | interesting . | 22:32 |
| jrosser | they way they implement the openstack -vpp stuff wants to own everything including the compute nodes | 22:32 |
| jrosser | yeah so mix/match is hard | 22:32 |
| jrosser | but the technology is incredible, that video is well worth a watch | 22:33 |
| cloudnull | ah, no lxb or ovs | 22:34 |
| * cloudnull watching | 22:34 | |
| openstackgerrit | Merged openstack/openstack-ansible-lxc_container_create master: Correct mac generation block https://review.openstack.org/517372 | 22:42 |
| *** thorst has joined #openstack-ansible | 22:44 | |
| *** thorst has quit IRC | 22:45 | |
| cloudnull | openstack-vpp looks like it just needs iommu enabled and an isolated vxlan capable nic. | 22:46 |
| cloudnull | if you want vxlan-gpe | 22:46 |
| cloudnull | else any vlan capable nic will work | 22:46 |
| cloudnull | looks cool | 22:46 |
| *** thorst has joined #openstack-ansible | 22:49 | |
| jrosser | the potential performance is so many orders of magnitude ahead of what you can do in other software with bridges/ovs or linux routers | 22:51 |
| cloudnull | my amd servers have /sys/class/iommu/ivhd0 so I may give this a try | 22:52 |
| jrosser | they use etc to share mac addresses between the compute and network nodes | 22:52 |
| jrosser | *etcd | 22:52 |
| jrosser | which i'm not sure what to make of | 22:53 |
| cloudnull | etcd is a thing, it works | 22:53 |
| cloudnull | we even have a role for it :D | 22:53 |
| *** sxc731 has joined #openstack-ansible | 22:53 | |
| jrosser | oh sure, but if you want networking-vpp to interoperate with linux-bridge or ovs..... | 22:54 |
| cloudnull | it looks like it has to be vpp or not | 22:54 |
| cloudnull | so mixed deployments don't seem to be possible | 22:54 |
| * cloudnull just looking at the docs | 22:55 | |
| jrosser | yes, ideally i'd like to be able to do mellanox/asap2/ovs on compute nodes and vpp on the network nodes | 22:55 |
| cloudnull | looks like odl has a vpp implementation too | 23:02 |
| *** thorst has quit IRC | 23:06 | |
| *** chyka_ has joined #openstack-ansible | 23:09 | |
| *** chyka has quit IRC | 23:12 | |
| *** chyka_ has quit IRC | 23:13 | |
| gunix | cloudnull: are you still online? | 23:14 |
| cloudnull | i am | 23:14 |
| gunix | you know the backup field in horizon, under volumes? | 23:15 |
| * cloudnull nerding out on vpp videos | 23:15 | |
| gunix | that makes a snapshot of the volume on cinder and after that moves it to swift | 23:15 |
| cloudnull | yes? | 23:15 |
| gunix | does it still work if you deploy OSA with CEPH for block storage and object storage (also with glance configured to use ceph), but without installing swift? | 23:16 |
| cloudnull | I believe cinder backup needs an object store to make that go. | 23:16 |
| cloudnull | but i think you could use ceph with RGW | 23:17 |
| cloudnull | as the swift store. | 23:17 |
| cloudnull | https://docs.openstack.org/cinder/latest/admin/blockstorage-volume-backups.html | 23:17 |
| cloudnull | looks like there's an NFS store option too | 23:18 |
| jrosser | cloudnull: practical nerding with vpp here https://wiki.fd.io/view/VPP#Use_Cases | 23:19 |
| jrosser | look at the "home gateway" one | 23:19 |
| gunix | cloudnull: so there must be a backup driver setting within cinder | 23:20 |
| gunix | wait, checking | 23:20 |
| gunix | cloudnull: found it: https://bpaste.net/show/dd1277f3b300 | 23:21 |
| gunix | and here is the docu: https://docs.openstack.org/mitaka/config-reference/block-storage/backup/ceph-backup-driver.html | 23:22 |
| gunix | thank you! | 23:22 |
| cloudnull | np :D | 23:23 |
| *** weezS has quit IRC | 23:28 | |
| gunix | cloudnull: also, found this: https://bpaste.net/show/53c6959b57f1 | 23:31 |
| cloudnull | ah look at that :) | 23:31 |
| cloudnull | looks like we already had ceph suppot | 23:32 |
| * cloudnull is an idiot | 23:32 | |
| *** sxc731 has quit IRC | 23:32 | |
| gunix | cloudnull: that code is genius, cause it seems ansible knows what you are using and configure cinder to use the appropriate driver | 23:33 |
| cloudnull | lol | 23:33 |
| cloudnull | https://docs.openstack.org/openstack-ansible-os_cinder/latest/configure-cinder.html#using-ceph-for-cinder-backups | 23:33 |
| Bico_Fino | cloudnull setup-openstack went fine | 23:33 |
| cloudnull | Bico_Fino: nice! | 23:33 |
| gunix | Bico_Fino: your first setup? | 23:33 |
| Bico_Fino | gunix nope. I played around with Ocata release | 23:34 |
| Bico_Fino | gunix http://bicofino.io/2017/09/05/openstack-ocata-deployment-with-openstack-ansible-and-ceph/ | 23:34 |
| gunix | Bico_Fino: in ocata i did only some manual installs. since i found out about OSA, i quit gaming. i am doing deplopys all day | 23:34 |
| cloudnull | gunix: https://github.com/openstack/openstack-ansible-os_cinder/blob/master/defaults/main.yml#L162-L164 | 23:34 |
| cloudnull | looks like you still have to set the driver from the role. | 23:35 |
| cloudnull | looking to see if we have some glue that automatically defines that | 23:35 |
| Bico_Fino | cloudnull trying to add the ovs-vsctl add-br br-provider | 23:35 |
| *** agrebennikov has quit IRC | 23:35 | |
| Bico_Fino | gunix I didn't tried manual installs yet, only Kolla and openstack-ansible | 23:36 |
| cloudnull | gunix: you'll need to set "cinder_service_backup_driver: cinder.backup.drivers.ceph" in your user_variables.yml file and it'll go | 23:36 |
| gunix | Bico_Fino: i like it that you have a blog. that's cool. however, my 2 cents: if you have enough experience to create blogs, try to work on the official openstack documentation and provide updates to it. it doesn't help a lot to have documentation fragmented in the community. | 23:36 |
| SamYaple | vroom vroom go | 23:36 |
| gunix | cloudnull: yea and i guess also ceph user | 23:37 |
| cloudnull | Bico_Fino: you using ghost ? | 23:38 |
| gunix | Bico_Fino: think about archlinux: that distro was nothing some years ago, but know they have a docu that makes it cooler than deb ubu centos and suse... a good wiki means a lot | 23:38 |
| gunix | *but NOW | 23:38 |
| * cloudnull reads the arch wiki all the time | 23:38 | |
| cloudnull | Bico_Fino: I use ghost, i quite like it | 23:39 |
| cloudnull | https://cloudnull.io/ | 23:39 |
| cloudnull | gunix: It's nice to blog about experiences, kinda get ideas out there. | 23:39 |
| cloudnull | ohai SamYaple | 23:39 |
| gunix | cloudnull: i used to read it a lot but i switched by desktop from archlinux back to debian. i am sorry, but debian is so stable that it makes life forget PC problems exist. i also like the upgrade philosophy ( i hate the fact ubuntu and arch need to get rebooted every week) | 23:40 |
| SamYaple | o7 | 23:40 |
| cloudnull | haha. I run openSUSE tumbleweed :P | 23:40 |
| SamYaple | do you really? | 23:40 |
| * cloudnull still reads the arch wiki | 23:40 | |
| cloudnull | SamYaple: I do | 23:40 |
| * cloudnull is a crazy person | 23:40 | |
| SamYaple | i shouldnt be suprised and yet i still am | 23:40 |
| cloudnull | I have it on my x1 | 23:40 |
| cloudnull | its great | 23:41 |
| gunix | cloudnull: i am 100% for blogging. it is cool. but if someone has good experience with something, working on the main wiki pe that software will help the community . that is my only point. arcticles could also be pasted from blog to community webite | 23:41 |
| gunix | *website | 23:41 |
| cloudnull | SamYaple: http://paste.openstack.org/show/625398/ | 23:41 |
| SamYaple | cloudnull: i get it, i run a fork of pfsense with a custom bsd kernel to break 802.1d rfc | 23:41 |
| cloudnull | :P | 23:41 |
| SamYaple | nice | 23:41 |
| cloudnull | its like arch, but with testing before a package is released | 23:42 |
| cloudnull | so it breaks every other day, instead of every day | 23:42 |
| SamYaple | thats crazy talk | 23:42 |
| gunix | yea suse tumbleweed is really cool but it's not as cool as archlinux because archlinux has a cool name. sorry, | 23:42 |
| SamYaple | i run arch, this install is up to a year without issues :) | 23:42 |
| SamYaple | i update when i remember | 23:42 |
| cloudnull | gunix: you're right | 23:42 |
| gunix | SamYaple: on archlinux i had issues with the steam tray. it was really annoying | 23:43 |
| SamYaple | gunix: yea i run a super stripped down eeverything | 23:43 |
| SamYaple | just switched to i3 from fluxbox a few months back | 23:43 |
| SamYaple | been on fluxbox for 10 years | 23:43 |
| gunix | i don't know if steam works on i3 :D | 23:43 |
| gunix | i am on gnome 3 | 23:44 |
| gunix | i use the desktop also for gaming and i hate dual booting windows. well, i hate windows. | 23:44 |
| cloudnull | gunix: I typically blog about a thing, and if i'm not being 100% insane (which takes a few days to get sorted) I'll contribute it back to the communities/products I'm blogging about. | 23:44 |
| cloudnull | It's fun . | 23:45 |
| cloudnull | hahaha. I run KDE :) | 23:45 |
| cloudnull | steam works great! | 23:45 |
| gunix | cloudnull: i think i have only one article ... this one: https://help.ubuntu.com/community/HALAMP | 23:45 |
| cloudnull | nice! | 23:45 |
| gunix | i was practicing this for openstack. tbh. i want to know how it works in the background | 23:46 |
| gunix | and while doing it i saw that docu online was horrible so i wrote that | 23:46 |
| gunix | i think i should write a small openstack docu for that ubuntu wiki | 23:47 |
| gunix | with some general info, like what each service does, and some links for quick start, like devstack and osa aio | 23:48 |
| cloudnull | Bico_Fino: If you are running ghost and want to have syntax highlighting on your posts https://gist.github.com/cloudnull/2e1ae0ccbd21209e89600af3f8024a4b -- I added that to the code-injection page and viola, syntax highlighting :) | 23:49 |
| gunix | what is gist? | 23:51 |
| Bico_Fino | cloudnull I need migrate to hugo. :P | 23:51 |
| cloudnull | its like pastebin | 23:51 |
| cloudnull | Bico_Fino: I just upgraded my blog to ghost 1.x | 23:52 |
| cloudnull | it was a pain | 23:52 |
| cloudnull | but ghost-cli is nic e | 23:52 |
| cloudnull | and upgrades since have been simple | 23:52 |
| cloudnull | I used to use pelican, but it was too much work | 23:53 |
| Bico_Fino | I really need to upgrade it | 23:53 |
| Bico_Fino | I used octopress | 23:54 |
| gunix | you guys have monitoring via ansible for openstack? | 23:54 |
| gunix | https://github.com/cloudnull/rpc-maas | 23:54 |
| gunix | sorry for stalking your github cloudnull | 23:54 |
| cloudnull | stalk away | 23:54 |
| cloudnull | I also throw things in https://github.com/os-cloud/ | 23:54 |
| cloudnull | the updated maas is here https://github.com/rcbops/rpc-maas | 23:55 |
| cloudnull | its deploys the monitors but all of that plugs into rackspace monitoring as a service | 23:55 |
| gunix | when i see your github i feel like i wasted my life | 23:55 |
| cloudnull | HOWEVER, i recently implemented a backend for all of the monitoring as a service checks to feed data into influxdb | 23:56 |
| openstackgerrit | Hengqing Hu proposed openstack/openstack-ansible-os_octavia stable/pike: Enable V2 Octavia API (Experimental) https://review.openstack.org/515698 | 23:56 |
| cloudnull | which can be visualized with grafana | 23:56 |
| gunix | so how do you monitor prod? ansible checks? or some zabbix nagios blabla? | 23:56 |
| cloudnull | the checks are a good reference but if your not using rackspace monitoring as a service they | 23:56 |
| cloudnull | **they're not super helpful | 23:56 |
| cloudnull | the checks feed all of their data back to a hosted monitoring service, kinda like newrelic | 23:57 |
| gunix | i was thinking about getting zabbix for monitoring | 23:57 |
| gunix | but it seems you have that part also figured out | 23:57 |
| cloudnull | there are a couple folks that use zabbix around here. | 23:58 |
| gunix | is it foss? | 23:58 |
| cloudnull | there's also https://github.com/openstack/openstack-ansible-ops/tree/master/cluster_metrics | 23:58 |
| cloudnull | rax-maas is not foss, at least I don't think it is... | 23:58 |
Generated by irclog2html.py 2.15.3 by Marius Gedminas - find it at mg.pov.lt!