| *** fusmu has quit IRC | 00:01 | |
| openstackgerrit | Merged openstack/openstack-ansible-lxc_container_create master: Always create containers with fixed MAC addresses https://review.openstack.org/550160 | 00:07 |
|---|---|---|
| *** masuberu has quit IRC | 00:35 | |
| *** masuberu has joined #openstack-ansible | 00:37 | |
| *** chyka has quit IRC | 00:38 | |
| logan- | pabelanger: the connection plugin just looks for a few vars in a host's inventory to enable the container namespace stuff... | 00:41 |
| logan- | on the container's inventory host, a physical_host var should point to the inventory_hostname of the bare metal host where the container is located | 00:42 |
| logan- | and the container_name var should contain the name of the container on that host | 00:42 |
| logan- | so usually it is sufficient to set container_name: "{{ inventory_hostname }}" | 00:42 |
| logan- | actually I think that might be defaulted anyway | 00:42 |
| logan- | if you run the task with -vvvvv you'll see some output from the connection plugin like "container confirmed" .. etc showing that it has successfully determined from your inventory that it is working against a container | 00:44 |
| *** kstev1 has quit IRC | 00:49 | |
| pabelanger | cool | 00:49 |
| pabelanger | like I said, totally want to try it out | 00:50 |
| *** lbragstad has quit IRC | 01:27 | |
| *** dave-mccowan has joined #openstack-ansible | 01:56 | |
| *** admin0 has joined #openstack-ansible | 02:08 | |
| *** admin0 has quit IRC | 02:13 | |
| *** markvoelker has quit IRC | 02:25 | |
| *** markvoelker has joined #openstack-ansible | 02:26 | |
| *** masuberu has quit IRC | 02:42 | |
| *** markvoelker has quit IRC | 02:53 | |
| *** markvoelker has joined #openstack-ansible | 02:53 | |
| *** masber has joined #openstack-ansible | 03:02 | |
| *** nurdie has quit IRC | 03:03 | |
| *** nurdie has joined #openstack-ansible | 03:03 | |
| yifei | hwoarang: ping | 03:20 |
| *** masber has quit IRC | 03:36 | |
| *** masber has joined #openstack-ansible | 03:47 | |
| openstackgerrit | Kevin Carter (cloudnull) proposed openstack/openstack-ansible-lxc_container_create master: Flush entire interface rather than just routes https://review.openstack.org/550269 | 04:02 |
| *** dave-mccowan has quit IRC | 04:08 | |
| *** masuberu has joined #openstack-ansible | 04:36 | |
| *** ANKITA has joined #openstack-ansible | 04:40 | |
| *** masber has quit IRC | 04:40 | |
| *** masber has joined #openstack-ansible | 04:40 | |
| *** masuberu has quit IRC | 04:42 | |
| *** radeks has joined #openstack-ansible | 04:47 | |
| openstackgerrit | Merged openstack/openstack-ansible-lxc_container_create master: Flush entire interface rather than just routes https://review.openstack.org/550269 | 04:48 |
| nsingh | while installing my playbook i am getting below error: | 05:05 |
| nsingh | fatal: [aio1]: FAILED! => {"changed": false, "connection": "close", "content": "<html>\r\n<head><title>404 Not Found</title></head>\r\n<body bgcolor=\"white\">\r\n<center><h1>404 Not Found</h1></center>\r\n<hr><center>nginx/1.10.3 (Ubuntu)</center>\r\n</body>\r\n</html>\r\n", "content_length": "178", "content_type": "text/html", "date": "Wed, 07 Mar 2018 05:04:15 GMT", "msg": "Status code was not [200]: HTTP Error 404: Not Found", "redirected": | 05:05 |
| nsingh | false, "server": "nginx/1.10.3 (Ubuntu)", "status": 404, "url": "http://172.29.236.100:8181/venvs/17.0.0/ubuntu-16.04-x86_64/masakari-17.0.0-x86_64.checksum" | 05:05 |
| nsingh | any help? | 05:06 |
| *** poopcat has quit IRC | 05:06 | |
| openstackgerrit | Kevin Carter (cloudnull) proposed openstack/openstack-ansible-plugins master: Resolve nsenter fork exception https://review.openstack.org/550331 | 05:06 |
| cloudnull | nsingh: where are you getting that error ? | 05:06 |
| cloudnull | is this in the aio create while running tempest ? | 05:07 |
| nsingh | cloudnull: i am getting this for the TASK [os_masakari : Retrieve checksum for venv download] | 05:08 |
| cloudnull | nsingh: I don't see that role https://github.com/openstack/openstack-ansible-os_masakari something you're working on ? | 05:09 |
| nsingh | i am creating POC for this only. | 05:10 |
| cloudnull | cool | 05:10 |
| *** udesale has joined #openstack-ansible | 05:10 | |
| cloudnull | the error in that context is likely trying to download a pre-built venv | 05:10 |
| cloudnull | assuming you're using the pattern we have in other os_ roles. | 05:11 |
| nsingh | yes | 05:11 |
| cloudnull | if so, you should be able to enable developer_mode | 05:11 |
| nsingh | if my role is on local machine how should i deal with this. | 05:12 |
| cloudnull | should look something like https://github.com/openstack/openstack-ansible-os_glance/blob/master/defaults/main.yml#L26 | 05:12 |
| cloudnull | which will bypass the venv lookup | 05:12 |
| cloudnull | and build everything on the target | 05:12 |
| nsingh | ok | 05:13 |
| cloudnull | your local role should have something lke https://github.com/openstack/openstack-ansible-os_glance/blob/master/tasks/glance_install.yml#L27-L47 | 05:15 |
| cloudnull | for developer mode | 05:15 |
| cloudnull | and https://github.com/openstack/openstack-ansible-os_glance/blob/master/tasks/glance_install.yml#L87-L104 | 05:15 |
| cloudnull | those same bits are most, if not all, of our os_ roles | 05:16 |
| nsingh | ANKITA: i am able to ping 172.29.236.100 | 05:25 |
| ANKITA | ? | 05:25 |
| *** chyka has joined #openstack-ansible | 05:26 | |
| cloudnull | ok, im off. take care all. | 05:27 |
| *** chyka has quit IRC | 05:30 | |
| *** masuberu has joined #openstack-ansible | 05:32 | |
| *** ThomasS has quit IRC | 05:33 | |
| *** masber has quit IRC | 05:35 | |
| *** dariko has joined #openstack-ansible | 05:45 | |
| *** dariko has quit IRC | 05:56 | |
| *** chyka has joined #openstack-ansible | 06:11 | |
| *** chyka has quit IRC | 06:11 | |
| *** chyka has joined #openstack-ansible | 06:12 | |
| *** chyka_ has joined #openstack-ansible | 06:15 | |
| *** chyka has quit IRC | 06:18 | |
| *** chyka_ has quit IRC | 06:19 | |
| *** chyka has joined #openstack-ansible | 06:19 | |
| *** fusmu has joined #openstack-ansible | 06:20 | |
| *** masuberu has quit IRC | 06:38 | |
| *** jamesdenton has quit IRC | 06:44 | |
| *** chyka has quit IRC | 06:46 | |
| *** ThomasS has joined #openstack-ansible | 06:47 | |
| *** Sha0000 has joined #openstack-ansible | 06:49 | |
| *** Sha0000 has quit IRC | 06:50 | |
| *** prometheanfire has quit IRC | 06:56 | |
| cloudnull | it looks like somethings gone wrong here https://github.com/openstack/openstack-ansible/blob/master/tests/roles/bootstrap-host/tasks/install_packages_debian.yml#L23-L64 | 06:57 |
| *** gkadam has joined #openstack-ansible | 06:58 | |
| cloudnull | seeing it generate odd urls http://logs.openstack.org/29/548929/2/check/openstack-ansible-deploy-aio_lxc-ubuntu-xenial/3f5b670/job-output.txt.gz#_2018-03-07_05_28_14_184586 | 06:58 |
| cloudnull | IE: 'deb :///ubuntu-cloud-archive xenial-updates/pike main' | 06:59 |
| cloudnull | the commands are working locally | 07:01 |
| cloudnull | but i suspect the infra repos are now different? | 07:01 |
| cloudnull | specifically https://github.com/openstack/openstack-ansible/blob/master/tests/roles/bootstrap-host/tasks/install_packages_debian.yml#L44 | 07:05 |
| cloudnull | which is setting https://github.com/openstack/openstack-ansible/blob/master/tests/roles/bootstrap-host/tasks/prepare_aio_config.yml#L68 | 07:06 |
| cloudnull | and setting ":///ubuntu-cloud-archive" | 07:06 |
| *** threestrands_ has joined #openstack-ansible | 07:09 | |
| *** threestrands_ has quit IRC | 07:09 | |
| *** threestrands_ has joined #openstack-ansible | 07:09 | |
| *** threestrands has quit IRC | 07:10 | |
| *** ivve has quit IRC | 07:12 | |
| openstackgerrit | Russell Tweed proposed openstack/ansible-hardening master: Add check to ClamAV task to prevent simeltaneous content updates https://review.openstack.org/550268 | 07:13 |
| *** threestrands_ has quit IRC | 07:13 | |
| cloudnull | looks like the commands are parsing the file correctly | 07:23 |
| cloudnull | accordingly to the repo file in the logs | 07:24 |
| cloudnull | http://logs.openstack.org/29/548929/2/check/openstack-ansible-deploy-aio_lxc-ubuntu-xenial/3f5b670/logs/openstack/instance-info/host_repo_info_05-27-41.log.txt.gz | 07:24 |
| openstackgerrit | Russell Tweed proposed openstack/ansible-hardening master: Add check to ClamAV task to prevent simeltaneous content updates https://review.openstack.org/550268 | 07:28 |
| *** Sha0000 has joined #openstack-ansible | 07:32 | |
| *** Sha0000 has quit IRC | 07:33 | |
| Taseer | cloudnull: looks like something fishy is happening with zuul | 07:34 |
| Taseer | I got all my jobs fail in under 10 minutes | 07:35 |
| *** admin0 has joined #openstack-ansible | 07:39 | |
| *** holser__ has joined #openstack-ansible | 07:44 | |
| *** armaan has joined #openstack-ansible | 07:51 | |
| *** pcaruana has joined #openstack-ansible | 07:54 | |
| *** pcaruana has quit IRC | 07:55 | |
| *** pcaruana has joined #openstack-ansible | 07:55 | |
| odyssey4me | Taseer yep, that seems to be a thing - adding the UCA repo crashes it when trying to apt-get update | 08:00 |
| *** epalper has joined #openstack-ansible | 08:02 | |
| *** gus has quit IRC | 08:02 | |
| PTO | Is it correct understood, that line 81 in https://github.com/openstack/openstack-ansible-os_neutron/blob/master/templates/neutron.conf.j2#L81 will be populated with the first defined provider networkin in the openstack_user_variables.yml? | 08:04 |
| *** gus has joined #openstack-ansible | 08:04 | |
| *** olivierbourdon3- has quit IRC | 08:05 | |
| *** chyka has joined #openstack-ansible | 08:08 | |
| *** olivierbourdon38 has joined #openstack-ansible | 08:09 | |
| *** chyka has quit IRC | 08:13 | |
| *** ivve has joined #openstack-ansible | 08:14 | |
| *** admin0 has quit IRC | 08:15 | |
| odyssey4me | ok, it looks like infra has moved from using a script to an ansible toolset, and broken our implementation of how we read the mirrors | 08:16 |
| odyssey4me | I'll work out some patches to get the gates going again later today | 08:16 |
| *** threestrands has joined #openstack-ansible | 08:32 | |
| *** threestrands has quit IRC | 08:32 | |
| *** threestrands has joined #openstack-ansible | 08:32 | |
| *** shardy has joined #openstack-ansible | 08:34 | |
| *** threestrands has quit IRC | 08:36 | |
| *** ANKITA has quit IRC | 08:43 | |
| *** electrofelix has joined #openstack-ansible | 08:47 | |
| openstackgerrit | Russell Tweed proposed openstack/ansible-hardening master: Add check to ClamAV task to prevent simeltaneous content updates https://review.openstack.org/550268 | 08:49 |
| *** mrch_ has joined #openstack-ansible | 09:10 | |
| *** mbuil has joined #openstack-ansible | 09:12 | |
| *** mrch_ has quit IRC | 09:15 | |
| openstackgerrit | Russell Tweed proposed openstack/ansible-hardening master: Add check to ClamAV task to prevent simeltaneous content updates https://review.openstack.org/550268 | 09:18 |
| *** olivierbourdon38 has quit IRC | 09:20 | |
| *** armaan has quit IRC | 09:24 | |
| hwoarang | good morning | 09:24 |
| *** armaan has joined #openstack-ansible | 09:25 | |
| *** stuartgr has joined #openstack-ansible | 09:26 | |
| ioni | hello guys | 09:28 |
| ioni | is there a setting that i need to enable in order to ansible to upgrade all packages inside the container? | 09:28 |
| *** PTO has quit IRC | 09:31 | |
| *** stuartgr has left #openstack-ansible | 09:32 | |
| *** olivierbourdon38 has joined #openstack-ansible | 09:34 | |
| *** olivierbourdon38 has quit IRC | 09:35 | |
| openstackgerrit | Russell Tweed proposed openstack/ansible-hardening master: Add check to ClamAV task to prevent simeltaneous content updates https://review.openstack.org/550268 | 09:36 |
| *** ANKITA has joined #openstack-ansible | 09:38 | |
| *** armaan has quit IRC | 09:38 | |
| *** armaan has joined #openstack-ansible | 09:38 | |
| *** olivierbourdon38 has joined #openstack-ansible | 09:47 | |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible master: [TEST] Check mirror info https://review.openstack.org/550416 | 09:57 |
| hwoarang | the gates do not look good today | 09:59 |
| odyssey4me | hwoarang yep, something changed in the way the apt sources are set - I'm on it | 09:59 |
| hwoarang | i also see failures outside of our domain | 10:00 |
| hwoarang | build-openstack-sphinx-docsRETRY_LIMIT in 39s | 10:00 |
| hwoarang | etc etc | 10:00 |
| *** holser___ has joined #openstack-ansible | 10:07 | |
| *** holser__ has quit IRC | 10:07 | |
| *** hw_wutianwei has quit IRC | 10:08 | |
| *** hw_wutianwei has joined #openstack-ansible | 10:08 | |
| *** kukacz has quit IRC | 10:17 | |
| *** hw_wutianwei has quit IRC | 10:23 | |
| *** lvdombrkr has joined #openstack-ansible | 10:25 | |
| *** hw_wutianwei has joined #openstack-ansible | 10:31 | |
| *** Sha0000 has joined #openstack-ansible | 10:32 | |
| openstackgerrit | Russell Tweed proposed openstack/openstack-ansible-os_keystone master: Add memcache setup to shibboleth2.xml https://review.openstack.org/550427 | 10:33 |
| *** Sha0000 has quit IRC | 10:43 | |
| *** mbuil has quit IRC | 10:43 | |
| *** gunix has joined #openstack-ansible | 10:54 | |
| gunix | hello guys | 10:54 |
| gunix | i installed OSA-AIO and after that downloaded the RC file from horizon but i'm getting an error when i try to use the CLI | 10:54 |
| gunix | when i do hypervisor list i get HTTPSConnectionPool(host='10.40.2.142', port=5000): Max retries exceeded with url: /v3/auth/tokens (Caused by SSLError(SSLError("bad handshake: Error([('SSL routines', 'tls_process_server_certificate', 'certificate verify failed')],)",),)) | 10:54 |
| gunix | am i missing something ? do i need to add another parameter to the rc file for ssl ingore self cert ? | 10:55 |
| *** ivve has quit IRC | 11:02 | |
| *** jwitko has quit IRC | 11:04 | |
| *** udesale_ has joined #openstack-ansible | 11:19 | |
| *** shardy has quit IRC | 11:20 | |
| *** ivve has joined #openstack-ansible | 11:21 | |
| *** shardy has joined #openstack-ansible | 11:21 | |
| *** udesale has quit IRC | 11:21 | |
| *** shardy has quit IRC | 11:21 | |
| *** udesale has joined #openstack-ansible | 11:21 | |
| *** shardy has joined #openstack-ansible | 11:22 | |
| *** udesale_ has quit IRC | 11:24 | |
| *** epalper has quit IRC | 11:25 | |
| odyssey4me | hwoarang back from a meeting - and yeah, I see that all jobs are hitting a retry limit | 11:29 |
| *** ANKITA has quit IRC | 11:30 | |
| *** udesale has quit IRC | 11:36 | |
| hwoarang | ok so we are forced to take a break today :) goodie | 11:44 |
| *** holser___ has quit IRC | 11:48 | |
| *** holser__ has joined #openstack-ansible | 11:48 | |
| *** Sha0000 has joined #openstack-ansible | 11:50 | |
| *** ivve has quit IRC | 12:00 | |
| *** Sha0000 has quit IRC | 12:10 | |
| *** karthi has joined #openstack-ansible | 12:12 | |
| *** gameon has joined #openstack-ansible | 12:16 | |
| *** speedmann has quit IRC | 12:18 | |
| *** speedmann has joined #openstack-ansible | 12:19 | |
| *** yolanda has quit IRC | 12:20 | |
| *** yolanda has joined #openstack-ansible | 12:21 | |
| *** mattoliverau has quit IRC | 12:21 | |
| *** mattoliverau has joined #openstack-ansible | 12:22 | |
| *** jamesdenton has joined #openstack-ansible | 12:25 | |
| *** armaan has quit IRC | 12:26 | |
| *** armaan has joined #openstack-ansible | 12:26 | |
| *** dave-mccowan has joined #openstack-ansible | 12:32 | |
| *** udesale has joined #openstack-ansible | 12:39 | |
| *** lbragstad has joined #openstack-ansible | 12:39 | |
| *** gameon has quit IRC | 12:43 | |
| *** chyka has joined #openstack-ansible | 12:46 | |
| *** karthi has quit IRC | 12:48 | |
| *** chyka has quit IRC | 12:50 | |
| *** armaan has quit IRC | 12:55 | |
| *** armaan has joined #openstack-ansible | 12:56 | |
| *** admin0 has joined #openstack-ansible | 12:56 | |
| *** admin0 has quit IRC | 12:57 | |
| mhayden | morning | 12:57 |
| *** karthi has joined #openstack-ansible | 13:09 | |
| *** Matias has quit IRC | 13:13 | |
| *** prometheanfire has joined #openstack-ansible | 13:20 | |
| *** karthi has quit IRC | 13:24 | |
| mhayden | odyssey4me / hwoarang: are these gate shenanigans something we can fix? or are they infra problems? | 13:24 |
| *** Matias has joined #openstack-ansible | 13:25 | |
| mhayden | oh nvm, i see in the jobs | 13:26 |
| mhayden | looks like the fix merged: https://review.openstack.org/#/c/550425/ | 13:29 |
| *** Matias has quit IRC | 13:29 | |
| *** Matias has joined #openstack-ansible | 13:30 | |
| openstackgerrit | Major Hayden proposed openstack/ansible-hardening master: Replace Fedora 26 with 27 https://review.openstack.org/550109 | 13:30 |
| *** ivve has joined #openstack-ansible | 13:34 | |
| *** Matias has quit IRC | 13:35 | |
| openstackgerrit | Andy McCrae proposed openstack/openstack-ansible master: Fix BOOTSTRAP_OPTS https://review.openstack.org/550473 | 13:35 |
| mhayden | yeah jobs seem to be running okay now | 13:38 |
| mhayden | odyssey4me: if you get time, could you review my idempotency patch for os-tests? https://review.openstack.org/550135 | 13:39 |
| *** Matias has joined #openstack-ansible | 13:39 | |
| openstackgerrit | Major Hayden proposed openstack/openstack-ansible-openstack_hosts master: Remove kernel modules: scsi_dh + x_tables https://review.openstack.org/549754 | 13:39 |
| openstackgerrit | Major Hayden proposed openstack/openstack-ansible-openstack_hosts master: Make role idempotent https://review.openstack.org/550102 | 13:39 |
| openstackgerrit | Andy McCrae proposed openstack/openstack-ansible master: Avoid putting rgw conf on ALL hosts https://review.openstack.org/549840 | 13:40 |
| *** woodard has quit IRC | 13:41 | |
| *** woodard has joined #openstack-ansible | 13:41 | |
| openstackgerrit | Major Hayden proposed openstack/openstack-ansible master: Use RDO Queens repo https://review.openstack.org/548641 | 13:42 |
| openstackgerrit | Major Hayden proposed openstack/openstack-ansible master: Use ARA instead of profile_tasks callback https://review.openstack.org/546271 | 13:42 |
| mhayden | hwoarang / xgerman_: if y'all get a chance to get https://review.openstack.org/549036 settled, that would be good :) | 13:43 |
| *** Matias has quit IRC | 13:46 | |
| *** esberglu has joined #openstack-ansible | 13:48 | |
| *** Matias has joined #openstack-ansible | 13:48 | |
| *** Matias has quit IRC | 13:53 | |
| *** jwitko_ has joined #openstack-ansible | 13:55 | |
| *** Matias has joined #openstack-ansible | 14:00 | |
| *** karthi has joined #openstack-ansible | 14:03 | |
| *** SerenaFeng has joined #openstack-ansible | 14:04 | |
| *** Matias has quit IRC | 14:06 | |
| hwoarang | mhayden: i am still wondering about the dib pin to 2.9.0 | 14:06 |
| hwoarang | do you know the reason? | 14:07 |
| hwoarang | 2.11.0 or any new release should work right? | 14:07 |
| *** karthi has quit IRC | 14:08 | |
| *** ivve has quit IRC | 14:09 | |
| *** esberglu has quit IRC | 14:14 | |
| *** lbragstad has quit IRC | 14:15 | |
| *** armaan has quit IRC | 14:16 | |
| *** Matias has joined #openstack-ansible | 14:18 | |
| *** armaan has joined #openstack-ansible | 14:20 | |
| openstackgerrit | Merged openstack/ansible-hardening master: Replace Fedora 26 with 27 https://review.openstack.org/550109 | 14:21 |
| *** lbragstad has joined #openstack-ansible | 14:21 | |
| *** ivve has joined #openstack-ansible | 14:22 | |
| *** armaan has quit IRC | 14:22 | |
| *** sar has joined #openstack-ansible | 14:23 | |
| *** Matias has quit IRC | 14:24 | |
| *** pylot has quit IRC | 14:27 | |
| jrosser | surprised at how complicated this is http://git.openstack.org/cgit/openstack/openstack-ansible-os_keystone/tree/tasks/keystone_db_setup.yml#n49 | 14:27 |
| jrosser | something like this can handle non existing services https://www.irccloud.com/pastebin/UVcGwTBh/ | 14:28 |
| *** throwsb1 has joined #openstack-ansible | 14:28 | |
| *** ivve has quit IRC | 14:29 | |
| *** esberglu has joined #openstack-ansible | 14:29 | |
| openstackgerrit | Major Hayden proposed openstack/openstack-ansible-openstack_hosts master: Make role idempotent https://review.openstack.org/550102 | 14:33 |
| *** chyka has joined #openstack-ansible | 14:37 | |
| *** Matias has joined #openstack-ansible | 14:41 | |
| *** chyka has quit IRC | 14:41 | |
| mhayden | hwoarang: no idea on that one:/ | 14:42 |
| *** ivve has joined #openstack-ansible | 14:43 | |
| *** lvdombrkr has quit IRC | 14:45 | |
| *** Matias has quit IRC | 14:45 | |
| *** electrofelix has quit IRC | 14:51 | |
| *** SerenaFeng has quit IRC | 14:51 | |
| cloudnull | mhayden: odyssey4me: evrardjp: any chance you've had a moment to take a look at http://logs.openstack.org/71/503971/34/check/openstack-ansible-deploy-aio_lxc-ubuntu-xenial/9d48ed4/job-output.txt.gz#_2018-03-07_12_34_04_418557 ? | 14:58 |
| *** mbuil has joined #openstack-ansible | 14:58 | |
| cloudnull | looks like this is still an issue in the gate which is not correctly setting the UCA repos. | 14:58 |
| mhayden | not i | 14:58 |
| *** mbuil has quit IRC | 15:00 | |
| cloudnull | seems to be picking up something like http://logs.openstack.org/29/548929/2/check/openstack-ansible-deploy-aio_lxc-ubuntu-xenial/3582610/job-output.txt.gz#_2018-03-07_03_15_03_849120 in the gate | 15:00 |
| *** Matias has joined #openstack-ansible | 15:01 | |
| *** epalper has joined #openstack-ansible | 15:01 | |
| openstackgerrit | Merged openstack/openstack-ansible-tests master: Add ansible-lint rule for apt_repository https://review.openstack.org/547497 | 15:01 |
| *** pabelanger has quit IRC | 15:04 | |
| ioni | guys | 15:04 |
| ioni | i have a problem with nova install | 15:04 |
| ioni | TASK [Gather software version list] | 15:05 |
| ioni | KeyError: 'ansible_local' | 15:05 |
| ioni | fatal: [ctrl1usf_nova_console_container-610db168]: FAILED! => {"failed": true, "msg": "Unexpected failure during module execution.", "stdout": ""} | 15:05 |
| cloudnull | ioni: are you running with a limit or tags? | 15:05 |
| ioni | https://paste.xinu.at/TMV/ | 15:06 |
| *** Matias has quit IRC | 15:06 | |
| cloudnull | I've seen the local facts not get set when running with a tag / limits | 15:06 |
| ioni | openstack-ansible os-nova-install.yml | 15:06 |
| cloudnull | ok. | 15:06 |
| ioni | i have two controllers that have the same problem | 15:06 |
| *** karthi has joined #openstack-ansible | 15:06 | |
| ioni | i'm running newton latest from yesterday, trying to update to it | 15:06 |
| *** Matias has joined #openstack-ansible | 15:07 | |
| cloudnull | ioni: is it failing anywhere else before it runs into this error? | 15:10 |
| *** kstev has joined #openstack-ansible | 15:10 | |
| cloudnull | the local facts are created / stored on the remote host | 15:11 |
| cloudnull | I'm wondering if there's a pre-condition that's not being met | 15:11 |
| ioni | cloudnull, hmm yes | 15:11 |
| ioni | on a compute node | 15:11 |
| ioni | some dependency problem | 15:12 |
| *** pabelanger has joined #openstack-ansible | 15:12 | |
| ioni | i'll fix it | 15:12 |
| ioni | i think from there is the problem | 15:12 |
| cloudnull | when you checked-out the new release did re-bootstrap ansible running scripts/bootstrap-ansible.sh | 15:12 |
| ioni | yes | 15:12 |
| *** karthi has quit IRC | 15:14 | |
| ioni | mariadb-client : Depends: mariadb-client-10.0 (= 10.0.34+maria-1~xenial) but 10.0.34-0ubuntu0.16.04.1 is installed | 15:14 |
| ioni | that's the problem i had on a lot of components | 15:14 |
| d34dh0r53 | moin | 15:15 |
| cloudnull | ioni: I thought we had a pin in for some versions of mariadb | 15:15 |
| cloudnull | maybe the old repos need to be cleaned out in favor of the new pinned versions ? | 15:16 |
| cloudnull | ioni: can you see if the apt preference pins are in place ? | 15:16 |
| *** radeks has quit IRC | 15:16 | |
| *** radeks has joined #openstack-ansible | 15:17 | |
| ioni | Package: * | 15:17 |
| ioni | Pin: release o=MariaDB | 15:17 |
| ioni | Pin-Priority: 1001 | 15:17 |
| *** epalper has quit IRC | 15:17 | |
| ioni | ? | 15:17 |
| ioni | ytere is | 15:17 |
| ioni | there is | 15:17 |
| ioni | but keep in mind that i have an old version of newton running | 15:17 |
| ioni | 14.2.2 or something like that | 15:17 |
| *** woodard has quit IRC | 15:18 | |
| cloudnull | I wonder if that pin needs to be removed ahead of this upgrade? | 15:19 |
| cloudnull | there was a version pin | 15:19 |
| cloudnull | Im looking for it | 15:19 |
| cloudnull | https://github.com/openstack/openstack-ansible-galera_server/blob/1075fe411aa7b37e6c536e1c1ea5d4bfd509d2b0/defaults/main.yml#L142-L156 | 15:20 |
| cloudnull | ioni: you should see something similar in the pins | 15:20 |
| ioni | nope, i only have the first entry | 15:21 |
| ioni | on compute node | 15:21 |
| *** epalper has joined #openstack-ansible | 15:21 | |
| ioni | also on galera container | 15:21 |
| ioni | https://paste.xinu.at/fjRqJ/ | 15:22 |
| *** Matias has quit IRC | 15:22 | |
| *** sar has quit IRC | 15:23 | |
| ioni | cloudnull, the issue is that i believe newton wants a specific version from a diferent repo rather than the one from xenial | 15:23 |
| ioni | 10.0.34+maria-1~xenial vs 10.0.34-0ubuntu0.16.04.1 | 15:23 |
| ioni | so remove the ubuntu one mariadb-client fixes the issue | 15:24 |
| ioni | for me | 15:24 |
| cloudnull | interesting. | 15:24 |
| cloudnull | so maybe we need a step in our inline upgrades to clear out the pins | 15:24 |
| ioni | root@ctrl1nl-galera-container-76204eb3:~# dpkg -l | grep mariadb-client | 15:24 |
| ioni | ii mariadb-client 10.0.34+maria-1~xenial all MariaDB database client (metapackage depending on the latest version) | 15:24 |
| ioni | ii mariadb-client-10.0 10.0.34+maria-1~xenial amd64 MariaDB database client binaries | 15:24 |
| ioni | ii mariadb-client-core-10.0 10.0.34+maria-1~xenial amd64 MariaDB database core client binaries | 15:24 |
| cloudnull | I wonder, did the deployment started life out before the pins or our repos got put in-place? | 15:27 |
| ioni | i don't know if the pins were there in 14.2.2 | 15:28 |
| *** karthi has joined #openstack-ansible | 15:29 | |
| *** armaan has joined #openstack-ansible | 15:32 | |
| *** udesale has quit IRC | 15:45 | |
| mardim | Hello all | 15:52 |
| mardim | oops wronf channel sorry | 15:53 |
| *** Matias has joined #openstack-ansible | 16:00 | |
| *** Matias has quit IRC | 16:04 | |
| *** Matias has joined #openstack-ansible | 16:07 | |
| *** pcaruana has quit IRC | 16:10 | |
| *** Matias has quit IRC | 16:12 | |
| openstackgerrit | Merged openstack/openstack-ansible-os_octavia stable/pike: Adds some skip linters to the tests https://review.openstack.org/549036 | 16:13 |
| mnaser | how can i force a rebuild of a venv? i updated the hash for nova and reran repo-install and it didnt rebuild it | 16:14 |
| openstackgerrit | Merged openstack/openstack-ansible-os_octavia stable/pike: Zuul: Remove project name https://review.openstack.org/541932 | 16:14 |
| openstackgerrit | Merged openstack/openstack-ansible-os_octavia stable/pike: Fix file path https://review.openstack.org/545336 | 16:14 |
| *** Matias has joined #openstack-ansible | 16:16 | |
| cloudnull | mnaser: you can run repo-build.yml | 16:16 |
| cloudnull | or you can enable developer mode on a given role | 16:16 |
| mnaser | had i ran repo-install | 16:18 |
| mnaser | my bad | 16:18 |
| mnaser | s/had/ah/ | 16:18 |
| cloudnull | there's also a flag to force rebuild everything if you need too | 16:19 |
| mnaser | cloudnull: i just changed a hash so i'm not sure it would recognize that the hash changed and rebuilds that specific venv | 16:19 |
| cloudnull | it should see the difference and rebuild the one venv. however `-e 'repo_build_venv_rebuild=true'` will force the rebuild | 16:21 |
| *** Matias has quit IRC | 16:31 | |
| *** Leo_m has joined #openstack-ansible | 16:32 | |
| *** woodard has joined #openstack-ansible | 16:32 | |
| *** woodard has quit IRC | 16:32 | |
| *** woodard has joined #openstack-ansible | 16:40 | |
| *** olivierbourdon38 has quit IRC | 16:40 | |
| mnaser | cloudnull: alright, i added "nova_git_install_branch: 0b7f8064df75d01370bb4405547f216b004d90da" and ran .. openstack-ansible repo-build.yml -e 'repo_build_venv_rebuild=true' .. and nova is not being rebuilt | 16:45 |
| mnaser | is there something else to use because it looks like the wheel is not being rebuilt | 16:45 |
| *** epalper has quit IRC | 16:45 | |
| *** fusmu has quit IRC | 16:45 | |
| andymccr | jmccrory: do you remember why we didn't adopt the upstream os_user/os_project modules instead of the keystone: one we keep in OSA? | 16:47 |
| *** karthi has quit IRC | 16:47 | |
| logan- | mnaser: delete /etc/ansible/facts.d/openstack_ansible.fact on the build container | 16:49 |
| logan- | or at least set need_wheel_build = True | 16:50 |
| cloudnull | anyone mind giving https://review.openstack.org/#/c/550331/ a review | 16:50 |
| cloudnull | should resolve the fork issue we're seeing randomely | 16:51 |
| * cloudnull still needs to look into what's causing the UCA breakage in the gate. | 16:51 | |
| cloudnull | local testing shows is passing, but the gate in master seems is very unhappy. | 16:51 |
| mnaser | logan-: i think that might do the trick, thank you | 16:51 |
| mnaser | cloudnull: can i help by another pair of eyes on logs? | 16:52 |
| cloudnull | sure. | 16:52 |
| cloudnull | http://logs.openstack.org/31/550331/1/experimental/openstack-ansible-integrated-deploy-aio/ac17d83/ | 16:52 |
| cloudnull | is a recent set I was looking at | 16:52 |
| mnaser | ok: [aio1] => (item={'state': 'absent', 'repo': 'deb :///ubuntu-cloud-archive xenial-updates/queens main', 'filename': 'ubuntu_cloud_archive_canonical_com_ubuntu'}) | 16:52 |
| mnaser | that cant be right | 16:52 |
| cloudnull | http://logs.openstack.org/31/550331/1/experimental/openstack-ansible-integrated-deploy-aio/ac17d83/job-output.txt.gz#_2018-03-07_07_28_33_720101 | 16:53 |
| cloudnull | yup | 16:53 |
| cloudnull | so we pull the base repo from the hsot | 16:53 |
| cloudnull | **host | 16:53 |
| cloudnull | and then pipe that into a fact for future use when setting the UCA repo. | 16:53 |
| mnaser | https://github.com/openstack/openstack-ansible-openstack_hosts/blob/master/tasks/openstack_hosts_configure_apt.yml#L44-L60 | 16:53 |
| cloudnull | this is an exchange that only exists in the gate | 16:53 |
| *** Matias has joined #openstack-ansible | 16:53 | |
| mnaser | which seems to be "_package_repos + (user_external_repos_list | default([]))" | 16:54 |
| mnaser | maybe that is not being set properly | 16:54 |
| cloudnull | for some more context http://eavesdrop.openstack.org/irclogs/%23openstack-ansible/latest.log.html#t2018-03-07T06:57:55 | 16:55 |
| cloudnull | it seems to be reading the sources list and storing the stdout | 16:55 |
| cloudnull | https://github.com/openstack/openstack-ansible/blob/master/tests/roles/bootstrap-host/tasks/prepare_aio_config.yml#L68 | 16:55 |
| mnaser | https://github.com/openstack/openstack-ansible/blob/e95bdb3fd65ba8561821b22f2c7f773869da7860/tests/roles/bootstrap-host/tasks/install_packages_debian.yml#L23-L30 | 16:56 |
| cloudnull | however a local test of the command found here https://github.com/openstack/openstack-ansible/blob/master/tests/roles/bootstrap-host/tasks/install_packages_debian.yml#L24 on the repos logged here http://logs.openstack.org/31/550331/1/experimental/openstack-ansible-integrated-deploy-aio/ac17d83/logs/openstack/instance-info/host_repo_info_07-28-00.log.txt.gz seems to work fine | 16:56 |
| mnaser | thats some serious regex | 16:56 |
| mnaser | :p | 16:56 |
| cloudnull | so something in between is either setting the repos wrong or munging the stdout | 16:57 |
| mnaser | http://logs.openstack.org/31/550331/1/experimental/openstack-ansible-integrated-deploy-aio/ac17d83/logs/ara/result/de09e872-438d-4e6f-9ec2-4ff03d541100/ | 16:57 |
| mnaser | i remember seeing a change adding trusted=yes lately. | 16:57 |
| cloudnull | yea, I'm curious if that' | 16:57 |
| cloudnull | is the source of our problems | 16:58 |
| cloudnull | due to this hack | 16:58 |
| mnaser | it's def that | 16:58 |
| mnaser | because if you look at the ara output | 16:58 |
| mnaser | its adding a weird bunch of strings before | 16:58 |
| mnaser | cloudnull: why dont we use zuul facts or the stuff in configure_mirrors.sh | 16:58 |
| cloudnull | I'm tempted to drop those commands / interactions for nwo | 16:58 |
| cloudnull | **now | 16:58 |
| cloudnull | mnaser: I do not know | 16:58 |
| cloudnull | what | 16:58 |
| cloudnull | ** what's in those files ? | 16:58 |
| cloudnull | maybe we can do this better | 16:59 |
| mnaser | cloudnull: all images in the openstack infra ship a file called configure_mirrors.sh -- when you source that file, it exports a set of env variables which include the URLs to ALL mirrors in openstack infra | 16:59 |
| cloudnull | IE https://github.com/openstack-infra/zuul-jobs/blob/master/roles/configure-mirrors/templates/etc/apt/sources.list.j2 | 16:59 |
| mnaser | cloudnull: nope, beyond that, one moment | 17:00 |
| *** Leo_m_ has joined #openstack-ansible | 17:00 | |
| *** chyka has joined #openstack-ansible | 17:00 | |
| *** Leo_m_ has quit IRC | 17:02 | |
| *** Leo_m has quit IRC | 17:02 | |
| *** Leo_m has joined #openstack-ansible | 17:02 | |
| mnaser | cloudnull: https://github.com/openstack-infra/openstack-zuul-jobs/tree/master/roles/mirror-info | 17:03 |
| mnaser | it templates this file https://github.com/openstack-infra/openstack-zuul-jobs/blob/master/roles/mirror-info/templates/mirror_info.sh.j2 out to /etc/ci/mirror_info.sh | 17:04 |
| mnaser | we can source that file and then do something like lookup('env', 'NODEPOOL_UCA_MIRROR') | default(uca_etc) | 17:05 |
| mnaser | let me push up something quick to try out | 17:05 |
| *** chyka has quit IRC | 17:05 | |
| *** karthi has joined #openstack-ansible | 17:08 | |
| *** chyka has joined #openstack-ansible | 17:11 | |
| *** gkadam has quit IRC | 17:11 | |
| *** karthi has quit IRC | 17:13 | |
| *** armaan has quit IRC | 17:15 | |
| *** armaan has joined #openstack-ansible | 17:15 | |
| openstackgerrit | Shannon Mitchell proposed openstack/openstack-ansible-ops master: Fix broken linuxbridge-agent config for the flat network in osa-ops mnaio https://review.openstack.org/550539 | 17:17 |
| *** SimAloo has joined #openstack-ansible | 17:17 | |
| openstackgerrit | Mohammed Naser proposed openstack/openstack-ansible master: Clean-up gate code to use mirror_info https://review.openstack.org/550542 | 17:20 |
| mnaser | cloudnull: ^ lets see how this will break terribly | 17:20 |
| cloudnull | that all looks good | 17:21 |
| cloudnull | if it gets past the openstack-hosts then https://www.myinstants.com/media/instants_images/boratgs.jpg | 17:22 |
| *** olivierbourdon38 has joined #openstack-ansible | 17:24 | |
| *** olivierbourdon38 has quit IRC | 17:24 | |
| *** fusmu has joined #openstack-ansible | 17:25 | |
| *** olivierbourdon38 has joined #openstack-ansible | 17:26 | |
| jmccrory | andymccr i don't. looks like those two have been around for a long time too | 17:30 |
| idlemind | yum -y update on host indicates priorities protections are excluding packages ... is this to keep a host in lock step? how do i apply any updates? | 17:30 |
| *** fusmu has quit IRC | 17:30 | |
| andymccr | yeah hmm | 17:31 |
| andymccr | ok might look into it | 17:31 |
| andymccr | may aswell ditch them if possible :D | 17:31 |
| *** olivierbourdon3- has joined #openstack-ansible | 17:32 | |
| *** olivierbourdon3- has quit IRC | 17:32 | |
| *** olivierbourdon3- has joined #openstack-ansible | 17:32 | |
| cloudnull | idlemind: is this something you're seeing in the playbooks ? | 17:33 |
| cloudnull | or just something you're wanting to run across the cloud ? | 17:33 |
| idlemind | nah if i log into a host built by osa and try to update it via yum -y update i see: 591 packages excluded due to repository priority protections | 17:33 |
| *** olivierbourdon38 has quit IRC | 17:33 | |
| cloudnull | hum. I' | 17:35 |
| cloudnull | **I'm not sure where those protections are coming from . | 17:35 |
| * cloudnull looking | 17:36 | |
| cloudnull | I wonder if this is something coming from our galera roles ? | 17:36 |
| cloudnull | mhayden: ^ | 17:36 |
| idlemind | priorities.conf in /etc/yum/pluginconf.d | 17:36 |
| spotz | evrardjp: How close are we to EOL? | 17:36 |
| cloudnull | http://paste.openstack.org/show/693689/ | 17:37 |
| cloudnull | idlemind: looks like we' | 17:37 |
| cloudnull | bah... | 17:37 |
| cloudnull | I can't type atm | 17:37 |
| spotz | hehe | 17:37 |
| cloudnull | looks like we're only setting the fastest mirror file | 17:38 |
| idlemind | hmm wonder how i got it | 17:38 |
| cloudnull | is there more priorities being set ? | 17:38 |
| cloudnull | idlemind: what release ? | 17:38 |
| idlemind | stable/pike | 17:38 |
| idlemind | https://github.com/openstack/openstack-ansible-openstack_hosts/blob/59112cbd84d6a8cb3652d780851d109d104692d6/tasks/openstack_hosts_configure_yum.yml | 17:40 |
| idlemind | at the bottom something about repo priorities i'll have to digest | 17:40 |
| *** armaan has quit IRC | 17:47 | |
| *** armaan has joined #openstack-ansible | 17:48 | |
| openstackgerrit | Major Hayden proposed openstack/openstack-ansible-openstack_hosts master: Make role idempotent https://review.openstack.org/550102 | 17:52 |
| openstackgerrit | Major Hayden proposed openstack/openstack-ansible-openstack_hosts master: Fix openstack_host_module_file typo https://review.openstack.org/550549 | 17:52 |
| *** armaan has quit IRC | 17:53 | |
| *** armaan has joined #openstack-ansible | 17:54 | |
| *** sar has joined #openstack-ansible | 17:54 | |
| idlemind | cloudnull alternatively just having it present might be the problem https://github.com/openstack/openstack-ansible-openstack_hosts/blob/45b2e96ec7dcade1e161868857084f9c436d6799/vars/redhat-7.yml#L91-L92 | 17:55 |
| idlemind | further down in there it sets priorities | 17:56 |
| *** pcaruana has joined #openstack-ansible | 17:57 | |
| idlemind | results in http://paste.openstack.org/show/693699/ | 17:57 |
| idlemind | end result seems to be that i don't get updates (although doing it from cli on a host may not be the osa way) | 17:58 |
| idlemind | i have to imagine i'm due some updates this cloud was built at least a month ago so something seems hinky | 18:04 |
| cloudnull | yea I think thats something is off | 18:04 |
| cloudnull | That paste shows quite a few pins | 18:05 |
| cloudnull | but i'm not 100% on why that's happening within cent | 18:05 |
| cloudnull | IMO I' | 18:05 |
| cloudnull | **I'd blow that preference file aways and run teh update | 18:05 |
| idlemind | brb after lunch; i was tinkering w/the priorities in yum.repos.d and yum update and all i get is more packages excluded for updates ... i'll disable it after lunch. | 18:05 |
| cloudnull | allowing the system to replce as needed | 18:05 |
| idlemind | it seems to be getting written for valid reason in the links i pasted | 18:06 |
| cloudnull | though idk if that'd cause more problems than solve | 18:06 |
| idlemind | it's set in the default variables and there is a task in the openstack-hosts | 18:06 |
| idlemind | play | 18:06 |
| idlemind | so somebody thought priorities were a good thing but it seems to break updates at least for me | 18:06 |
| cloudnull | yea I suspect that if its breaking for you it'll break for everyone | 18:09 |
| *** osnaya has joined #openstack-ansible | 18:12 | |
| osnaya | @cloudnull Hi there... Got your reference from Victor P at meetup last evening. | 18:15 |
| cloudnull | o/ | 18:15 |
| osnaya | @cloudnull looking into openstack-ansible to deploy pike in our lab... | 18:16 |
| cloudnull | cool | 18:17 |
| osnaya | @cloudnull have openstack-ansible downloaded and configuring for deployment... want to deploy on Ubuntu16.04 host (not using containers at this stage). deploy it directly on Ubuntu | 18:18 |
| cloudnull | osnaya: yes you can do that | 18:20 |
| cloudnull | evrardjp: has recently made the tests so that's running in the gate | 18:21 |
| osnaya | @cloudnull there is a mention in the doc... that it can be deployed on host (config file needs to be created in /etc.d)..... can you point to any examples on how to bypass the container deployment such as aio_metal.yml | 18:21 |
| cloudnull | osnaya: you need to setup your environment to run all on metal | 18:22 |
| cloudnull | https://github.com/openstack/openstack-ansible/blob/8a9bbaab1d01d05c1053db75658bb7c91ce70689/etc/openstack_deploy/env.d/aio_metal.yml.example | 18:22 |
| cloudnull | if you put that file into the /etc/openstack_deploy/env.d/ location the inventory will not build any containers | 18:23 |
| mnaser | hmm | 18:23 |
| mnaser | it looks like ceph roles broke ansible linter jobs? | 18:23 |
| cloudnull | a new linter was added | 18:23 |
| cloudnull | maybe related ? | 18:23 |
| mnaser | exception: no action detected in task. This often indicates a misspelled module name, or incorrect module path. | 18:23 |
| mnaser | - name: configure crush hierarchy | 18:24 |
| mnaser | in ceph-mon/tasks/crush_rules.yml | 18:24 |
| osnaya | @cloudnull ...looked at that file... which has "container_skel:" section in the example | 18:24 |
| mnaser | looks like 'ceph_crush' module is a thing? | 18:24 |
| mnaser | but i guess it's not being loaded when the linter is running? | 18:24 |
| mnaser | ok so other than the gate being broken, the patch fixed that issue cloudnull | 18:26 |
| cloudnull | woot! | 18:26 |
| cloudnull | its the wheel of progress | 18:26 |
| cloudnull | :) | 18:27 |
| osnaya | @cloudnull .... I want to know for deployment on the host, do we keep "component_skel: and physical_skel:" sections blank for that example or do we have entries in there? | 18:27 |
| *** olivierbourdon3- has quit IRC | 18:27 | |
| *** olivierbourdon38 has joined #openstack-ansible | 18:28 | |
| cloudnull | osnaya: you will need all of those sections | 18:28 |
| cloudnull | the override file is what you'd need to simply instruct the dynamic inventory to not construct the container host entries | 18:29 |
| cloudnull | if you have a generated inventory already you will have to flush it out and recreate it | 18:30 |
| *** olivierbourdon38 has quit IRC | 18:30 | |
| osnaya | @cloudnull I have not generated inventory yet... | 18:30 |
| cloudnull | ok | 18:31 |
| *** olivierbourdon38 has joined #openstack-ansible | 18:31 | |
| *** shardy has quit IRC | 18:31 | |
| cloudnull | so you should just need to put that aio_metal.yml file in place and the rest should "just work"-tm | 18:31 |
| cloudnull | after that file is in place follow the regular deployment document | 18:31 |
| osnaya | @cloudnull question I have is, in my case can I use the "component_skel and physical_skel" section from extra_container_yml.example as it in aio_metal.yml or the entries in these section for deployment on host will be different? | 18:32 |
| *** olivierbourdon38 has quit IRC | 18:32 | |
| *** dariko has joined #openstack-ansible | 18:32 | |
| *** olivierbourdon38 has joined #openstack-ansible | 18:32 | |
| cloudnull | you can have those sections however those shouldn't be needed | 18:32 |
| cloudnull | the core environment files are housed with the openstack-ansible code base | 18:33 |
| osnaya | @cloudnull cool... will give it a try and find out... | 18:33 |
| cloudnull | so the override file will just tickle the core environments to make them all deploy on metal | 18:34 |
| osnaya | @cloudnull when you and docs refer to deploy on metal is actually deploy on Ubuntu host, right? | 18:34 |
| cloudnull | osnaya: let me know how it goes. | 18:34 |
| cloudnull | yes, its to deploy on the baremetal without any containerization | 18:35 |
| cloudnull | you can use ubuntu, suse, or centos | 18:35 |
| osnaya | @cloudnull yes, definitely.... will share the findings :) thank you | 18:35 |
| osnaya | @cloudnull thanks for the pointers... | 18:35 |
| osnaya | exit | 18:44 |
| *** Leo_m has quit IRC | 18:44 | |
| osnaya | quit | 18:47 |
| *** osnaya has quit IRC | 18:47 | |
| goldenfri | Is openvswitch not starting on 17.0.0.0rc1 (AIO) a known bug? | 18:50 |
| idlemind | cloudnull w/priorities disabled i get updates and no more excluded messages not a lot and they all come from epel | 18:55 |
| *** sdake has joined #openstack-ansible | 18:59 | |
| *** sdake has quit IRC | 18:59 | |
| *** sdake has joined #openstack-ansible | 18:59 | |
| cloudnull | goldenfri: not one that im aware of | 19:01 |
| goldenfri | oh ok, I'll create one then :) | 19:01 |
| goldenfri | well a bug report, not an actual bug | 19:01 |
| cloudnull | you could create a bug too if you wanted :) | 19:02 |
| cloudnull | idlemind: so disabling the priorities across the cluster was needed to get updates flowing again? | 19:03 |
| idlemind | cloudnull i'm not positive; it was a small amount of updates (6) and they all came from epel so it's possible they were being ignored intentionally because a preferred repo had them at an older version | 19:03 |
| cloudnull | interesting | 19:04 |
| idlemind | cloudnull i'm wondering if updates are triggering automatically otherwise | 19:04 |
| idlemind | maybe mhayden could chime in | 19:04 |
| idlemind | if his updates are working as expected in his centos land | 19:04 |
| cloudnull | we do use the package state of "latest" for the deployment so when running the playbooks the packages across the cluster should be updated when the playbooks are executed | 19:04 |
| *** Leo_m has joined #openstack-ansible | 19:05 | |
| cloudnull | but, that might now catch normal system upgradwes | 19:05 |
| cloudnull | **upgrades | 19:05 |
| cloudnull | mhayden: would be a good person to ask though | 19:05 |
| cloudnull | especially as it pertains to cent | 19:05 |
| idlemind | ya usually a yum "*" state: latest would do the upgrade (along w/update_cache: yes) I believe ... those could be nestled in the playbooks along the way too | 19:06 |
| mhayden | so we have package_state=present for centos 7 in queens/master right now | 19:06 |
| mhayden | to speed things up | 19:07 |
| mhayden | what's the question/problem? | 19:07 |
| *** jra has joined #openstack-ansible | 19:07 | |
| *** poopcat has joined #openstack-ansible | 19:09 | |
| *** olivierbourdon38 has quit IRC | 19:09 | |
| *** olivierbourdon38 has joined #openstack-ansible | 19:10 | |
| jra | Hi all, I'm working on a new OSA install of Pike with the intention of using OpenvSwitch rather than Linux Birdges, and I'm encountering a few issues after a successful OSA deployment. Notably, it looks like /var/run/openvswitch/ doesn't get created, so ovsdb can't start up; and neutron-openvswitch-agent wants to talk to ovsdb on localhost:6604, but it's only configured to listen on its socket, not that TCP port. Are these k | 19:13 |
| *** olivierbourdon38 has quit IRC | 19:13 | |
| *** Leo_m has quit IRC | 19:14 | |
| *** Leo_m has joined #openstack-ansible | 19:15 | |
| goldenfri | @mhayden what is the proper way to set a custom login_banner (couldn't find it in the docs), should I just change /etc/ansible/roles/ansible-hardening/files/login_banner.txt? | 19:15 |
| *** olivierbourdon38 has joined #openstack-ansible | 19:16 | |
| *** olivierbourdon38 has quit IRC | 19:18 | |
| cloudnull | jra: is that temp path required for ovs ? | 19:18 |
| cloudnull | I would think that' | 19:18 |
| cloudnull | **that's something ovs would create on start | 19:18 |
| cloudnull | but maybe it's something we need to add to the tempfiles creation unit files | 19:18 |
| jra | cloudnull: it doesn't seem to create the path on its own. once the path exists, it happily creates the socket file there | 19:18 |
| odyssey4me | andymccr we used our modules because the upstream ones weren't capable of doing all the things we needed, but that's likely changed now - we should use as many as we can - at the PTG we also discussed using the openstack CLI to fill any gaps left so that we can remove the federaiton stuff too | 19:19 |
| *** olivierbourdon38 has joined #openstack-ansible | 19:19 | |
| mhayden | goldenfri: https://github.com/openstack/ansible-hardening/blob/a10fae4fe13c2f056cf3bde058653978c86ce187/defaults/main.yml#L281 | 19:19 |
| mhayden | goldenfri: security_login_banner_text is the variable you want to change | 19:19 |
| andymccr | odyssey4me: sweet thanks! I knew there was a historic reason i just couldn't remember | 19:19 |
| andymccr | cos we removed a bunch of the other modules | 19:19 |
| odyssey4me | andymccr there was also an issue with shade not being in u-c in older releases, but that's solved now... so we're a full go now for any upstream modules that do what we need | 19:20 |
| odyssey4me | mnaser thanks for https://review.openstack.org/#/c/550542/1 - mind if I tweak it so that it can still work outside of openstack CI | 19:20 |
| odyssey4me | ? | 19:21 |
| jra | cloudnull: even once the socket is created, though, the agent isn't actually configured to connect to it. It's expecting a listener on localhost:6604, but ovsdb is only configured to listen on the socket. | 19:21 |
| cloudnull | jra: mind pasting the files under /etc/tmpfiles.d/openstack-* ? | 19:21 |
| cloudnull | from one of the compute hosts | 19:21 |
| mnaser | odyssey4me: sure, my theory was bootstrap_host_ubuntu_repo and bootstrap_host_ubuntu_security_repo would be defined somehow externally and it'll still work | 19:21 |
| andymccr | sweet - i might give that a go | 19:21 |
| andymccr | thanks odyssey4me | 19:21 |
| mnaser | because of "bootstrap_host_ubuntu_repo is not defined" and "bootstrap_host_ubuntu_security_repo is not defined" | 19:21 |
| jra | This issue is appearing on the neutron servers, actually | 19:21 |
| cloudnull | ok from that node then . | 19:22 |
| jra | is there a preferred pastebin for this channel? | 19:22 |
| cloudnull | I use http://paste.openstack.org/ | 19:23 |
| cloudnull | but any will do | 19:23 |
| goldenfri | @mhayden perfect thank you! | 19:23 |
| jra | http://paste.openstack.org/show/693711/ | 19:23 |
| odyssey4me | mnaser yeah, the thing rewrites the sources to a known and optimised way - whether in openstack CI or not... obviously something's changed which has broken the grep... but you're right that some things in the bootstrap-host role need changing to better use mirror info | 19:23 |
| odyssey4me | I actually had it on my TODO with the hope of fixing this week, even before this break happened | 19:24 |
| odyssey4me | have we isolated what changed that caused this? | 19:24 |
| mnaser | odyssey4me: yes, infra added trusted=yes which breaks things | 19:24 |
| cloudnull | jra: ok, looks like we need to add that path to the neutron role so that we ensure teh temp path is present | 19:24 |
| mnaser | odyssey4me: http://logs.openstack.org/31/550331/1/experimental/openstack-ansible-integrated-deploy-aio/ac17d83/logs/ara/result/de09e872-438d-4e6f-9ec2-4ff03d541100/ | 19:24 |
| jra | @cloudnull: how can I help with that? | 19:25 |
| cloudnull | I dont run ovs so I cant say why that happened, maybe someone here might be able to chime in on their experiences however I'd be happy to throw a pr at it if needed. | 19:25 |
| cloudnull | jra: if you want to add something like https://github.com/openstack/openstack-ansible-os_neutron/blob/master/templates/neutron-systemd-tmpfiles.j2#L3 | 19:25 |
| odyssey4me | mnaser ah ok, I think we can safely just remove the rewriting of the sources and leave the host as it is | 19:26 |
| cloudnull | to the role when ovs is in use it'd be super helpful | 19:26 |
| odyssey4me | it was necessary back with trusty where we required the backports, but we don't do trust any more so meh | 19:26 |
| cloudnull | also if you could document your experience and what you did to solve the issue in a launchpad bug it'd be appreciated | 19:26 |
| mnaser | odyssey4me: your final call :) | 19:26 |
| odyssey4me | mnaser happy for me to modify the patch? | 19:26 |
| mnaser | odyssey4me: its all yours! | 19:26 |
| cloudnull | going to get a coffee brb | 19:27 |
| idlemind | mhayden i was seeing nothing but excluded packages on my 2 hosts because of yum priorities plugin ... i disabled it and i only had 6 updates to do ... not certain the last time i did an update but wanted to make sure your hosts were updating correctly w/it .. maybe there's a state: latest "*" somewhere that's doing updates when i re-run openstack-ansible to apply the day's configuration tweaks as i play | 19:27 |
| evrardjp | hey everyone | 19:27 |
| odyssey4me | apologies for not getting to it sooner - we had a team outing today | 19:27 |
| evrardjp | yeah same. | 19:27 |
| jra | cloudnull: sure, I'll get a bug filed and see about submitting a patch for it | 19:29 |
| *** olivierb_ has joined #openstack-ansible | 19:29 | |
| mhayden | idlemind: so the priorities are there to ensure you get the correct packages installed | 19:29 |
| *** holser__ has quit IRC | 19:29 | |
| mhayden | there's reasons why we prefer packages from RDO over EPEL | 19:30 |
| mhayden | :) | 19:30 |
| idlemind | mhayden k so i just have to figure out how / when my boxes have been updating themselves :) | 19:30 |
| *** Leo_m has quit IRC | 19:30 | |
| mhayden | hmm, perhaps you have automated updates turned on? that's part of the ansible-hardening role | 19:31 |
| *** Leo_m has joined #openstack-ansible | 19:31 | |
| *** Leo_m has quit IRC | 19:32 | |
| *** rromans has quit IRC | 19:33 | |
| *** RandomTech has joined #openstack-ansible | 19:36 | |
| *** olivierbourdon38 has quit IRC | 19:36 | |
| *** olivierbourdon38 has joined #openstack-ansible | 19:36 | |
| RandomTech | Hello. I am having an issue with my openstack-ansible cluster set-up. | 19:38 |
| idlemind | mhayden that's likely where it's coming from; i'll check that role to verify | 19:38 |
| RandomTech | I am encountering this issue when running the setup-infrastructure playbook: | 19:38 |
| RandomTech | fatal: [ctrl-02_galera_container-fbb8def4]: FAILED! => {"attempts": 3, "changed": false, "failed": true, "msg": "Unable to start service mysql: Job for mariadb.service failed because the control process exit ed with error code. See \"systemctl status mariadb.service\" and \"journalctl -xe\" for details.\n"} | 19:38 |
| RandomTech | Any ideas? | 19:38 |
| *** olivierbourdon38 has quit IRC | 19:38 | |
| *** Leo_m has joined #openstack-ansible | 19:39 | |
| *** olivierbourdon38 has joined #openstack-ansible | 19:39 | |
| *** Leo_m has quit IRC | 19:39 | |
| *** Leo_m has joined #openstack-ansible | 19:42 | |
| jra | cloudnull: https://bugs.launchpad.net/openstack-ansible/+bug/1754144 | 19:42 |
| openstack | Launchpad bug 1754144 in openstack-ansible "OpenvSwitch run path not created" [Undecided,New] | 19:42 |
| *** olivierbourdon38 has quit IRC | 19:43 | |
| *** olivierbourdon38 has joined #openstack-ansible | 19:43 | |
| jra | Any thoughts on the tcp-port-vs-unix-socket issue? It looks to me like the OVS portions of the deploy don't create a functioning environment out of the box - is this a rare deployment type? | 19:44 |
| *** Leo_m_ has joined #openstack-ansible | 19:45 | |
| evrardjp | yes jra | 19:45 |
| evrardjp | patches welcome. | 19:45 |
| evrardjp | We are working on it with OPNFV crew to make it part of our integrated gates | 19:46 |
| evrardjp | properly tested and all | 19:46 |
| *** Leo_m_ has quit IRC | 19:46 | |
| *** Leo_m has quit IRC | 19:47 | |
| *** olivierbourdon38 has quit IRC | 19:47 | |
| *** Leo_m has joined #openstack-ansible | 19:47 | |
| *** olivierbourdon38 has joined #openstack-ansible | 19:48 | |
| jra | evrardjp: thanks, that's great to know | 19:49 |
| jra | we're trying to decide whether to use OVS for a fairly large deployment that needs to get out the door in the next few months; if OVS isn't production-ready from OSA at this point, we'd probably choose to stick to Linux Bridges. Do you have any advice on that decision? | 19:50 |
| *** olivierb_ has quit IRC | 19:51 | |
| idlemind | jra what's the reasoning for ovs? | 19:51 |
| *** olivierbourdon38 has quit IRC | 19:51 | |
| *** olivierbourdon38 has joined #openstack-ansible | 19:51 | |
| *** poopcat1 has joined #openstack-ansible | 19:51 | |
| *** jafeha__ has joined #openstack-ansible | 19:52 | |
| jra | idlemind: mostly future-proofing; LB doesn't give us the prospect of hardware offloading, so parallel HPC-type workloads might not be an option if we go that way | 19:52 |
| *** woodard has quit IRC | 19:52 | |
| *** poopcat has quit IRC | 19:52 | |
| *** jafeha has quit IRC | 19:52 | |
| *** radeks has quit IRC | 19:52 | |
| *** radeks has joined #openstack-ansible | 19:52 | |
| *** olivierbourdon38 has quit IRC | 19:53 | |
| *** throwsb11 has joined #openstack-ansible | 19:53 | |
| *** olivierbourdon38 has joined #openstack-ansible | 19:53 | |
| *** throwsb1 has quit IRC | 19:53 | |
| *** jamesdenton has quit IRC | 19:54 | |
| *** woodard has joined #openstack-ansible | 19:55 | |
| *** rromans has joined #openstack-ansible | 19:55 | |
| evrardjp | thanks for the bug :) | 19:56 |
| evrardjp | jra: sticking to linux bridge is wiser for ops. | 19:57 |
| evrardjp | jra: I think you can use ovs already, but you might hit some bugs | 19:57 |
| evrardjp | also first queens release will be out of the door soon | 19:57 |
| evrardjp | so you might want to try this too :) | 19:57 |
| evrardjp | if you don't need ovs, just don't use it, linux bridge has been around for ages, and it works :) | 19:58 |
| jra | evrardjp: we are planning to move up to queens before doing our big deployment, I'm trying to establish a stable dev before then | 19:58 |
| jra | but your advice helps a lot, and I'll discuss with the team whether we want to do the development work that might be necessary to get a functional OVS versus just getting a working LB setup out the door | 19:58 |
| evrardjp | yeah it all depends on what you need to have | 19:59 |
| evrardjp | and your workloads | 19:59 |
| *** woodard_ has joined #openstack-ansible | 20:00 | |
| *** woodard has quit IRC | 20:02 | |
| *** ivve has quit IRC | 20:08 | |
| idlemind | is it possible to integrate this patch (https://review.openstack.org/#/c/545783/) into my openstack-ansible (stable/pike) cleanly? or do i have to wait for it to merge and backported to stable/pike of neutron and picked up there w/a commit-id update | 20:15 |
| idlemind | and is there a link to old bug triage irc meeting notes? | 20:18 |
| openstackgerrit | Jean-Philippe Evrard proposed openstack/openstack-ansible stable/queens: Bump A-R-R before first queens release https://review.openstack.org/550589 | 20:18 |
| *** RandomTech has quit IRC | 20:23 | |
| idlemind | found the irc meeting logs | 20:24 |
| *** jra has quit IRC | 20:26 | |
| *** throwsb11 has quit IRC | 20:31 | |
| *** hw_wutianwei has quit IRC | 20:31 | |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible master: Clean-up gate code to use mirror_info https://review.openstack.org/550542 | 20:32 |
| *** hw_wutianwei has joined #openstack-ansible | 20:32 | |
| openstackgerrit | Jean-Philippe Evrard proposed openstack/openstack-ansible stable/pike: Update all SHAs for 16.0.10 https://review.openstack.org/550593 | 20:35 |
| odyssey4me | mnaser ^ that's a start - I'll work on doing some more cleanup and simplifying of bootstrap-host later this week... it would seem it's time | 20:35 |
| odyssey4me | of course, if you spot some things that you think will help simplify, please go ahead and propose patches :) | 20:36 |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible-tests stable/queens: Add ansible-lint rule for apt_repository https://review.openstack.org/550594 | 20:37 |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible-tests stable/pike: Add ansible-lint rule for apt_repository https://review.openstack.org/550595 | 20:37 |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible-tests stable/ocata: Add ansible-lint rule for apt_repository https://review.openstack.org/550596 | 20:37 |
| mnaser | odyssey4me: that cleaned up nicely, the shell/source/write file bothers me but i cant think of a better cleaner way :D | 20:39 |
| odyssey4me | mnaser yep, it's for tests only so I'm trying not to think too hard about it | 20:40 |
| mnaser | odyssey4me: yeah, that's valid, resources better spend in other more important things :) | 20:40 |
| odyssey4me | if there's a better way, I'm all ears | 20:40 |
| odyssey4me | if this passes, we'll need to port this back all the way to ocata (newton is about to EOL)... but there are subtle differences that need to be made for ocata. | 20:41 |
| mnaser | odyssey4me: i was thinking source the file in the gate shell job and use lookup('env', 'NODEPOOL_FOO') to create the file with a template | 20:42 |
| mnaser | but i don't have the time to do that right now, so this is good enough by me :) | 20:43 |
| odyssey4me | mnaser we're trying to reduce the amount of bash ;) | 20:43 |
| odyssey4me | that would work, but then it's also another template | 20:43 |
| odyssey4me | meh - it's kinda the same | 20:43 |
| mnaser | odyssey4me: in some ideal world, zuul calls the playbook directly, zuul exposes the mirror info via variables | 20:43 |
| mnaser | and we do this with variables only, but yeah :D | 20:44 |
| odyssey4me | mnaser the trouble with zuul native jobs is that we have no way to run them outside of openstack-ci | 20:45 |
| mnaser | odyssey4me: yeah, that's the one thing i struggled with, but my thought process was that you'd end up with a simple ./run_tests.sh wrapper which install ansible locally and then starts the zuul playbook | 20:46 |
| odyssey4me | so our ideal situation is to have a thin shim that gets us to a place where we can run the job the way zuul runs it - that shim should be usable for vagrant, etc - but can be ignored by zuul | 20:46 |
| mnaser | my thoughts exactly :> | 20:46 |
| *** Guest52 has joined #openstack-ansible | 20:46 | |
| openstackgerrit | Jean-Philippe Evrard proposed openstack/openstack-ansible stable/ocata: Update all SHAs for 15.1.18 https://review.openstack.org/550599 | 20:47 |
| odyssey4me | to do that we'd ideally need to install the same version of zuul, which is achievable - and I'd like to discuss that option and figure out an easy way to reference the pin | 20:47 |
| odyssey4me | even better would be for zuul to publish the venv containing the right ansible version... so we just download, fix the shebangs, and execute | 20:48 |
| odyssey4me | go one better and provide a curl | bash script to do it all for you ;) | 20:48 |
| evrardjp | mnaser: glad you are in sync :D | 20:48 |
| persia | In another channel, almost right now, there is a discussion about how venvs are not relocatable, which might break that model | 20:48 |
| odyssey4me | :) persia we use venvs, and we know there are some limitations - but it works for us | 20:49 |
| evrardjp | they can be relocatable , depending on some cases | 20:49 |
| *** Guest52 has left #openstack-ansible | 20:50 | |
| persia | Fair enough :) | 20:50 |
| odyssey4me | they're not universally relocatable - it depends on c bindings and all that | 20:50 |
| * persia was mostly amused at the synchonicity, and doesn't actually understand all the details about venv relocation | 20:50 | |
| odyssey4me | but anyway - I need to go, time to !computer for the day | 20:50 |
| odyssey4me | persia it is actually quite weird how that happens :) | 20:50 |
| odyssey4me | night all | 20:53 |
| *** dariko has quit IRC | 20:53 | |
| *** dariko has joined #openstack-ansible | 20:55 | |
| *** dariko has quit IRC | 20:55 | |
| *** hvvclint has joined #openstack-ansible | 21:01 | |
| *** osnaya has joined #openstack-ansible | 21:07 | |
| osnaya | @cloudnull a quick q about ansible deployment on Ubuntu Host .... in env.d folder now I have aio_metal.yml... Do I need to remove the default extra_container.yml file or can it stay there but will be ignored | 21:11 |
| *** pcaruana has quit IRC | 21:12 | |
| cloudnull | its just an example file , | 21:13 |
| cloudnull | you can remove it if you like | 21:13 |
| osnaya | @cloudnull.... ok, so even if it is present, it will be ignore because aio_metal.yml will be picked up for execution right? | 21:14 |
| *** Leo_m has quit IRC | 21:17 | |
| cloudnull | yes | 21:20 |
| openstackgerrit | Kevin Carter (cloudnull) proposed openstack/openstack-ansible-os_neutron master: Create ovs temp directories when using ovs https://review.openstack.org/550609 | 21:25 |
| cloudnull | jra: ^ | 21:25 |
| mnaser | ok.. i'll kindly ask for spare eyes... 'nova_git_install_branch: 0b7f8064df75d01370bb4405547f216b004d90da' in user_variables.yml, then openstack-ansible repo-build.yml ... but osa doesn't seem to build that new wheel/venv | 21:27 |
| mnaser | but it looks like for some reason, the variable isn't even being passed? | 21:27 |
| mnaser | skipping: [controller-1-dev_repo_container-e6c3f913] => (item={'name': u'nova', 'project_group': u'nova_all', 'fragment': None, 'url': u'https://git.openstack.org/openstack/nova', 'version': u'9adf97c0132c3e0d34bbb2e33b59c6e5b00bec57', 'original': u'git+https://git.openstack.org/openstack/nova@9adf97c0132c3e0d34bbb2e33b59c6e5b00bec57#egg=nova&gitname=nova&projectgroup=nova_all', 'egg_name': u'nova'}) | 21:27 |
| *** osnaya has quit IRC | 21:28 | |
| *** radeks has quit IRC | 21:52 | |
| cloudnull | mnaser: if the underlying version of OSA hasn't changed and the venv already exists it wont rebuild unless the option "repo_build_venv_rebuild=yes" is set | 22:00 |
| mnaser | =yes or =true? i tried to run with =true but i guess i can try =yes | 22:00 |
| mnaser | openstack-ansible repo-build.yml -e 'repo_build_venv_rebuild=true' | 22:00 |
| cloudnull | either will work | 22:00 |
| mnaser | i'll give it another shot, maybe i missed something | 22:00 |
| *** Leo_m has joined #openstack-ansible | 22:01 | |
| openstackgerrit | Major Hayden proposed openstack/openstack-ansible-tests master: Improve idempotency https://review.openstack.org/550135 | 22:02 |
| openstackgerrit | Major Hayden proposed openstack/openstack-ansible-openstack_hosts master: Make role idempotent https://review.openstack.org/550102 | 22:08 |
| mnaser | cloudnull: this is strange, i ran it with that and it didnt attempt to rebuild stuff... | 22:09 |
| cloudnull | hum... that should be all that | 22:09 |
| cloudnull | *that's needed. | 22:09 |
| mnaser | cloudnull: what's weirder is.. its trying to clone the git commit that is in the original openstack_services.yml .. but not the one in my user_variables.yml | 22:09 |
| mnaser | skipping: [controller-1-dev_repo_container-e6c3f913] => (item={'name': u'nova', 'project_group': u'nova_all', 'fragment': None, 'url': u'https://git.openstack.org/openstack/nova', 'version': u'9adf97c0132c3e0d34bbb2e33b59c6e5b00bec57', 'original': u'git+https://git.openstack.org/openstack/nova@9adf97c0132c3e0d34bbb2e33b59c6e5b00bec57#egg=nova&gitname=nova&projectgroup=nova_all', 'egg_name': u'nova'}) | 22:09 |
| mnaser | repo_build : Build package requirements file <= this tasks seems to just say 'ok' right away.. so not saying that it's changing anything | 22:10 |
| cloudnull | curious, running `openstack-ansible repo-build.yml -e 'repo_build_venv_rebuild=true'` ? | 22:10 |
| mnaser | i also checked | 22:10 |
| mnaser | the requirements.txt file it builds .. "nova # 9adf97c0132c3e0d34bbb2e33b59c6e5b00bec57" | 22:11 |
| mnaser | so that file is not even changing | 22:11 |
| mnaser | # grep nova_git /etc/openstack_deploy/user_variables.yml => nova_git_install_branch: 0b7f8064df75d01370bb4405547f216b004d90da | 22:11 |
| mnaser | i'll give that a shot but it looks to me that it's pretty much ignoring that | 22:11 |
| mnaser | by that, ignoring my 'nova_git_install_branch' | 22:12 |
| cloudnull | with "nova_git_install_branch: $SOMESHA" in your user_variables.yml file it should just go. | 22:12 |
| cloudnull | though you could also specify that on the cli | 22:12 |
| cloudnull | `openstack-ansible repo-build.yml -e 'repo_build_venv_rebuild=true' -e 'nova_git_install_branch=$SOMESHA'` | 22:12 |
| mnaser | i'll try that as a second attempt | 22:13 |
| mnaser | once this run finishes | 22:13 |
| *** openstackstatus has quit IRC | 22:13 | |
| *** Leo_m has quit IRC | 22:13 | |
| *** openstack has joined #openstack-ansible | 22:17 | |
| *** ChanServ sets mode: +o openstack | 22:17 | |
| mnaser | the load local packages task has resulted with `git+https://git.openstack.org/openstack/nova@9adf97c0132c3e0d34bbb2e33b59c6e5b00bec57#egg=nova&gitname=nova&projectgroup=nova_all` in remote_packages | 22:17 |
| cloudnull | mnaser: maybe you also need https://github.com/openstack/openstack-ansible-repo_build/blob/master/defaults/main.yml#L65 | 22:17 |
| mnaser | cloudnull: does the 'load local packages' task load things up to pass to repo_build? if they don't have the updated hash, i think the repo_build would never be able to solve the issue in the first place? | 22:18 |
| cloudnull | is "nova_git_install_branch" set within any of your existing user variable files ? | 22:18 |
| *** Leo_m has joined #openstack-ansible | 22:18 | |
| mnaser | cloudnull: only user_secrets and user_variables, and it's set in user_variables to tip of stable/pike (0b7f8064df75d01370bb4405547f216b004d90da) but the value is not being used | 22:18 |
| cloudnull | ok. | 22:18 |
| mnaser | cloudnull: http://paste.openstack.org/show/693938/ and that was with -e nova_git_install_branch= | 22:19 |
| mnaser | oh, it got truncated. | 22:20 |
| mnaser | http://paste.openstack.org/show/693941/ | 22:21 |
| mnaser | this got truncated too but it has the git hashes which dont match.. | 22:21 |
| cloudnull | hum, that's not great. can recursively grep for nova_git_install_branch in the /etc/openstack_deploy directory | 22:23 |
| cloudnull | what you have on the cli should always win | 22:23 |
| mnaser | grep -ir nova_git_install_branch /etc/openstack_deploy => /etc/openstack_deploy/user_variables.yml:nova_git_install_branch: 0b7f8064df75d01370bb4405547f216b004d90da | 22:23 |
| mnaser | yeah thats what i know with ansible too | 22:23 |
| mhayden | hwoarang / cloudnull: on opensuse, what's the filesystem permissions for stuff in /etc/cron.d/ ? | 22:24 |
| mnaser | i'm gonna add a debug msg= nova_git_install_branch | 22:25 |
| cloudnull | root:root | 22:25 |
| cloudnull | mhayden: ^ | 22:25 |
| mhayden | the perms | 22:25 |
| mhayden | trying to fiure out why it shows the cron as changed | 22:25 |
| cloudnull | drwxr-xr-x | 22:25 |
| mhayden | http://logs.openstack.org/02/550102/10/check/openstack-ansible-functional-opensuse-423/c09e034/job-output.txt.gz#_2018-03-07_22_16_21_582371 | 22:25 |
| mhayden | i'm thinking opensuse might use different file perms for stuff in /etc/cron.d/ | 22:25 |
| mhayden | 755 | 22:25 |
| mhayden | okay | 22:25 |
| mhayden | no, wait, the files inside /etc/cron.d/ | 22:26 |
| mhayden | sorry | 22:26 |
| mhayden | :P | 22:26 |
| cloudnull | its drwxr-xr-x on my machine running opensuse tumbleweed | 22:26 |
| mnaser | ok.. added a debug.. "msg": "0b7f8064df75d01370bb4405547f216b004d90da" .. so the var is making its way | 22:26 |
| cloudnull | 0644 for the filles | 22:26 |
| mhayden | ah, 0644! | 22:26 |
| mhayden | that's it | 22:26 |
| mhayden | thank you sir | 22:26 |
| mhayden | centos, ubuntu and suse are all different :| | 22:26 |
| cloudnull | lovely | 22:26 |
| cloudnull | maybe we should use systemd timners | 22:26 |
| cloudnull | **timers | 22:26 |
| openstackgerrit | Major Hayden proposed openstack/openstack-ansible-openstack_hosts master: Make role idempotent https://review.openstack.org/550102 | 22:27 |
| mhayden | hopefully this works ^^ | 22:27 |
| *** woodard has joined #openstack-ansible | 22:27 | |
| *** hvvclint has quit IRC | 22:29 | |
| *** woodard_ has quit IRC | 22:31 | |
| *** woodard has quit IRC | 22:32 | |
| mnaser | mhayden: https://github.com/openstack/openstack-ansible-plugins/blob/master/lookup/py_pkgs.py#L325-L340 | 22:33 |
| mnaser | oops | 22:33 |
| mnaser | cloudnull: ^ | 22:33 |
| mnaser | it looks like when the files are filtered to be parsed | 22:33 |
| mnaser | the order of defaults vs user variable files | 22:33 |
| mnaser | doesnt matter? | 22:33 |
| mnaser | ill add a print locally but i think thats whats happening.. | 22:34 |
| cloudnull | all of the user_ files should be parsed last | 22:34 |
| mnaser | cloudnull: but that code doesnt seem to do any ordering | 22:35 |
| *** esberglu has quit IRC | 22:35 | |
| *** fusmu has joined #openstack-ansible | 22:37 | |
| cloudnull | it should be ordered by the file_name_words however we could zip those lists to filter more strictly | 22:38 |
| mnaser | cloudnull: but the loop seems to happen over the file names provided then appended if they're part of file_name_words | 22:39 |
| mnaser | so if i understand the code.. if the order of files comes incorrect (user before than role default), this won't fix it | 22:39 |
| cloudnull | i see . | 22:40 |
| mnaser | i added a print and i see this as the first line | 22:40 |
| mnaser | ['/opt/openstack-ansible/etc/openstack_deploy/user_secrets.yml', '/opt/openstack-ansible/etc/openstack_deploy/user_variables.yml', '/opt/openstack-ansible/playbooks/defaults/repo_packages/gnocchi.yml', '/opt/openstack-ansible/playbooks/defaults/repo_packages/nova_consoles.yml', '/opt/openstack-ansible/playbooks/defaults/repo_packages/openstack_services.yml', | 22:40 |
| mnaser | '/opt/openstack-ansible/playbooks/defaults/repo_packages/openstack_testing.yml', '/opt/openstack-ansible/playbooks/roles/system_crontab_coordination/defaults/main.yml', '/opt/openstack-ansible/tests/roles/bootstrap-host/defaults/main.yml', '/opt/openstack-ansible/tests/roles/bootstrap-host/vars/redhat.yml', '/opt/openstack-ansible/tests/roles/bootstrap-host/vars/ubuntu.yml', | 22:40 |
| mnaser | '/opt/openstack-ansible/tests/roles/bootstrap-host/vars/suse.yml'] | 22:40 |
| mnaser | oops that was a lot, but you see it took it first, then openstack_services was after | 22:40 |
| *** fusmu has quit IRC | 22:42 | |
| *** openstackstatus has quit IRC | 22:42 | |
| *** openstack has joined #openstack-ansible | 22:46 | |
| *** ChanServ sets mode: +o openstack | 22:46 | |
| cloudnull | well that would work if the '/etc/openstack_deploy/' was last | 22:46 |
| mnaser | yeah. i'm doing a bit more troubleshooting now | 22:46 |
| mnaser | so when it parses the '/opt/openstack-ansible/playbooks/../' path .. it generates all the git requirements, but when it does the iteration around /etc/ansible/roles and /etc/openstack_deploy .. it gets nothing | 22:50 |
| mnaser | i understand why /etc/ansible/roles gets nothing, but /etc/openstack_deploy should create the nova one | 22:50 |
| *** esberglu has joined #openstack-ansible | 22:50 | |
| cloudnull | indeed it should | 22:51 |
| mnaser | cloudnull: ha | 22:51 |
| mnaser | got it | 22:51 |
| mnaser | https://github.com/openstack/openstack-ansible-plugins/blob/master/lookup/py_pkgs.py#L554-L561 | 22:51 |
| mnaser | only if a git_repo is defined it will actually process git | 22:51 |
| mnaser | which means that i will need to point the git repo too (even though it's just upstream) | 22:51 |
| cloudnull | ah ha! | 22:52 |
| mnaser | i guess i should add the nova_git_project_group too.. | 22:52 |
| cloudnull | hum... that's kinda a bummer. so both variables need to be defined | 22:53 |
| cloudnull | I guess most folks setting the sha have also set the repo url ]. | 22:53 |
| mnaser | all 3 i guess | 22:53 |
| mnaser | oh | 22:53 |
| mnaser | no, not necessarily, itll get the override | 22:54 |
| mnaser | maybe the check should be based on sha too | 22:54 |
| mnaser | as that's most likely to be overridden more often than repo | 22:54 |
| cloudnull | ++ | 22:54 |
| cloudnull | add git_install_branch to the key check and it should be ok. | 22:55 |
| mnaser | and i guess extra vars dont matter | 22:55 |
| mnaser | because this thing parses them outside ansible | 22:55 |
| cloudnull | yea the plugin feeds ansible. my bad. I was misleading you | 22:56 |
| cloudnull | with the -e bit s | 22:56 |
| mnaser | no worries | 22:56 |
| mnaser | it wasnt straightforward :D | 22:56 |
| cloudnull | We've been using this lookup for a few years now but I believe its one of the things that odyssey4me was wanting to remove given the pip install simplification effort | 22:57 |
| cloudnull | so hopefully we can make these types of changes a whole simpler in the near future | 22:57 |
| *** armaan has quit IRC | 22:58 | |
| *** armaan has joined #openstack-ansible | 22:59 | |
| *** pmannidi has joined #openstack-ansible | 22:59 | |
| openstackgerrit | Mohammed Naser proposed openstack/openstack-ansible-plugins master: Check for keys ending with git_install_branch https://review.openstack.org/550640 | 23:02 |
| mnaser | cloudnull: ^ | 23:03 |
| mnaser | everything worked just fine after | 23:03 |
| mnaser | no need for any extra vars, ansible figured out itself that it needed to rebuild wheels and venvs | 23:03 |
| *** esberglu has quit IRC | 23:09 | |
| *** threestrands has joined #openstack-ansible | 23:12 | |
| *** threestrands has quit IRC | 23:12 | |
| *** threestrands has joined #openstack-ansible | 23:12 | |
| *** kstev has quit IRC | 23:19 | |
| *** armaan has quit IRC | 23:33 | |
| *** armaan has joined #openstack-ansible | 23:34 | |
| *** threestrands_ has joined #openstack-ansible | 23:34 | |
| *** threestrands_ has quit IRC | 23:35 | |
| *** threestrands_ has joined #openstack-ansible | 23:36 | |
| *** threestrands has quit IRC | 23:36 | |
| *** Leo_m has quit IRC | 23:38 | |
| *** Leo_m_ has joined #openstack-ansible | 23:39 | |
| *** osnaya has joined #openstack-ansible | 23:53 | |
| osnaya | openstack-ansible setup-hosts.yml --syntax-check throws error "Invalid extra vars data supplied. '@/etc/openstack_deploy/user_secrets.yml' could not be made into a dictionary" - any ideas or pointers please? | 23:55 |
Generated by irclog2html.py 2.15.3 by Marius Gedminas - find it at mg.pov.lt!
