| *** maharg101 has joined #openstack-ansible | 00:20 | |
| *** rf0lc0 has quit IRC | 00:22 | |
| *** maharg101 has quit IRC | 00:25 | |
| *** gyee has quit IRC | 00:31 | |
| *** ianychoi__ is now known as ianychoi | 00:44 | |
| *** cshen has joined #openstack-ansible | 01:07 | |
| *** cshen has quit IRC | 01:12 | |
| *** maharg101 has joined #openstack-ansible | 02:21 | |
| *** maharg101 has quit IRC | 02:26 | |
| *** macz_ has joined #openstack-ansible | 02:41 | |
| *** macz_ has quit IRC | 02:46 | |
| *** NewJorg has quit IRC | 02:58 | |
| *** NewJorg has joined #openstack-ansible | 02:59 | |
| *** cshen has joined #openstack-ansible | 03:07 | |
| *** cshen has quit IRC | 03:12 | |
| *** maharg101 has joined #openstack-ansible | 04:22 | |
| *** maharg101 has quit IRC | 04:26 | |
| *** evrardjp has quit IRC | 04:33 | |
| *** evrardjp has joined #openstack-ansible | 04:33 | |
| *** cshen has joined #openstack-ansible | 05:08 | |
| *** cshen has quit IRC | 05:12 | |
| *** nurdie has joined #openstack-ansible | 05:43 | |
| *** cshen has joined #openstack-ansible | 05:45 | |
| *** nurdie has quit IRC | 05:47 | |
| *** cshen has quit IRC | 05:50 | |
| *** nurdie has joined #openstack-ansible | 06:03 | |
| *** nurdie has quit IRC | 06:08 | |
| *** macz_ has joined #openstack-ansible | 06:18 | |
| *** miloa has joined #openstack-ansible | 06:20 | |
| *** macz_ has quit IRC | 06:22 | |
| *** maharg101 has joined #openstack-ansible | 06:23 | |
| *** maharg101 has quit IRC | 06:27 | |
| *** masterpe has quit IRC | 06:35 | |
| *** ioni has quit IRC | 06:35 | |
| *** csmart has quit IRC | 06:35 | |
| *** fridtjof[m] has quit IRC | 06:35 | |
| *** cshen has joined #openstack-ansible | 06:38 | |
| *** openstackgerrit has joined #openstack-ansible | 06:40 | |
| openstackgerrit | Jonathan Rosser proposed openstack/openstack-ansible-lxc_hosts master: Include libpython and rsync for centos in lxc base image https://review.opendev.org/756587 | 06:40 |
|---|---|---|
| *** cshen has quit IRC | 06:42 | |
| *** csmart has joined #openstack-ansible | 06:44 | |
| *** ioni has joined #openstack-ansible | 07:10 | |
| *** fridtjof[m] has joined #openstack-ansible | 07:10 | |
| *** masterpe has joined #openstack-ansible | 07:10 | |
| *** maharg101 has joined #openstack-ansible | 07:22 | |
| *** cshen has joined #openstack-ansible | 07:23 | |
| *** rpittau|afk is now known as rpittau | 07:27 | |
| *** andrewbonney has joined #openstack-ansible | 07:27 | |
| *** miloa has quit IRC | 07:33 | |
| *** jamesdenton has quit IRC | 07:33 | |
| *** miloa has joined #openstack-ansible | 07:33 | |
| jrosser | morning | 07:35 |
| *** jamesdenton has joined #openstack-ansible | 07:38 | |
| openstackgerrit | Jonathan Rosser proposed openstack/openstack-ansible-os_cinder master: Add rsync to required packages for redhat based OS https://review.opendev.org/756651 | 07:45 |
| openstackgerrit | Dmitriy Rabotyagov (noonedeadpunk) proposed openstack/openstack-ansible master: Add infra testing scenario https://review.opendev.org/755497 | 07:51 |
| *** tosky has joined #openstack-ansible | 07:54 | |
| recyclehero | morning | 08:06 |
| recyclehero | images dont get uploaded. I have set debug: true and didnt to any glance specefic configs. I see up until the post request in the logs but after that nothing. | 08:14 |
| recyclehero | How can I investigate | 08:14 |
| recyclehero | i didnt create the LV that was told in the docs? | 08:16 |
| recyclehero | but I think glance will then use file, right? | 08:16 |
| recyclehero | do I need to override anything | 08:16 |
| *** pto has joined #openstack-ansible | 08:21 | |
| *** pto has joined #openstack-ansible | 08:24 | |
| pto | The Ussuri docs states that you should clone master (git clone -b master https://opendev.org/openstack/openstack-ansible /opt/openstack-ansible) should that not be a tag instead? | 08:25 |
| noonedeadpunk | recyclehero: I can suggest you setting `glance_use_uwsgi: false` and trying that out | 08:25 |
| noonedeadpunk | Not sure, but maybe you're trying to use interoperable import as an image upload method, which is known not to work with uwsgi | 08:26 |
| noonedeadpunk | pto it should | 08:26 |
| pto | noonedeadpunk: 21.0.1 tag? | 08:26 |
| recyclehero | noonedeadpunk: thanks. I was checking in the browser and saw some CORS error. I was using the internal lb address to connect to it. | 08:27 |
| recyclehero | yes 21.0.1 | 08:27 |
| *** pto has quit IRC | 08:27 | |
| noonedeadpunk | oh, it's totally time for release... | 08:27 |
| *** CeeMac has joined #openstack-ansible | 08:31 | |
| recyclehero | its a CORS error which prevent me from uploading image, tring to disable cors in firefox | 08:41 |
| jrosser | recyclehero: https://github.com/openstack/openstack-ansible-os_glance/blob/master/releasenotes/notes/add-cors-config-6326223fe7fa7423.yaml | 08:45 |
| jrosser | as an end user with a browser it's probably best to use the external lb address | 08:46 |
| recyclehero | I changed to use the load balancer address like 200.200.200.110 | 08:46 |
| recyclehero | but the CORS complaint about port mismatch I guess | 08:47 |
| recyclehero | 200.200.200.200.119:9292 | 08:47 |
| recyclehero | 0 | 08:47 |
| recyclehero | 200.200.200.110:9292 | 08:47 |
| recyclehero | I know the address is not private, I was wrong when I set up that netowrk | 08:47 |
| recyclehero | https://200.200.200.110:9292/v2/images/ba0d5ec8-bc4b-4b17-91f2-ec968445910d/file | 08:48 |
| jrosser | anyway, my point is that there is already provision in the glance role to set the cors headers | 08:48 |
| recyclehero | jrosser: thank u I will get into that | 08:48 |
| *** johnsom has quit IRC | 09:25 | |
| *** johnsom has joined #openstack-ansible | 09:25 | |
| *** macz_ has joined #openstack-ansible | 09:55 | |
| *** macz_ has quit IRC | 09:59 | |
| *** SecOpsNinja has joined #openstack-ansible | 10:01 | |
| recyclehero | are these still relevant? https://docs.openstack.org/project-deploy-guide/openstack-ansible/ocata/app-advanced-config-override.html | 10:03 |
| recyclehero | The general format for the variable names used for overrides is <service>_<filename>_<file extension>_overrides | 10:03 |
| recyclehero | ? | 10:03 |
| noonedeadpunk | generally yes, but you'd better double check naming inside the role were overrides will be applied, ie for os_nova it will be https://opendev.org/openstack/openstack-ansible-os_nova/src/branch/master/defaults/main.yml#L391-L398 | 10:07 |
| noonedeadpunk | sorry, wrong selection | 10:08 |
| noonedeadpunk | this one is valid https://opendev.org/openstack/openstack-ansible-os_nova/src/branch/master/defaults/main.yml#L498-L504 | 10:08 |
| recyclehero | noonedeadpunk: thank u, after doing this override, which one of setup_hosts, infrastructure or openstack should be execeuted? | 10:11 |
| recyclehero | all? | 10:11 |
| recyclehero | I think just openstack | 10:13 |
| noonedeadpunk | well, I'd say depends on what exactly you're doing:) in case of overriding one of nova configs, worth shooting just os-nova-install.yml | 10:13 |
| noonedeadpunk | and we have a playbook and role for each specific service | 10:27 |
| recyclehero | noonedeadpunk: can u please chew it a little bit more for me? how to shoot one off playbook | 10:29 |
| recyclehero | openstack-ansible os-glance.yml | 10:29 |
| recyclehero | or I should use ansible commands? ansible ansble-playbook | 10:29 |
| recyclehero | openstack-ansbile /etc/ansible/roles/glance/tasks/main.yml ? | 10:31 |
| recyclehero | ERROR! 'include_vars' is not a valid attribute for a Play | 10:33 |
| noonedeadpunk | openstack-ansible /opt/openstack-ansible/playbooks/os-glance-install.yml | 10:37 |
| noonedeadpunk | openstack-ansible as well as ansible-playbook (openstack-ansible is eventually just a wrapper around ansible-playbook) can launch plays, which in their turn include roles, which are stored in /etc/ansible/roles/ | 10:39 |
| noonedeadpunk | but you can't run role directly without a play | 10:39 |
| recyclehero | so I should write a play myself indicating hosts which are the glance container in my case | 10:44 |
| noonedeadpunk | well, as I said, we have all sets of playbooks | 10:45 |
| noonedeadpunk | ie glance one is here https://opendev.org/openstack/openstack-ansible/src/branch/master/playbooks/os-glance-install.yml | 10:46 |
| recyclehero | yes, I just saw them. thank you. | 10:46 |
| noonedeadpunk | and setup-openstack.yml is just an include of all these playbooks in correct order for deployment/upgrade | 10:46 |
| *** rh-jlabarre has quit IRC | 10:49 | |
| recyclehero | yeah it went ok and I saw it replacing configs. but does it restart the service or make it re read its configuration | 10:57 |
| recyclehero | ? | 10:57 |
| recyclehero | cause its still comlaining about CORS | 10:57 |
| noonedeadpunk | yep, it should do that in case config is changed via handlers | 10:57 |
| noonedeadpunk | you should see service restart in the output | 10:57 |
| noonedeadpunk | well, cors is pretty tricky. if you reach via domain name - you should place it for CORS | 10:58 |
| recyclehero | no I reach via IP and put the ip in there. allowed_origin: https://200.200.200.110. I checked the request in the browser and its what I ahve specified | 10:59 |
| noonedeadpunk | if you upload via horizon, you may also want to try setting legacy upload method | 11:00 |
| recyclehero | noonedeadpunk: handlers doesnt write in their output things like [ok] [changed] [skipped] | 11:00 |
| recyclehero | ? | 11:00 |
| noonedeadpunk | they do when executed | 11:00 |
| recyclehero | so for stop and start they didnt write anything under them | 11:01 |
| noonedeadpunk | it should be called `Execute service restart` | 11:02 |
| noonedeadpunk | ok, we might have a bug here... | 11:03 |
| noonedeadpunk | can you check `systemctl status gnocchi-api`? | 11:03 |
| recyclehero | on infra1? | 11:03 |
| recyclehero | it seems I dont have this servic | 11:04 |
| recyclehero | I commented when on "Execute service restart" thats what I can do ;d | 11:06 |
| recyclehero | noonedeadpunk: TASK [Disable the service restart requirement] | 11:16 |
| *** lkoranda has joined #openstack-ansible | 11:17 | |
| noonedeadpunk | recyclehero: on infra inside glance container | 11:19 |
| noonedeadpunk | oh well | 11:19 |
| noonedeadpunk | `systemctl status glance-api` ofc, sorry | 11:19 |
| recyclehero | its active | 11:20 |
| recyclehero | but I just restart it to check my config | 11:20 |
| noonedeadpunk | yeah, I mean time when it was launched (when it was restarted last time) | 11:20 |
| recyclehero | it seems it was restarted but I have played with the playbook | 11:22 |
| recyclehero | do u want me to check with the orig playbook and report back | 11:22 |
| recyclehero | ? | 11:22 |
| SecOpsNinja | hello. atm in my openstack deployment i have only 1 infra note + 1 compute + storage, but im sessing that the majority of memory in infra hosts is consumed by uwsgi. is there any way to reduce this? the infra node has 16GB but with only 349 free mem and its using swap. i was checking each lxc contianer memory in there and i saw the bigger ones were nova_api (3,36G), heat-api (2.17G), cinder | 11:22 |
| SecOpsNinja | (1,57G), keystone (1,42G), glance (1,23G) and magnum (1,06G) | 11:22 |
| recyclehero | but why isnt it respecting my config | 11:22 |
| noonedeadpunk | recyclehero: you won't get it restarted unless config is changed | 11:22 |
| noonedeadpunk | with original playbook/role | 11:23 |
| noonedeadpunk | SecOpsNinja: every role with uwsgi has variable like `{{service_name }}_wsgi_processes` which I guess set to 16. Decreasing this number will help in decreasing ram usage as well | 11:25 |
| noonedeadpunk | recyclehero: you can try setting `horizon_images_upload_mode: legacy` and run os-horizon-install playbook | 11:26 |
| recyclehero | noonedeadpunk: I changed the config and its playing with orig conf | 11:27 |
| noonedeadpunk | this will make uploads to be done through horizon container and does not require cors | 11:27 |
| recyclehero | noonedeadpunk: ok I will tryt hat | 11:27 |
| noonedeadpunk | it's not the best solution, but it should work. after you can rollback and continue playing with cors :) | 11:28 |
| recyclehero | I changed the config and it havent restarted. thought u should know | 11:29 |
| noonedeadpunk | ok, I see.... | 11:30 |
| SecOpsNinja | noonedeadpunk, thanks. i was checking nova default/main and yep i found nova_wsgi_processes: "{{ [[ansible_processor_vcpus|default(1), 1] | max * 2, nova_wsgi_processes_max] | min }}" but couldn't understand what would be the processors valour for exmaple with a machine with 8 processores (proc/cpuinfo) | 11:30 |
| SecOpsNinja | and i need to set them in all container roles right? there isn't a global variable to override them all | 11:31 |
| noonedeadpunk | no, there's no global setting unfortunatelly | 11:32 |
| noonedeadpunk | I think it will be 16 | 11:33 |
| noonedeadpunk | you can also set nova_wsgi_processes_max but it will be kind of the same result here | 11:33 |
| SecOpsNinja | yep for what i understand the value is going to pik the mininal od cpus *2 and the max value defined right? | 11:34 |
| noonedeadpunk | yep | 11:35 |
| recyclehero | noonedeadpunk: horizon_images_upload_mode in user_varialbes. just like that? and play horizon afterwards ? | 11:36 |
| admin0 | checking if anyone can guide me in removing .novalocal from hostnames -- new and existing | 11:36 |
| SecOpsNinja | ok i will make a few test to try to reduce this memory usage. im going to add 2 more infra host but i would like to have the same lxc containers in all of them for redundancy that is why im try to mess with each service to reduce its memory ocupation | 11:36 |
| noonedeadpunk | admin0: use nova overrides and set `dhcp_domain` in DEFAULT section | 11:38 |
| noonedeadpunk | regarding existing - not sure | 11:38 |
| noonedeadpunk | SecOpsNinja: well, 16Gb is not much for controllers tbh | 11:40 |
| noonedeadpunk | I think it's ok until you have really basic set of services | 11:40 |
| noonedeadpunk | but also keep in mind that decresing amount of processes will reflect on amount of request that could be process at moment of time | 11:41 |
| recyclehero | SecOpsNinja: [[ansible_processor_vcpus|default(1), 1] | max * 2, | 11:41 |
| recyclehero | if you found out how to read this please share it with me too | 11:41 |
| recyclehero | is it yaml? | 11:42 |
| noonedeadpunk | it's jinja I think) | 11:43 |
| *** macz_ has joined #openstack-ansible | 11:43 | |
| noonedeadpunk | imo, max filter here is pointless | 11:43 |
| noonedeadpunk | recyclehero: you're right, we have pretty serious bug here.... | 11:44 |
| * recyclehero receives bug bounty | 11:46 | |
| recyclehero | noonedeadpunk: can u give me a link so I can follow what u are working on | 11:46 |
| noonedeadpunk | well, I didn't create one yet:) | 11:47 |
| recyclehero | I meant where u have identified the bug | 11:48 |
| recyclehero | but ok | 11:48 |
| *** macz_ has quit IRC | 11:48 | |
| noonedeadpunk | we don't trigger service restart when glance is ran with uwsgi | 11:49 |
| noonedeadpunk | I will push PR to cover that | 11:50 |
| openstackgerrit | Dmitriy Rabotyagov (noonedeadpunk) proposed openstack/openstack-ansible-os_glance master: Trigger uwsgi restart https://review.opendev.org/756681 | 11:53 |
| noonedeadpunk | recyclehero: ^ | 11:53 |
| openstackgerrit | Dmitriy Rabotyagov (noonedeadpunk) proposed openstack/openstack-ansible-os_aodh master: Trigger uwsgi restart https://review.opendev.org/756684 | 12:04 |
| *** shyam89 has joined #openstack-ansible | 12:04 | |
| *** shyamb has joined #openstack-ansible | 12:04 | |
| *** shyamb has quit IRC | 12:05 | |
| *** shyam89 has quit IRC | 12:05 | |
| *** shyamb has joined #openstack-ansible | 12:06 | |
| openstackgerrit | Dmitriy Rabotyagov (noonedeadpunk) proposed openstack/openstack-ansible-os_barbican master: Trigger uwsgi restart https://review.opendev.org/756686 | 12:08 |
| *** lkoranda has quit IRC | 12:10 | |
| openstackgerrit | Dmitriy Rabotyagov (noonedeadpunk) proposed openstack/openstack-ansible-os_barbican master: Cleanup stop handler https://review.opendev.org/756689 | 12:11 |
| *** cshen has quit IRC | 12:14 | |
| openstackgerrit | Dmitriy Rabotyagov (noonedeadpunk) proposed openstack/openstack-ansible-os_cinder master: Remove unecessary apache tasks https://review.opendev.org/756690 | 12:15 |
| *** rf0lc0 has joined #openstack-ansible | 12:22 | |
| recyclehero | noonedeadpunk: but the problem with my situation concerning CORS doesnt realte to this, right? I have restarted the galnce-api itself. I think uWSGI are just a medium in between | 12:22 |
| openstackgerrit | Merged openstack/openstack-ansible-os_nova master: Bind novncproxy host and port to defined variables https://review.opendev.org/753774 | 12:23 |
| noonedeadpunk | recyclehero: well yeah, it was regarding service not restarting | 12:23 |
| noonedeadpunk | regarding cors - I can only suggest hat you set it in some wrong way... | 12:23 |
| recyclehero | i will figure it out later. first I need to have a basic setup working | 12:25 |
| recyclehero | i did this btw http://paste.openstack.org/show/798849/ | 12:25 |
| recyclehero | at the end of user_variables.yml | 12:25 |
| noonedeadpunk | recyclehero: well, you could just provide `glance_cors_allowed_origin: https://200.200.200.110` as we have specific variable for it:) | 12:26 |
| noonedeadpunk | and you're accessing horizon with https://200.200.200.110? | 12:26 |
| recyclehero | yes | 12:26 |
| recyclehero | didnt see that, but they have the same result I guess | 12:27 |
| noonedeadpunk | yeah, they are | 12:28 |
| recyclehero | I went in the glance container and checked the config | 12:28 |
| *** shyamb has quit IRC | 12:28 | |
| recyclehero | noonedeadpunk: I am uploading to glance via legacy. I was wondering if both legacy and direct use haproxy and the mgmt-br, right? | 12:33 |
| noonedeadpunk | yeah | 12:33 |
| noonedeadpunk | the difference is that in case of direct, horizon doesn't take part and you streaming content directly to glance endpoint | 12:33 |
| noonedeadpunk | while with legacy you upload to horizon container, and then horizon uploads to glance | 12:34 |
| recyclehero | great, but for both situation ha-proxy is invloved and br-mgmt is used? | 12:37 |
| noonedeadpunk | yes | 12:37 |
| recyclehero | thank you | 12:37 |
| *** lkoranda has joined #openstack-ansible | 12:39 | |
| admin0 | noonedeadpunk, can it be left blank ? | 12:43 |
| admin0 | in dhcp_domain ? | 12:43 |
| noonedeadpunk | have no idea | 12:43 |
| openstackgerrit | Merged openstack/openstack-ansible-rabbitmq_server master: Require the use of community.rabbitmq ansible collection https://review.opendev.org/754657 | 12:50 |
| *** cshen has joined #openstack-ansible | 12:53 | |
| openstackgerrit | Dmitriy Rabotyagov (noonedeadpunk) proposed openstack/openstack-ansible master: Remove glance_registry from inventory https://review.opendev.org/756318 | 13:02 |
| openstackgerrit | Dmitriy Rabotyagov (noonedeadpunk) proposed openstack/openstack-ansible master: Switch integrated linters to focal https://review.opendev.org/755759 | 13:03 |
| openstackgerrit | Dmitriy Rabotyagov (noonedeadpunk) proposed openstack/openstack-ansible master: Fix octavia tempest tests https://review.opendev.org/755737 | 13:03 |
| *** shyamb has joined #openstack-ansible | 13:05 | |
| *** shyam89 has joined #openstack-ansible | 13:09 | |
| *** shyamb has quit IRC | 13:12 | |
| mgariepy | https://zuul.opendev.org/t/openstack/build/3b9ab58deab3468696dd8618c439b555/log/job-output.txt#1569-1612 | 13:16 |
| mgariepy | some error on the ara generation dmsimard ^? | 13:17 |
| *** vakuznet has quit IRC | 13:19 | |
| mgariepy | galaxy.ansible.com is having issues ? | 13:23 |
| mgariepy | https://zuul.opendev.org/t/openstack/build/3b9ab58deab3468696dd8618c439b555/log/job-output.txt#1528-1529 | 13:23 |
| dmsimard | mgariepy: it looks like it's trying to generate reports but there's no data (or database) | 13:27 |
| dmsimard | I think the real issue is slightly above with the timed out download, yeah | 13:28 |
| *** macz_ has joined #openstack-ansible | 13:31 | |
| *** macz_ has quit IRC | 13:36 | |
| *** shyam89 has quit IRC | 13:40 | |
| *** nurdie has joined #openstack-ansible | 13:44 | |
| jrosser | looks like ansible galaxy is really broken somehow | 14:01 |
| *** lkoranda has quit IRC | 14:20 | |
| openstackgerrit | Dmitriy Rabotyagov (noonedeadpunk) proposed openstack/openstack-ansible-os_cinder master: Trigger uwsgi restart https://review.opendev.org/756719 | 14:20 |
| noonedeadpunk | jrosser: yeah it is which is pretty disapointing... | 14:21 |
| openstackgerrit | Dmitriy Rabotyagov (noonedeadpunk) proposed openstack/openstack-ansible-os_cloudkitty master: Remove unused api handler https://review.opendev.org/756721 | 14:27 |
| *** soutr has joined #openstack-ansible | 14:31 | |
| openstackgerrit | Dmitriy Rabotyagov (noonedeadpunk) proposed openstack/openstack-ansible-os_heat master: Trigger uwsgi restart https://review.opendev.org/756724 | 14:33 |
| *** MickyMan77 has joined #openstack-ansible | 14:35 | |
| soutr | To retrieve the value of password from the below file:————clouds: dal-prod-openstackv1: auth: username: ‘user’ password: ‘XXX’ auth_url: 'https://authurl:5000/v3'——————I am using the below code——————- - name: Parse auth file set_fact: auth_content: "{{ lookup('file', ‘user.ini') | | 14:39 |
| soutr | from_yaml }}" delegate_to: localhost- name: Obtain password set_fact: cluster_pass: "{{ auth_content['clouds’][‘dal-prod-openstackv1’][‘auth']['password'] }}"———————Is there any way I can combine the 2 into one? | 14:39 |
| openstackgerrit | Dmitriy Rabotyagov (noonedeadpunk) proposed openstack/openstack-ansible-os_ironic master: Trigger uwsgi restart https://review.opendev.org/756780 | 14:41 |
| noonedeadpunk | soutr: um, I'm not sure why you need this tbh. as in case of OSA you have all passwords already defined as variables. Even if not, openstack ansible modules use clouds.yaml nicely and you can just specify cloud name there | 14:45 |
| *** soutr has quit IRC | 14:50 | |
| openstackgerrit | Dmitriy Rabotyagov (noonedeadpunk) proposed openstack/openstack-ansible-os_nova master: Trigger uwsgi restart https://review.opendev.org/756861 | 14:50 |
| noonedeadpunk | jrosser: btw seems functional tests are somehow broken in train:( | 14:51 |
| noonedeadpunk | specifically tox and py2 | 14:51 |
| *** soutr has joined #openstack-ansible | 14:52 | |
| soutr | noonedeadpunk I am trying to automate the creation of projects and users using ansible. The creation of users and projects will be done as openstack admin user. I am saving the the auth details for users and projects. And then parsing the file to retrieve the password, so that I can use them with os_user module. | 14:54 |
| soutr | - name: Add openstack users os_user: domain: "{{ item.user_domain }}" default_project: "{{ item.default_project }}" name: "{{ item.name }}" description: "{{ item.description }}" password: "{{ item.password }}"# update_password: always environment: "{{ env }}" with_items: "{{ openstack_common.users }}" | 14:54 |
| noonedeadpunk | soutr: well, you can avoid retrieving password for this for sure. you may look at example https://opendev.org/openstack/openstack-ansible-tests/src/branch/stable/train/sync/tasks/service_setup.yml | 14:55 |
| noonedeadpunk | so the thing that needs to be done - this file should be put in /etc/openstack/clouds.yaml or ~/.config/openstack/clouds.yaml ie https://docs.openstack.org/python-openstackclient/ussuri/configuration/index.html | 14:56 |
| noonedeadpunk | and ansible modules will just pick them, when you provide cloud name as parameter | 14:57 |
| noonedeadpunk | `cloud: default` in our example | 14:57 |
| noonedeadpunk | for you it will be `dal-prod-openstackv1` | 14:57 |
| openstackgerrit | Dmitriy Rabotyagov (noonedeadpunk) proposed openstack/openstack-ansible-os_panko master: Cleanup apache tasks https://review.opendev.org/756864 | 14:59 |
| openstackgerrit | Merged openstack/openstack-ansible-os_mistral master: Remove support for lxc2 config keys https://review.opendev.org/756249 | 15:00 |
| soutr | Thankyou noonedeadpunk . I will take a look in those documents. | 15:00 |
| noonedeadpunk | you might need them if you will be using so specific service clients which don't read clouds.yaml, but I can't think about which are not currently | 15:02 |
| openstackgerrit | Dmitriy Rabotyagov (noonedeadpunk) proposed openstack/openstack-ansible-os_placement master: Trigger service restart https://review.opendev.org/756866 | 15:05 |
| jrosser | noonedeadpunk: do yo have a link to the failing train tests? | 15:05 |
| noonedeadpunk | https://review.opendev.org/#/c/690342/ | 15:06 |
| noonedeadpunk | btw I dunno how we really lived without service restart on config change.... | 15:06 |
| jrosser | noonedeadpunk: internet seems to say virtualenv 20 is maybe the problem | 15:16 |
| noonedeadpunk | yeah, I googled | 15:17 |
| jrosser | we are still running python2 there for those functional tests i think | 15:17 |
| noonedeadpunk | but I can't recall we've bumped it since release? | 15:17 |
| noonedeadpunk | we do for sure | 15:17 |
| jrosser | maybe not on train i guess | 15:17 |
| noonedeadpunk | and tox should be bumped as well... | 15:18 |
| noonedeadpunk | I mean tox should be fixed | 15:18 |
| jrosser | where is this set for the openstack-ansible-tests repo..... | 15:19 |
| noonedeadpunk | oh... | 15:19 |
| *** macz_ has joined #openstack-ansible | 15:19 | |
| noonedeadpunk | I think we take https://opendev.org/openstack/openstack-ansible/src/branch/master/global-requirement-pins.txt but there's neither tox nor virtualenv | 15:20 |
| noonedeadpunk | well... we don't bump tox | 15:22 |
| noonedeadpunk | https://opendev.org/openstack/openstack-ansible-tests/src/branch/stable/train/run_tests_common.sh#L85 | 15:22 |
| *** soutr has quit IRC | 15:24 | |
| *** macz_ has quit IRC | 15:24 | |
| *** macz_ has joined #openstack-ansible | 15:29 | |
| masterpe | I'm upgrading rabbitmq from Stein to Train, see that it brings down every rabbitmq node for the upgrade. Is that by design? | 15:30 |
| noonedeadpunk | masterpe: it should bring one node at a time | 15:31 |
| noonedeadpunk | as we have serial:1 here https://opendev.org/openstack/openstack-ansible/src/branch/master/playbooks/rabbitmq-install.yml#L24 | 15:31 |
| masterpe | I use the -e 'rabbitmq_upgrade=true' -e 'placement_migrate_flag=true' options and I followed the rabbitmqctl cluster_status on every node | 15:32 |
| noonedeadpunk | ah, well, yes, there might be short downtime of all members when upgrading first | 15:33 |
| noonedeadpunk | as we need to stop all instances extept the single one, and while upgrading it it might be restarted, during this time period rabbit might be down | 15:34 |
| noonedeadpunk | https://www.rabbitmq.com/upgrade.html#rabbitmq-cluster-configuration | 15:35 |
| masterpe | thanks | 15:36 |
| noonedeadpunk | but it should be really ~1min downtime iirc | 15:36 |
| noonedeadpunk | but probably you're right and we can do rolling upgrades nowadays since we're on 3.8 | 15:37 |
| masterpe | downtime was a little bit more then one minute but i did not clock it. | 15:38 |
| jrosser | noonedeadpunk: maybe we need to pin virtualenv < 20 on train | 15:42 |
| jrosser | just tons of complexity with where to put that | 15:42 |
| noonedeadpunk | and we had pin in python_venv_build only? | 15:43 |
| jrosser | also we bring a *very* old version of pip in the functional tests | 15:43 |
| jrosser | i think the error you see is just in getting the tox environment to work though | 15:44 |
| jrosser | well before anything actually to do with ansible stuff | 15:44 |
| noonedeadpunk | it is | 15:45 |
| noonedeadpunk | I just trying to recall where I saw virtualenv binding :) | 15:45 |
| jrosser | kind of unrelated i think but https://bootstrap.pypa.io/3.4/ would get us to pip version 19.x.y | 15:47 |
| *** gyee has joined #openstack-ansible | 16:00 | |
| *** rpittau is now known as rpittau|afk | 16:01 | |
| noonedeadpunk | jrosser: virtualenv==16.7.10 works nicely | 16:02 |
| jrosser | noonedeadpunk: i think we are doing identical things :) http://paste.openstack.org/show/798875/ | 16:03 |
| jrosser | i get the same version 'Successfully installed virtualenv-16.7.10' | 16:04 |
| noonedeadpunk | pip version is really unrelated | 16:04 |
| jrosser | yeah, was just interested to see what happened | 16:05 |
| noonedeadpunk | but yeah, I like your solution | 16:05 |
| noonedeadpunk | but I didn't get what exactly installs virtualenv | 16:05 |
| noonedeadpunk | dependency to tox? | 16:05 |
| jrosser | it's a dependancy of tox | 16:05 |
| noonedeadpunk | yeah, I see) | 16:05 |
| jrosser | https://zuul.opendev.org/t/openstack/build/5c5a7a206ef346ba9e0fdc9949b42f19/log/job-output.txt#1227 | 16:06 |
| noonedeadpunk | well, I'd say let's bump tox as well? | 16:06 |
| noonedeadpunk | as they're pretty close with >16.0.0 <20.0.0 | 16:07 |
| noonedeadpunk | it's kind of one major release | 16:07 |
| jrosser | looks like latest tox still supports python2 | 16:08 |
| noonedeadpunk | it is | 16:08 |
| noonedeadpunk | sugggest return to this once it won't? | 16:08 |
| jrosser | if we want to bump tox version that sounds like something to do on master and backport | 16:09 |
| jrosser | fixing virtualenv pin we should just do straight to train right now | 16:09 |
| noonedeadpunk | well, if we want to bump it in master, we need to do that in test-requirements.txt and include it as contraints | 16:10 |
| noonedeadpunk | yeah, ok, let's bump virtualenv, backport to stein | 16:10 |
| noonedeadpunk | and return to this later :p | 16:10 |
| jrosser | you happy we do it like my diff? we don't have a constraints file really in the tests repo | 16:11 |
| noonedeadpunk | I'm pretty happy | 16:13 |
| * jrosser makes patch | 16:14 | |
| openstackgerrit | Jonathan Rosser proposed openstack/openstack-ansible-tests stable/train: Pin virtualenv<20 for python2 functional tests https://review.opendev.org/756883 | 16:14 |
| openstackgerrit | Dmitriy Rabotyagov (noonedeadpunk) proposed openstack/openstack-ansible-lxc_hosts stable/train: Updated from OpenStack Ansible Tests https://review.opendev.org/690342 | 16:18 |
| noonedeadpunk | ah, depends on doesn't for for tests repo | 16:18 |
| jrosser | CI seems wedged up for this https://review.opendev.org/#/c/756587/ looks like the centos-8 job is stuck waiting to timeout | 16:25 |
| jrosser | well there is a lesson - no sooner do i say that, the job passes | 16:27 |
| noonedeadpunk | heh:) | 16:28 |
| openstackgerrit | Merged openstack/openstack-ansible-lxc_hosts master: Include libpython and rsync for centos in lxc base image https://review.opendev.org/756587 | 16:47 |
| *** cshen has quit IRC | 17:19 | |
| *** maharg101 has quit IRC | 17:24 | |
| *** cshen has joined #openstack-ansible | 17:25 | |
| *** MickyMan77 has quit IRC | 17:28 | |
| *** MickyMan77 has joined #openstack-ansible | 17:29 | |
| *** cshen has quit IRC | 17:29 | |
| openstackgerrit | Jonathan Rosser proposed openstack/openstack-ansible master: Update integrated ansible-lint rules https://review.opendev.org/756121 | 17:31 |
| openstackgerrit | Jonathan Rosser proposed openstack/openstack-ansible master: Use nodepool epel mirror in CI for systemd-networkd package https://review.opendev.org/754706 | 17:31 |
| openstackgerrit | Jonathan Rosser proposed openstack/openstack-ansible master: Deprecate os_congress role https://review.opendev.org/742521 | 17:31 |
| openstackgerrit | Jonathan Rosser proposed openstack/openstack-ansible master: Remove "when" statement from vars_prompt https://review.opendev.org/755824 | 17:31 |
| openstackgerrit | Jonathan Rosser proposed openstack/openstack-ansible master: Bump SHAs for master https://review.opendev.org/755973 | 17:32 |
| openstackgerrit | Jonathan Rosser proposed openstack/openstack-ansible master: Added Openstack Adjutant role deployment https://review.opendev.org/756310 | 17:32 |
| *** fghaas has joined #openstack-ansible | 17:41 | |
| *** fghaas has quit IRC | 17:44 | |
| recyclehero | hey guys, sometimes I get 504 when I want to login. and this implentation is slower than my queens manual deployment. | 18:04 |
| recyclehero | I remember noonedeadpunk told me before about adjusting number of sth | 18:05 |
| recyclehero | that sth is processes or threads or both? | 18:05 |
| recyclehero | the system is mostly used by me, what should I consider? | 18:06 |
| recyclehero | another housekeeping q for me, is heat a totally optional service, can I remove neutron-metering-agent? | 18:06 |
| recyclehero | should I focus on ram or mostly or cpu for controller? | 18:07 |
| recyclehero | I use NFS cinder volume is better to reside on infra1 or compute1, give me some reasoning for this one. | 18:08 |
| recyclehero | uWSGI | 18:08 |
| *** maharg101 has joined #openstack-ansible | 18:24 | |
| *** maharg101 has quit IRC | 18:31 | |
| openstackgerrit | Georgina Shippey proposed openstack/openstack-ansible-galera_server master: Ability to take mariadb backups using mariabackup https://review.opendev.org/755261 | 18:40 |
| *** SecOpsNinja has left #openstack-ansible | 18:48 | |
| djhankb | recyclehero: I had run into the 504 issue quite a bit back when I was doing my first test cluster on basic Desktop-class hardware. I set some of the services to use less "workers" to try and save on resources which helped a bit. You can add some stuff from here http://paste.openstack.org/show/798881/ to user_variables.yml | 18:49 |
| recyclehero | djhankb: thanks I needed something to look for. u set this for a 8 core cpu? | 18:52 |
| recyclehero | in general whats the reasoning for chosing threads and processes? or better process to thread ratio? | 18:55 |
| djhankb | recyclehero I think so - they were i5 desktop machines, IIRC | 18:55 |
| recyclehero | I think I5 has 4 cores though. but I am going for I5 anyway | 18:55 |
| *** spatel has joined #openstack-ansible | 18:56 | |
| *** MickyMan77 has quit IRC | 18:56 | |
| *** andrewbonney has quit IRC | 19:00 | |
| recyclehero | this is goodhttps://www.mathworks.com/help/parallel-computing/choose-between-thread-based-and-process-based-environments.html | 19:01 |
| *** miloa has quit IRC | 19:21 | |
| noonedeadpunk | damn it, 756244 failed in gates (╯°□°)╯︵ ┻━┻ | 19:29 |
| jrosser | oh dear | 19:36 |
| *** d34dh0r53 has quit IRC | 19:49 | |
| *** d34dh0r53 has joined #openstack-ansible | 19:52 | |
| *** nurdie has quit IRC | 20:00 | |
| *** nurdie has joined #openstack-ansible | 20:39 | |
| *** nurdie has quit IRC | 20:45 | |
| *** spatel has quit IRC | 20:57 | |
| recyclehero | i followed this variable for cors setting of glance | 21:05 |
| recyclehero | glance_cors_allowed_origin: "{{ (glance_show_multiple_locations | bool) | ternary(openstack_service_publicuri_proto | default('http') + '://' + external_lb_vip_address, None) }}" | 21:05 |
| recyclehero | I found them not good | 21:05 |
| recyclehero | glance_show_multiple_locations: "{{ glance_default_store == 'rbd' }}" | 21:06 |
| recyclehero | it means if you have a default file store you cant upload from horizon! | 21:06 |
| recyclehero | I will set glance_show-multiple_locations: true and leave the rest as is. | 21:09 |
| recyclehero | cause you see the cors was complaining about the port mismatch. | 21:10 |
| *** cshen has joined #openstack-ansible | 21:16 | |
| *** cshen has quit IRC | 21:21 | |
| *** d34dh0r53 has quit IRC | 21:28 | |
| *** d34dh0r53 has joined #openstack-ansible | 21:28 | |
| *** rf0lc0 has quit IRC | 21:30 | |
| *** MickyMan77 has joined #openstack-ansible | 21:31 | |
| *** maharg101 has joined #openstack-ansible | 22:28 | |
| *** maharg101 has quit IRC | 22:37 | |
| *** nurdie has joined #openstack-ansible | 22:40 | |
| *** nurdie has quit IRC | 22:45 | |
| *** tosky has quit IRC | 22:59 | |
| *** pmannidi has joined #openstack-ansible | 23:23 | |
| *** macz_ has quit IRC | 23:32 | |
| *** rfolco has joined #openstack-ansible | 23:38 | |
| *** rfolco has quit IRC | 23:40 | |
| *** rfolco has joined #openstack-ansible | 23:41 | |
| *** rfolco has quit IRC | 23:45 | |
| *** gyee has quit IRC | 23:50 | |
Generated by irclog2html.py 2.17.2 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!