| *** rfolco has joined #openstack-ansible | 00:02 | |
| *** tinwood has quit IRC | 00:03 | |
| *** tinwood has joined #openstack-ansible | 00:03 | |
| *** Muran has joined #openstack-ansible | 00:16 | |
| *** Muran has quit IRC | 00:20 | |
| *** gyee has quit IRC | 00:24 | |
| *** rfolco has quit IRC | 00:30 | |
| *** cshen has joined #openstack-ansible | 00:31 | |
| *** cshen has quit IRC | 00:36 | |
| *** MickyMan77 has quit IRC | 01:13 | |
| *** MickyMan77 has joined #openstack-ansible | 01:14 | |
| *** MickyMan77 has quit IRC | 01:23 | |
| *** MickyMan77 has joined #openstack-ansible | 01:59 | |
| *** MickyMan77 has quit IRC | 02:08 | |
| *** rfolco has joined #openstack-ansible | 02:14 | |
| *** rfolco has quit IRC | 02:18 | |
| *** mcarden has quit IRC | 02:18 | |
| *** karanveersingh56 has joined #openstack-ansible | 02:20 | |
| *** rh-jelabarre has quit IRC | 02:20 | |
| *** rh-jelabarre has joined #openstack-ansible | 02:21 | |
| karanveersingh56 | Hi guys , facing issues in openstack deployment | 02:21 |
|---|---|---|
| karanveersingh56 | can anybody share a openstack_user_config.yml format , is any other file i have to make changes | 02:21 |
| karanveersingh56 | root@node3:/opt/openstack-ansible/playbooks# openstack-ansible setup-hosts.yml -vvvv --syntax-check | 02:22 |
| karanveersingh56 | '/etc/ansible/roles/plugins/library', '/etc/ansible/roles/ceph-ansible/library'] | 02:22 |
| karanveersingh56 | setting up inventory plugins | 02:22 |
| karanveersingh56 | as it did not pass its verify_file() method | 02:22 |
| karanveersingh56 | execution error: | 02:22 |
| karanveersingh56 | 169, in parse | 02:22 |
| karanveersingh56 | "/opt/ansible-runtime/lib/python3.8/site-packages/ansible/inventory/manager.py", line 289, in parse_source | 02:22 |
| karanveersingh56 | please give some pointers so that i can proceed | 02:22 |
| *** cshen has joined #openstack-ansible | 02:32 | |
| karanveersingh56 | can anybody share a openstack_user_config.yml format , is any other file i have to make changes | 02:32 |
| karanveersingh56 | --------------------------------------------------------------------------------------------------------------- | 02:32 |
| karanveersingh56 | root@node3:/opt/openstack-ansible/playbooks# openstack-ansible setup-hosts.yml -vvvv --syntax-check | 02:32 |
| karanveersingh56 | '/etc/ansible/roles/plugins/library', '/etc/ansible/roles/ceph-ansible/library'] | 02:32 |
| karanveersingh56 | setting up inventory plugins | 02:32 |
| karanveersingh56 | as it did not pass its verify_file() method | 02:32 |
| karanveersingh56 | execution error: | 02:32 |
| karanveersingh56 | 169, in parse | 02:32 |
| karanveersingh56 | "/opt/ansible-runtime/lib/python3.8/site-packages/ansible/inventory/manager.py", line 289, in parse_source | 02:33 |
| *** cshen has quit IRC | 02:36 | |
| *** MickyMan77 has joined #openstack-ansible | 02:36 | |
| *** cshen has joined #openstack-ansible | 02:41 | |
| *** MickyMan77 has quit IRC | 02:44 | |
| *** cshen has quit IRC | 02:46 | |
| *** rfolco has joined #openstack-ansible | 02:54 | |
| *** rfolco has quit IRC | 02:59 | |
| *** karanveersingh56 has quit IRC | 03:04 | |
| *** MickyMan77 has joined #openstack-ansible | 03:15 | |
| *** recyclehero has quit IRC | 03:36 | |
| *** MickyMan77 has quit IRC | 04:19 | |
| *** evrardjp has quit IRC | 04:33 | |
| *** evrardjp has joined #openstack-ansible | 04:33 | |
| *** cshen has joined #openstack-ansible | 04:41 | |
| *** cshen has quit IRC | 04:46 | |
| *** nurdie has quit IRC | 05:10 | |
| *** karanveersingh56 has joined #openstack-ansible | 05:11 | |
| karanveersingh56 | anyone , need help on openstack ansible deployment | 05:12 |
| karanveersingh56 | i am facing issue in the initial phase | 05:13 |
| *** MickyMan77 has joined #openstack-ansible | 05:15 | |
| *** MickyMan77 has quit IRC | 05:24 | |
| *** cshen has joined #openstack-ansible | 05:30 | |
| *** cshen has quit IRC | 05:35 | |
| *** jbadiapa has joined #openstack-ansible | 05:46 | |
| *** recyclehero has joined #openstack-ansible | 05:53 | |
| *** MickyMan77 has joined #openstack-ansible | 05:57 | |
| *** rh-jelabarre has quit IRC | 06:05 | |
| *** MickyMan77 has quit IRC | 06:05 | |
| *** rh-jelabarre has joined #openstack-ansible | 06:06 | |
| *** recyclehero has quit IRC | 06:07 | |
| *** macz_ has joined #openstack-ansible | 06:19 | |
| noonedeadpunk | morning | 06:19 |
| noonedeadpunk | returning to our biggest issue at the moment - ubuntu images | 06:20 |
| noonedeadpunk | I think we shouldn't use versioned images now once they drop them | 06:21 |
| noonedeadpunk | as it means that one day we will face with that again and it will be pretty soon actually | 06:21 |
| noonedeadpunk | when all our releases will be broken due to this | 06:21 |
| noonedeadpunk | so choice is between lxd image which is not really base, or daily build of base | 06:22 |
| *** macz_ has quit IRC | 06:24 | |
| *** noonedeadpunk has quit IRC | 06:31 | |
| *** cshen has joined #openstack-ansible | 06:32 | |
| *** cshen has quit IRC | 06:36 | |
| *** noonedeadpunk has joined #openstack-ansible | 06:39 | |
| noonedeadpunk | ah, well, or write soma nasty thing to pick up always the latest release | 06:40 |
| *** pto has quit IRC | 06:42 | |
| *** pto has joined #openstack-ansible | 06:43 | |
| *** recyclehero has joined #openstack-ansible | 06:44 | |
| *** cshen has joined #openstack-ansible | 06:44 | |
| karanveersingh56 | Hi Dmitriy | 06:46 |
| karanveersingh56 | https://bugs.launchpad.net/openstack-ansible/+bug/1900884 | 06:46 |
| openstack | Launchpad bug 1900884 in openstack-ansible "ubuntu-base-20.04-base-amd64.tar.gz is missing" [Undecided,In progress] - Assigned to Dmitriy Rabotyagov (noonedeadpunk) | 06:46 |
| noonedeadpunk | karanveersingh56: we were trying to cover that with https://review.opendev.org/#/c/758992/ | 06:47 |
| noonedeadpunk | but it's not final | 06:47 |
| noonedeadpunk | issue here that ubuntu has dropped all images except latest ones, which it used to store | 06:47 |
| karanveersingh56 | correct , I was trying to deploy the stack | 06:49 |
| *** cshen has quit IRC | 06:49 | |
| karanveersingh56 | any workaround where I can change it to ubuntu-base-20.04.1-base-amd64.tar.gz | 06:49 |
| noonedeadpunk | yep, you can override it with lxc_hosts_container_image_url: http://cdimage.ubuntu.com/ubuntu-base/releases/20.04/release/ubuntu-base-20.04.1-base-{{ lxc_cache_map.arch }}.tar.gz | 06:50 |
| karanveersingh56 | which file I can find this variable | 06:51 |
| noonedeadpunk | currently it's here https://opendev.org/openstack/openstack-ansible-lxc_hosts/src/branch/master/vars/ubuntu-20.04.yml#L16 | 06:53 |
| noonedeadpunk | bu I mean you should put it in user_variables and that's it | 06:53 |
| noonedeadpunk | *you can put | 06:53 |
| karanveersingh56 | let me give a try , thanks buddy | 06:54 |
| *** rpittau|afk is now known as rpittau | 06:54 | |
| *** jawad_axd has joined #openstack-ansible | 06:54 | |
| *** recyclehero has quit IRC | 07:02 | |
| *** recyclehero has joined #openstack-ansible | 07:03 | |
| *** Muran has joined #openstack-ansible | 07:03 | |
| jrosser | maybe stat the file and only download the latest if it's not missing | 07:04 |
| *** karanveersingh56 has quit IRC | 07:04 | |
| *** pto has quit IRC | 07:04 | |
| *** pto has joined #openstack-ansible | 07:05 | |
| jrosser | we already have # Set this option to true to pull a new cached image. | 07:05 |
| jrosser | lxc_image_cache_refresh: false | 07:05 |
| *** pto has quit IRC | 07:06 | |
| *** karanveersingh56 has joined #openstack-ansible | 07:06 | |
| *** pto has joined #openstack-ansible | 07:06 | |
| *** pto has quit IRC | 07:09 | |
| *** pto has joined #openstack-ansible | 07:10 | |
| noonedeadpunk | yeah but we set explicitly it's tll | 07:10 |
| noonedeadpunk | *ttl | 07:10 |
| noonedeadpunk | so you mean drop it and storebase image until lxc_image_cache_refresh is explicitly true? | 07:10 |
| jrosser | something like that | 07:11 |
| noonedeadpunk | I started thinking about the way to get available image out of checksums | 07:11 |
| jrosser | and we can add that maybe to minor upgrade instructions | 07:11 |
| noonedeadpunk | ie http://cdimage.ubuntu.com/ubuntu-base/releases/20.04/release/SHA256SUMS | 07:11 |
| jrosser | ah yes | 07:11 |
| recyclehero | morning | 07:12 |
| noonedeadpunk | but probably your option is good as well and the way easier | 07:12 |
| noonedeadpunk | as ofc if we get checksum it's easy for us to verify integrity of the imge | 07:13 |
| noonedeadpunk | but overriding of the source would be pretty much painful then | 07:13 |
| jrosser | I do that now | 07:13 |
| jrosser | have some offline installation | 07:13 |
| noonedeadpunk | and I guess you don't store SHAs?:) | 07:14 |
| noonedeadpunk | what makes things worse that this should be something backportable | 07:15 |
| noonedeadpunk | so option with removal of image ttl seems better here | 07:16 |
| jrosser | I have to be away for a bit, back in hour or so | 07:17 |
| noonedeadpunk | sure | 07:17 |
| karanveersingh56 | Guys , just a basic Q , do we need to create these bridges manually and wat all are mandatory bridges | 07:18 |
| karanveersingh56 | br-mgmt , br-storage , br-vxlan , br-vlan | 07:18 |
| noonedeadpunk | yep, we don't touch underlying networks, so they need to be created manualy before the deployment | 07:19 |
| noonedeadpunk | I think you named all of them if you don't deploy octavia or designate as they may require additional ones | 07:20 |
| karanveersingh56 | yea , I did expose these in user config yaml but haven't created so all the containers failed | 07:22 |
| karanveersingh56 | lxc-start node1_rabbit_mq_container-67075c91 20201022071431.293 ERROR cgfsng - cgroups/cgfsng.c:mkdir_eexist_on_last:1143 - File exists - Failed to create directory "/sys/fs/cgroup/cpuset//lxc.monitor.node1_rabbit_mq_container-67075c91" | 07:22 |
| karanveersingh56 | cgroups/cgfsng.c:mkdir_eexist_on_last:1143 - File exists - Failed to create directory "/sys/fs/cgroup/cpuset//lxc.payload.node1_rabbit_mq_container-67075c91" | 07:22 |
| karanveersingh56 | 20201022071431.304 ERROR network - network.c:lxc_ovs_attach_bridge:2387 - Failed to attach "67075c91_eno1" to openvswitch bridge "br-mgmt": lxc-start: node1_rabbit_mq_container-67075c91: | 07:22 |
| karanveersingh56 | "67075c91_eno1" to bridge "br-mgmt" | 07:22 |
| karanveersingh56 | network | 07:22 |
| karanveersingh56 | lxc-start node1_rabbit_mq_container-67075c91 20201022071431.349 ERROR lxc_start - tools/lxc_start.c:main:311 - To get more details, run the container in foreground mode | 07:22 |
| karanveersingh56 | and --logpriority options | 07:22 |
| karanveersingh56 | can I get any sample user config file as well , which I can take as a reference , like minimalistic openstack to start with | 07:23 |
| *** andrewbonney has joined #openstack-ansible | 07:24 | |
| *** pto has quit IRC | 07:30 | |
| *** pto has joined #openstack-ansible | 07:31 | |
| *** cshen has joined #openstack-ansible | 07:32 | |
| *** pto has quit IRC | 07:33 | |
| *** pto has joined #openstack-ansible | 07:33 | |
| *** karanveersingh56 has quit IRC | 07:36 | |
| *** tosky has joined #openstack-ansible | 07:44 | |
| *** pto has quit IRC | 07:49 | |
| *** pto has joined #openstack-ansible | 07:50 | |
| *** pto has quit IRC | 07:53 | |
| *** pto has joined #openstack-ansible | 07:57 | |
| *** macz_ has joined #openstack-ansible | 08:07 | |
| *** macz_ has quit IRC | 08:12 | |
| *** pto_ has joined #openstack-ansible | 08:17 | |
| *** pto_ has quit IRC | 08:18 | |
| *** pto_ has joined #openstack-ansible | 08:19 | |
| *** pto_ has quit IRC | 08:20 | |
| *** pto_ has joined #openstack-ansible | 08:20 | |
| *** pto has quit IRC | 08:20 | |
| *** nurdie has joined #openstack-ansible | 08:27 | |
| *** sshnaidm|afk is now known as sshnaidm|rover | 08:29 | |
| *** nurdie has quit IRC | 08:30 | |
| *** pto_ has quit IRC | 08:38 | |
| *** pto has joined #openstack-ansible | 08:39 | |
| *** nurdie has joined #openstack-ansible | 08:41 | |
| *** klamath_atx has joined #openstack-ansible | 08:43 | |
| *** nurdie has quit IRC | 08:45 | |
| *** pto has quit IRC | 08:46 | |
| *** pto has joined #openstack-ansible | 08:46 | |
| *** pto has quit IRC | 08:52 | |
| *** pto has joined #openstack-ansible | 08:53 | |
| *** gshippey has joined #openstack-ansible | 08:59 | |
| recyclehero | in which cases the interfaces should be unnumbered? | 09:03 |
| *** pto has quit IRC | 09:03 | |
| *** pto has joined #openstack-ansible | 09:04 | |
| recyclehero | in particular can an interface be presented for physical flat network mapping via host_bind_interface and also get an address to be used as external_lb_address | 09:08 |
| *** pto has quit IRC | 09:09 | |
| *** pto has joined #openstack-ansible | 09:10 | |
| *** pto has quit IRC | 09:21 | |
| noonedeadpunk | jrosser: oh, well, maybe we should just return to legacy url fetch process | 09:21 |
| *** pto has joined #openstack-ansible | 09:22 | |
| jrosser | let me just see what we do right now for this | 09:24 |
| noonedeadpunk | https://uk.images.linuxcontainers.org/images/ubuntu/focal/amd64/default/20201022_07:42/ | 09:25 |
| noonedeadpunk | but it's not really base image as well | 09:25 |
| openstackgerrit | Dmitriy Rabotyagov (noonedeadpunk) proposed openstack/openstack-ansible-lxc_hosts master: Determine latest base image available https://review.opendev.org/759229 | 09:26 |
| noonedeadpunk | jrosser: that's what I come up with | 09:27 |
| noonedeadpunk | but honestly not sure in this solution - it really looks like this legacy part | 09:27 |
| noonedeadpunk | except legacy part is more reliable | 09:27 |
| jrosser | right currently i override lxc_hosts_container_image_url to some local URL | 09:29 |
| noonedeadpunk | Well I think the biggest question now is what image do we want to take as base one | 09:31 |
| noonedeadpunk | well, probably worth creating etherpad | 09:31 |
| * jrosser meetings all morning :( | 09:31 | |
| noonedeadpunk | ah, I see | 09:31 |
| noonedeadpunk | https://etherpad.opendev.org/p/osa-ubuntu-image-source | 09:45 |
| *** pto has quit IRC | 09:57 | |
| *** pto has joined #openstack-ansible | 09:58 | |
| *** pto has quit IRC | 09:58 | |
| *** pto has joined #openstack-ansible | 09:59 | |
| *** pto has joined #openstack-ansible | 10:00 | |
| *** pto has quit IRC | 10:01 | |
| *** pto has joined #openstack-ansible | 10:01 | |
| *** pto has quit IRC | 10:05 | |
| *** pto has joined #openstack-ansible | 10:06 | |
| *** pto has quit IRC | 10:42 | |
| *** pto has joined #openstack-ansible | 10:42 | |
| *** mgariepy has quit IRC | 11:04 | |
| *** pto has quit IRC | 11:08 | |
| *** pto has joined #openstack-ansible | 11:09 | |
| recyclehero | the only way I can think of wihtout adding a nic to the picture is to hapxory_ssl: flase | 11:09 |
| recyclehero | I dont hava an ip to assign for external_lb | 11:09 |
| recyclehero | the provider netowrk in infra one is used by the linux bridge agent so I shouldnt assign a address to it | 11:10 |
| *** jawad_axd has quit IRC | 11:11 | |
| recyclehero | but when I did this and rna haproxy playbook I am getting "regerer is insecure while host is secure " | 11:12 |
| *** pto has quit IRC | 11:14 | |
| *** pto has joined #openstack-ansible | 11:15 | |
| recyclehero | aha I serached haproxy_ssl in other roles so I should redplooy those too as it affect them | 11:23 |
| recyclehero | but I wish It was a way that I could have an ip address on the external in infra1 | 11:23 |
| *** pto has quit IRC | 11:35 | |
| *** pto has joined #openstack-ansible | 11:36 | |
| recyclehero | cd /var/log | 11:37 |
| recyclehero | ls | 11:37 |
| * recyclehero feels stupid | 11:37 | |
| jrosser | recyclehero: i think you should take a look at how this is done in the AIO where the whole deployment is on one host | 11:43 |
| jrosser | there is an extra interface eth12 which is given to neutron, and wired with a veth to the 'flat network' | 11:43 |
| jrosser | that lets you achieve what you are describing | 11:43 |
| *** spatel has joined #openstack-ansible | 11:46 | |
| *** spatel has quit IRC | 11:51 | |
| *** mgariepy has joined #openstack-ansible | 11:59 | |
| *** nurdie has joined #openstack-ansible | 12:01 | |
| *** pto has quit IRC | 12:06 | |
| *** pto_ has joined #openstack-ansible | 12:06 | |
| *** nurdie has quit IRC | 12:06 | |
| *** pto_ has quit IRC | 12:10 | |
| *** pto has joined #openstack-ansible | 12:10 | |
| *** jawad_axd has joined #openstack-ansible | 12:19 | |
| *** rfolco has joined #openstack-ansible | 12:20 | |
| *** pto has quit IRC | 12:27 | |
| *** pto has joined #openstack-ansible | 12:28 | |
| *** nurdie has joined #openstack-ansible | 12:36 | |
| *** nurdie has quit IRC | 12:53 | |
| *** pto has quit IRC | 13:01 | |
| *** pto has joined #openstack-ansible | 13:03 | |
| *** sep has quit IRC | 13:10 | |
| dmsimard | ara 1.5.2 is out with the handler recording and performance improvements | 13:18 |
| dmsimard | let me know if you see any problems, thanks :) | 13:19 |
| *** nurdie has joined #openstack-ansible | 13:24 | |
| *** jawad_axd has quit IRC | 13:28 | |
| noonedeadpunk | do we install just latest?:) | 13:30 |
| dmsimard | I don't think you pin so it should install latestyeah | 13:32 |
| *** sep has joined #openstack-ansible | 13:48 | |
| *** sshnaidm|rover is now known as sshnaidm|afk | 13:54 | |
| mgariepy | anyone here have seen : ERROR! Unexpected templating type error occurred on ({{ (lxc_container_networks_combined.values() | list) | sort(attribute='interface') }}): '<' not supported between instances of 'str' and 'NoneType' ? | 14:05 |
| mgariepy | huh.. silly me. missing interface on the host.. | 14:07 |
| mgariepy | lol. | 14:07 |
| *** spatel has joined #openstack-ansible | 14:19 | |
| kleini | @noonedeadpunk, is that image problem an acute problem? tried to deploy 21.1.0 on Ubuntu 18.04 and it fails downloading a 18.04.3 base image. So it seems to me, that I need to overwrite lxc_hosts_container_image_url | 14:20 |
| noonedeadpunk | kleini: you can set lxc_hosts_container_image_download_legacy: true and this should work out | 14:21 |
| *** cshen has quit IRC | 14:21 | |
| * noonedeadpunk working on proper fix | 14:21 | |
| kleini | Does it help to test https://review.opendev.org/759229 ? | 14:28 |
| noonedeadpunk | well, yes:) | 14:29 |
| kleini | Anything else I can help with as I am already stumbling over this?= | 14:29 |
| noonedeadpunk | I'm currently looking into what is wrong with centos, I think it's the only point for the moment | 14:30 |
| noonedeadpunk | actually I have etherpad here with options how to work this through https://etherpad.opendev.org/p/osa-ubuntu-image-source | 14:30 |
| noonedeadpunk | and if you have some input - feel free to share | 14:30 |
| kleini | I read that etherpad but actually I do not have any good idea, how to improve the situation? Your current solution in the patch looks good to me. I only found a typo, if it is a typo: "cheksum" instead of "checksum" | 14:38 |
| *** LowKey has joined #openstack-ansible | 14:38 | |
| *** d34dh0r53 has quit IRC | 14:40 | |
| noonedeadpunk | ah, good catch | 14:43 |
| noonedeadpunk | well, if current patch is good for you it's also great input | 14:43 |
| *** d34dh0r53 has joined #openstack-ansible | 14:43 | |
| noonedeadpunk | just need more eyes so asked:) | 14:43 |
| *** macz_ has joined #openstack-ansible | 14:45 | |
| *** macz_ has quit IRC | 14:45 | |
| *** macz_ has joined #openstack-ansible | 14:45 | |
| *** sshnaidm|afk is now known as sshnaidm|rover | 14:48 | |
| *** pto has quit IRC | 14:50 | |
| *** pto has joined #openstack-ansible | 14:52 | |
| *** jawad_axd has joined #openstack-ansible | 14:55 | |
| jamesdenton | noonedeadpunk i've not seen linters errors like these before. any suggestions? or just a recheck? https://zuul.opendev.org/t/openstack/build/e397aba2473a4c1392e4b44d72e0b91b | 14:59 |
| *** pto has quit IRC | 14:59 | |
| noonedeadpunk | jamesdenton: ah. it's not happy with how when statement are used ie here https://review.opendev.org/#/c/736336/5/tasks/ironic_conductor_post_install.yml | 15:01 |
| noonedeadpunk | there shouldn't be "{{ }}" in when | 15:01 |
| jamesdenton | gotcha | 15:03 |
| jamesdenton | thank you | 15:03 |
| noonedeadpunk | we've replaced ansible-lint excludes with warnings, to remeber about technical debt :p | 15:04 |
| noonedeadpunk | but the real issue is in this ascii box | 15:05 |
| *** jawad_axd has quit IRC | 15:05 | |
| jamesdenton | No Jinja2 in when | 15:05 |
| openstackgerrit | Dmitriy Rabotyagov (noonedeadpunk) proposed openstack/openstack-ansible-lxc_hosts master: Determine latest base image available https://review.opendev.org/759229 | 15:08 |
| jamesdenton | ok, i didn't know how to decipher the box. makes sense now | 15:10 |
| openstackgerrit | James Denton proposed openstack/openstack-ansible-os_ironic master: Add iPXE support to Ironic Conductor https://review.opendev.org/736336 | 15:24 |
| *** gregwork has joined #openstack-ansible | 15:26 | |
| noonedeadpunk | uh, damn it | 15:26 |
| noonedeadpunk | we're in bade shape with centos mages as well.... | 15:26 |
| noonedeadpunk | they don't ship layers as a standalone image - only docker ones https://cloud.centos.org/centos/8/x86_64/images/ | 15:26 |
| noonedeadpunk | seems I crossed out option 4 too early | 15:28 |
| recyclehero | jrosser: thank u jrosser. it was what I was looking for. | 15:28 |
| jrosser | recyclehero: excellent! | 15:29 |
| openstackgerrit | Dmitriy Rabotyagov (noonedeadpunk) proposed openstack/openstack-ansible-lxc_hosts master: Determine latest base image available https://review.opendev.org/759229 | 15:30 |
| jamesdenton | what is layers? | 15:33 |
| noonedeadpunk | well, it's docker thing. eventually it's how docker packs images | 15:35 |
| noonedeadpunk | and centos was providing only rootfs previously, but modern image has only this http://paste.openstack.org/show/799288/ | 15:35 |
| noonedeadpunk | and layer.tar is rootfs | 15:35 |
| noonedeadpunk | I think https://medium.com/@jessgreb01/digging-into-docker-layers-c22f948ed612 kind of explains | 15:36 |
| noonedeadpunk | https://cameronlonsdale.com/2018/11/26/whats-in-a-docker-image/ is probably better | 15:36 |
| noonedeadpunk | ok. let's close this with pretty backportable https://review.opendev.org/#/c/759229 and will talk about how to resolve this in a better way during PTG | 15:52 |
| openstackgerrit | Dmitriy Rabotyagov (noonedeadpunk) proposed openstack/openstack-ansible-lxc_hosts master: Determine latest base image available https://review.opendev.org/759229 | 15:53 |
| *** rpittau is now known as rpittau|afk | 15:57 | |
| *** yann-kaelig has joined #openstack-ansible | 15:58 | |
| openstackgerrit | amolkahat proposed openstack/openstack-ansible-os_tempest master: Migrate to content provider jobs/templates https://review.opendev.org/759287 | 16:00 |
| *** mgariepy has quit IRC | 16:02 | |
| openstackgerrit | amolkahat proposed openstack/openstack-ansible-os_tempest master: Migrate to content provider jobs/templates https://review.opendev.org/759287 | 16:02 |
| noonedeadpunk | so, should we merge https://review.opendev.org/#/c/759229/4 ? | 16:29 |
| openstackgerrit | Dmitriy Rabotyagov (noonedeadpunk) proposed openstack/openstack-ansible-lxc_hosts stable/ussuri: Determine latest base image available https://review.opendev.org/759298 | 16:47 |
| openstackgerrit | Dmitriy Rabotyagov (noonedeadpunk) proposed openstack/openstack-ansible-lxc_hosts stable/train: Determine latest base image available https://review.opendev.org/759299 | 16:50 |
| recyclehero | do u have plans to update ussuri docs to checkout 21.1.0 instead of 21.0.1 | 16:54 |
| openstackgerrit | Dmitriy Rabotyagov (noonedeadpunk) proposed openstack/openstack-ansible-lxc_hosts stable/stein: Determine latest base image available https://review.opendev.org/759300 | 16:54 |
| noonedeadpunk | recyclehero: docs are designed to be auto updated but it's not working at the moment for $reason | 16:55 |
| *** mgariepy has joined #openstack-ansible | 16:55 | |
| noonedeadpunk | which I can't really recall, but it's something we need to look into | 16:55 |
| * jrosser tries to review 759229 | 16:56 | |
| jrosser | been very very long day :/ | 16:56 |
| noonedeadpunk | to sum up - I think it's just small backportable change we will need to replace pretty soon | 16:56 |
| noonedeadpunk | as I've started to work on option 4 https://etherpad.opendev.org/p/osa-ubuntu-image-source and it's pretty deep dive | 16:56 |
| noonedeadpunk | well not so much, but we have https://opendev.org/openstack/openstack-ansible-lxc_hosts/src/branch/master/defaults/main.yml#L200 which we made as list for $reason | 16:58 |
| noonedeadpunk | and have no idea at the moment what to do with that.... | 16:58 |
| noonedeadpunk | and probably we can discuss more during ptg | 16:58 |
| jrosser | the conditional on here is hurting my head https://review.opendev.org/#/c/759229/4/tasks/lxc_cache_prestage.yml | 16:59 |
| noonedeadpunk | yeah... so we run it only if all required variables are set, and we don't have lxc_hosts_container_image_url | 17:00 |
| jrosser | the first two conditions are somehow combined in a complex way | 17:01 |
| jrosser | becasue of whats in defaults/main.yml | 17:01 |
| jrosser | or an override will change that | 17:01 |
| noonedeadpunk | _lxc_hosts_container_image_url_base is defined only for suse... | 17:02 |
| noonedeadpunk | and regarding lxc_hosts_container_image_url - yes it will be overriden either with user overrides, or with _lxc_hosts_container_image_url for centos | 17:03 |
| jrosser | oh my | 17:03 |
| jrosser | like needs some whole document writing above that task | 17:04 |
| noonedeadpunk | I hope we will replace it tbh.... as we have issue with centos images as well | 17:04 |
| noonedeadpunk | but what I come up with.. breaks my head as well... | 17:05 |
| noonedeadpunk | but I'm ok with documenting it | 17:05 |
| jrosser | theres like option 5 being DIB | 17:06 |
| jrosser | don't know if thats even remotely sane idea | 17:06 |
| noonedeadpunk | not sure if dib can make us tar archive | 17:07 |
| noonedeadpunk | we can bind mount raw or qcow... | 17:07 |
| noonedeadpunk | but that will resolve issue with our custom scripts | 17:08 |
| noonedeadpunk | that we run against retrieved image | 17:08 |
| noonedeadpunk | but it may take the way more time | 17:09 |
| jrosser | yeah, it was something jesse talked about a long time ago | 17:09 |
| noonedeadpunk | lets write it down?:) | 17:09 |
| *** cshen has joined #openstack-ansible | 17:12 | |
| *** jawad_axd has joined #openstack-ansible | 17:17 | |
| *** cshen has quit IRC | 17:19 | |
| spatel | noonedeadpunk: can i deploy victoria using OSA or its still under work? | 17:20 |
| noonedeadpunk | spatel: under work:( | 17:21 |
| noonedeadpunk | we need to merge several things and https://review.opendev.org/#/c/755973/ | 17:21 |
| spatel | next week i am planning to rollout my production cloud. I can go with ussuri and later upgrade to victoria when stable ready available | 17:21 |
| noonedeadpunk | after that I think we will make some milestone release (ie beta) | 17:22 |
| spatel | cool, thank you | 17:22 |
| *** ChiTo has joined #openstack-ansible | 17:24 | |
| ChiTo | Hi OSA team | 17:24 |
| spatel | ChiTo: Hello | 17:24 |
| ChiTo | is there any mechanism to remove a specific set of packages that previously were installed on a bare metal node within the venvs? | 17:24 |
| noonedeadpunk | nope:( | 17:25 |
| ChiTo | got it, manually only right? | 17:25 |
| ChiTo | I mean just removing my venvs related to those projects | 17:25 |
| noonedeadpunk | well, you can run ad-hoc ansible command | 17:25 |
| ChiTo | true, that makes sense too | 17:25 |
| ChiTo | since I am relocating some services I decided to not use someones who were previously insstalled on a different server | 17:26 |
| ChiTo | but I want to purge everythiing related to those services | 17:26 |
| *** andrewbonney has quit IRC | 17:26 | |
| noonedeadpunk | yeah, we don't have anything prepared for that | 17:26 |
| ChiTo | it is fine I think I can just remove them to avoid being registered against my controllers, and previously disable them to avoid struggling with the new ones | 17:27 |
| ChiTo | my main goal is because I am deploying my compute nodes to act as neutron node as well, but I would like to separate the Neutron APIs from the data plane services like metadata, dhcp, l3, l2, agents in general, etc, but I am afraid that is not possible right? | 17:28 |
| ChiTo | because the host group is only referred as a network_hosts from the inventory perspective | 17:28 |
| ChiTo | but not sure if I can break down the neutron-server component which exposes the API and separate the other services to be deployed on my compute nodes | 17:28 |
| openstackgerrit | Jonathan Rosser proposed openstack/openstack-ansible-tests stable/stein: Pin virtualenv<20 for python2 functional tests https://review.opendev.org/759308 | 17:28 |
| ChiTo | I am doing a HA/DVR on computes | 17:28 |
| noonedeadpunk | well by default neutron-api is separated from all other agents | 17:29 |
| noonedeadpunk | *neutron-server | 17:29 |
| ChiTo | finally I was able to deploy the VXLAN properly for tunnels on the computes but now I am dealing with the br-provider/br-vlan for my external network and I would like to make the compute nodes to route their VMs | 17:29 |
| jrosser | gerritbot seems to be gone but this is needed for unbreaking stein https://review.opendev.org/#/c/759308/1 | 17:30 |
| noonedeadpunk | pretty sure it won't pass due to broken ubuntu images | 17:30 |
| ChiTo | noonedeadpunk: Agreed, but how can I group them separately in my openstack_user_config.yaml because I only have my network_hosts which contain the compute nodes so far | 17:30 |
| jrosser | yeah | 17:30 |
| jrosser | noonedeadpunk: imho we should change neutron-api to always be on the infra host | 17:32 |
| ChiTo | jrosser: exactly something like tht | 17:32 |
| jrosser | everyone here who ever makes dedicated network nodes always wants to move them | 17:32 |
| ChiTo | just to separate my API access for my haproxies | 17:32 |
| *** jawad_axd has quit IRC | 17:32 | |
| noonedeadpunk | have nothing against it | 17:33 |
| jrosser | ChiTo: leaving aside all the dvr and stuff it's totally possible to adjust env.d to do this | 17:33 |
| noonedeadpunk | we kind of have neutron_server_container but yeah, it's part of general network_containers | 17:33 |
| jrosser | it's because the agents used to be in containers, so kind of history that is long now removed | 17:34 |
| ChiTo | jrosser: understood, so i have to create my own variables to group something like "neutron-api" on x or y server within my env.d on that specific group? | 17:35 |
| ChiTo | I mean I have to customize my inventory right? | 17:35 |
| noonedeadpunk | I still a bit confused in what we can set in openstack_user_config and what not from env.d | 17:35 |
| jrosser | if CeeMac is around he had a example in a paste of what to do | 17:35 |
| jrosser | i'm not on my work computer to copy this from my deployment right now | 17:36 |
| noonedeadpunk | but I thought that we can set either from container_skel or component_skel ? | 17:36 |
| CeeMac | jrosser: to move neutron server to controller? | 17:37 |
| jrosser | yeah, you shared a recipie a loooong time ago to do that | 17:37 |
| CeeMac | I'm just making dinner but I can dig it out later. | 17:37 |
| ChiTo | I tried something like http://paste.openstack.org/show/799290 but not sure if that makes sense | 17:38 |
| CeeMac | Should probably put it in an ether pad or something | 17:38 |
| CeeMac | ChiTo: will you be around for a while? | 17:38 |
| jrosser | is that how dvr works? | 17:38 |
| jrosser | i thought the centralised L3 agent was still needed for some traffic | 17:39 |
| ChiTo | CeeMac: sure thing!, thx a lot | 17:39 |
| * jrosser never done this though | 17:39 | |
| CeeMac | Pretty sure central l3 agent still sits on network nodes too in DVR, I prepped my environment for it but didn't implement it yet | 17:40 |
| CeeMac | I should probably scroll up and read the conversation history for context :) | 17:41 |
| jrosser | something something snat | 17:41 |
| ChiTo | jrosser, CeeMac: In my case I just set the compute nodes to work as network hosts too, so essentially my l3 agents are working on HA mode per compute | 17:41 |
| ChiTo | jrosser: yep, usually dvr_snat as the mode on the computes | 17:41 |
| ChiTo | and also enable_distributed_routing = True in the openvswitch_agent.ini | 17:42 |
| ChiTo | so essentially "I don't have" neutron nodes separately | 17:43 |
| *** sshnaidm|rover is now known as sshnaidm|off | 17:45 | |
| CeeMac | ChiTo: makes sense. I have dedicated network nodes to limit point of presence on the Internet, but with DVR that becomes moot anyway I guess | 17:52 |
| ChiTo | CeeMac: Agreed, at the end the physical network infrastructure admin is the main responsible to track 100% on what is going on those nodes. Also I find it useful because of the zero-SPOF and eventually better L3 optimization techniques per compute when it comes to larger clouds | 17:54 |
| CeeMac | ChiTo: true, but that guy is me too >_< | 17:55 |
| ChiTo | CeeMac: hehe, agreed, when we have full control I rather suggest a separated architecture to control the neutron hosts but the cloud I am working on has a a lot of siloed-approaches when it comes to operations. | 17:57 |
| openstackgerrit | Dmitriy Rabotyagov (noonedeadpunk) proposed openstack/openstack-ansible-lxc_hosts master: Determine latest base image available v4 https://review.opendev.org/759310 | 18:02 |
| noonedeadpunk | jrosser: well, it's option from variant 4 which gets from images.linuxcontainers.org | 18:02 |
| *** alvinstarr has quit IRC | 18:04 | |
| *** MickyMan77 has joined #openstack-ansible | 18:05 | |
| openstackgerrit | Dmitriy Rabotyagov (noonedeadpunk) proposed openstack/openstack-ansible-lxc_hosts master: Determine latest base image available v4 https://review.opendev.org/759310 | 18:13 |
| *** MickyMan77 has quit IRC | 18:13 | |
| recyclehero | guys, whats unbound? | 18:32 |
| noonedeadpunk | dns | 18:32 |
| recyclehero | like bind but unbound? | 18:33 |
| *** KurtB has quit IRC | 18:35 | |
| openstackgerrit | Dmitriy Rabotyagov (noonedeadpunk) proposed openstack/openstack-ansible-lxc_hosts master: Determine latest base image available v4 https://review.opendev.org/759310 | 18:35 |
| noonedeadpunk | well yes | 18:35 |
| noonedeadpunk | just lightweight and simple | 18:35 |
| noonedeadpunk | https://github.com/NLnetLabs/unbound | 18:36 |
| recyclehero | thanks | 18:36 |
| CeeMac | ChiTo: http://paste.openstack.org/show/799291/ this should do the trick | 18:40 |
| noonedeadpunk | I guess valid env.d should be like this http://paste.openstack.org/show/799292/ | 18:45 |
| noonedeadpunk | except c omponent_skel not omponent_skel | 18:45 |
| noonedeadpunk | * component_skel | 18:46 |
| *** MickyMan77 has joined #openstack-ansible | 18:48 | |
| CeeMac | noonedeadpunk: I used the one in the paste and it worked? *shrug* | 18:50 |
| CeeMac | Not sure if it would merge in bits from the default neutron-server? | 18:51 |
| noonedeadpunk | well, if you add it to already generated inventory, it probably should work... | 18:52 |
| noonedeadpunk | as you have agents already in place... | 18:52 |
| CeeMac | I'm happy to bow to your experience though if the override should be complete as you put it. | 18:53 |
| CeeMac | Noonedeadpu | 18:53 |
| CeeMac | Bah | 18:53 |
| CeeMac | noonedeadpunk: I see, hadn't thought of that | 18:53 |
| noonedeadpunk | anyway better safe then sorry) | 18:54 |
| CeeMac | I'll update my records accordingly :) | 18:54 |
| ChiTo | noonedeadpunk, CeeMac: Thanks guys, I think I have to include the containers_skel independently if I use metal only | 18:54 |
| CeeMac | True fact | 18:54 |
| noonedeadpunk | I think we should make this change upstream | 18:54 |
| ChiTo | I am going to try them during my day | 18:54 |
| noonedeadpunk | but we will need to update group names as well... | 18:54 |
| CeeMac | I think it's more intuitive to keep the neutron api with the other api containers | 18:55 |
| *** MickyMan77 has quit IRC | 18:55 | |
| noonedeadpunk | it totally is | 18:55 |
| noonedeadpunk | just agents were in containers one day, so it's kind of historical thing | 18:55 |
| CeeMac | Not sure what convention you would use for group names | 18:55 |
| CeeMac | Yeah, evolution etc | 18:56 |
| noonedeadpunk | who knows... probably just replace `-` with `_` | 18:56 |
| noonedeadpunk | as having dash in group names is obsolete in terms of ansible for a qwhile | 18:57 |
| CeeMac | I have a feeling I copied the formating from compute-infra_containers, but I could be wrong | 18:57 |
| noonedeadpunk | yeah, we have that type across all services I think | 18:57 |
| noonedeadpunk | including repo, cinder, placement, etc | 18:58 |
| CeeMac | Oh, so it's an upstream ansible convention? | 18:58 |
| noonedeadpunk | osa upstream or ansible upstream?:) | 18:58 |
| CeeMac | Hah I mean the reasoning to remove the - | 18:59 |
| noonedeadpunk | ansible is about to prohivit usage of `-` in group names, so we're balancing here | 18:59 |
| noonedeadpunk | yeah, it is | 18:59 |
| CeeMac | Right, makes sense | 18:59 |
| *** yann-kaelig has quit IRC | 18:59 | |
| CeeMac | I have so much to learn still haha | 18:59 |
| noonedeadpunk | they have even this https://docs.ansible.com/ansible/latest/reference_appendices/config.html#transform-invalid-group-chars | 19:00 |
| noonedeadpunk | well, here's the list http://paste.openstack.org/show/799293/ :p | 19:05 |
| CeeMac | Hah, just a couple to clear up! | 19:14 |
| *** cshen has joined #openstack-ansible | 19:15 | |
| openstackgerrit | Dmitriy Rabotyagov (noonedeadpunk) proposed openstack/openstack-ansible-lxc_hosts master: Determine latest base image available v4 https://review.opendev.org/759310 | 19:17 |
| *** cshen has quit IRC | 19:19 | |
| *** gyee has joined #openstack-ansible | 19:26 | |
| openstackgerrit | Dmitriy Rabotyagov (noonedeadpunk) proposed openstack/openstack-ansible-lxc_hosts master: Use lxc_image_cache_server_mirrors as image source https://review.opendev.org/759310 | 19:40 |
| *** wpp has quit IRC | 19:42 | |
| *** Muran has quit IRC | 19:47 | |
| *** Muran has joined #openstack-ansible | 19:48 | |
| *** Muran has quit IRC | 19:52 | |
| openstackgerrit | Dmitriy Rabotyagov (noonedeadpunk) proposed openstack/openstack-ansible-lxc_hosts master: Use lxc_image_cache_server_mirrors as image source https://review.opendev.org/759310 | 20:05 |
| *** Muran has joined #openstack-ansible | 20:17 | |
| *** Muran has quit IRC | 20:22 | |
| recyclehero | where unbound host are defined, | 20:22 |
| recyclehero | doh, accidental enter! | 20:23 |
| *** yolanda has quit IRC | 20:24 | |
| *** yolanda has joined #openstack-ansible | 20:24 | |
| *** nurdie has quit IRC | 20:26 | |
| *** Muran has joined #openstack-ansible | 20:29 | |
| *** Muran has quit IRC | 20:34 | |
| recyclehero | CeeMac: I was following what your paste for learning purposes. --limit network-infa_all: does the all have special syntax meaning as I dont see it. and is it necessary to use limit with openstack-ansbile os-neutron-install.ynl ? | 20:42 |
| *** mgariepy has quit IRC | 20:48 | |
| *** mgariepy has joined #openstack-ansible | 21:02 | |
| *** rfolco has quit IRC | 21:03 | |
| *** nurdie has joined #openstack-ansible | 21:13 | |
| *** nurdie has quit IRC | 21:14 | |
| *** cshen has joined #openstack-ansible | 21:15 | |
| CeeMac | recyclehero: that's one of the dynamic inventory groups | 21:16 |
| CeeMac | You dont have to use limit, but it's useful to restrict the playbook to specific hosts or containers | 21:17 |
| *** cshen has quit IRC | 21:20 | |
| CeeMac | recyclehero: you can view the dynamic inventory to see the various groups etc and then the inventory groups can be used to limit the plays. https://docs.openstack.org/openstack-ansible/queens/reference/inventory/manage-inventory.html | 21:21 |
| CeeMac | Or https://docs.openstack.org/openstack-ansible/latest/reference/inventory/manage-inventory.html for the latest version even | 21:21 |
| *** sum12 has joined #openstack-ansible | 21:35 | |
| *** spatel has quit IRC | 21:50 | |
| *** odyssey4me has quit IRC | 21:53 | |
| *** nurdie has joined #openstack-ansible | 21:54 | |
| *** nurdie has quit IRC | 21:58 | |
| *** tosky has quit IRC | 22:54 | |
| *** rh-jelabarre has quit IRC | 23:15 | |
| *** cshen has joined #openstack-ansible | 23:16 | |
| *** cshen has quit IRC | 23:20 | |
| *** gshippey has quit IRC | 23:21 | |
| recyclehero | CeeMac: thanks | 23:22 |
| recyclehero | I am having trouble with this allocation_ratio | 23:22 |
| recyclehero | I cant set it to anything and it isnt as noonedeadpunk said that I could lunch no matter what shows in horzion hypervisor section | 23:22 |
| recyclehero | I deployed clean, applyed jrosser patch. set cpu_allocation_ratio: '' | 23:23 |
| recyclehero | set initial_cpu_allocation_ratio: 16 with via overriders in user_variable.yml | 23:23 |
| recyclehero | but no luck | 23:24 |
| *** macz_ has quit IRC | 23:24 | |
| recyclehero | when I use the openstack cli openstack resource provider inventory show vcpu it shows allocation_ratio as 16.0 but there is problem with total | 23:25 |
| *** macz_ has joined #openstack-ansible | 23:25 | |
| recyclehero | total is still 8 | 23:25 |
| recyclehero | I am doing api calls which have resource_genration which is some int that resource tracker keeps I guess for knowing where is it | 23:25 |
| recyclehero | after a successful api call I can check with cli that total is set to 128 and generation++ | 23:26 |
| recyclehero | but then it override that and I saw another generation++ and total back to 8 | 23:26 |
| recyclehero | I have read every thing I could get hold of | 23:27 |
| recyclehero | nova/resource_tracker.py and OSA are my suspects | 23:28 |
| recyclehero | I need to lunch with allocation ratio more than 1 | 23:28 |
| * recyclehero goes to sleep | 23:28 | |
| *** macz_ has quit IRC | 23:35 | |
Generated by irclog2html.py 2.17.2 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!