| NeilHanlon | jrosser: my crystal ball is telling me python3.6 will be available for the lifetime of 8.x; there are updates for 3.8 and 3.9 coming in 8.6, and 9.x will have python 3.9 as the default. I suspect 3.10 will arrive _eventually_ | 05:30 |
|---|---|---|
| noonedeadpunk | jrosser: I'd say depends on permissions:) as basically nobody should stop from using "root" mount as remote fs I think | 06:51 |
| jrosser | can the two mounts overlap? | 06:52 |
| noonedeadpunk | oh, you mean that... | 06:52 |
| noonedeadpunk | yeah, indeed, you're right | 06:52 |
| jrosser | I don’t remember why we mount /var/www like that | 06:53 |
| noonedeadpunk | NeilHanlon: well, I think main question was that when 3.6 is not supported by openstack, there's no option right now to have python-libvirt libraries for 3.8 | 06:53 |
| noonedeadpunk | so basically you can't use 3.8 in real world | 06:53 |
| noonedeadpunk | jrosser: so content remain if container rebuilts | 06:53 |
| jrosser | ok - so with a shared file system that’s no longer necessary | 06:54 |
| jrosser | btw I think the biggest difficulty with glusterfs (or anything like that) is deleting/recreating a container | 06:55 |
| jrosser | there’s quite a lot of $stuff to do to replace a previous node | 06:55 |
| noonedeadpunk | and then also we kind of need the way of adding new controller... | 06:58 |
| noonedeadpunk | which kind of leads me to idea if my idea about setting gluster inside containers was stupid one | 07:02 |
| noonedeadpunk | and indeed it should jsut run on controllers and we jsut bind mount stuff inside repo containers... | 07:03 |
| jrosser | i think adding a controller might be OK | 07:31 |
| jrosser | as thats the same playbook as deploying first time, just with an extra host in the group | 07:31 |
| jrosser | oh and just for fun have you seen this https://review.opendev.org/c/openstack/devstack/+/837659 | 07:33 |
| opendevreview | Jonathan Rosser proposed openstack/openstack-ansible-plugins master: Add role for creating simple glusterfs clients/servers https://review.opendev.org/c/openstack/openstack-ansible-plugins/+/837582 | 08:07 |
| noonedeadpunk | I've read ML regarding that. I kind of wonder if release tool is broken as it relied on same behaviour | 08:08 |
| jrosser | i wonder if we do similar when using an AIO for development | 08:10 |
| opendevreview | Dmitriy Rabotyagov proposed openstack/openstack-ansible-openstack_hosts master: Add nova dependency repo for distro install https://review.opendev.org/c/openstack/openstack-ansible-openstack_hosts/+/803475 | 08:10 |
| jrosser | i connect vscode by ssh as ubuntu | 08:10 |
| jrosser | but the deploy / boostrap-ansible is done as root | 08:10 |
| * noonedeadpunk still can't find time to figure out remote debug stuff in vscode | 08:19 | |
| opendevreview | Dmitriy Rabotyagov proposed openstack/openstack-ansible master: Switch services to Yoga https://review.opendev.org/c/openstack/openstack-ansible/+/837701 | 08:30 |
| opendevreview | Dmitriy Rabotyagov proposed openstack/openstack-ansible-openstack_hosts master: Switch codename to Yoga https://review.opendev.org/c/openstack/openstack-ansible-openstack_hosts/+/837702 | 08:40 |
| jrosser | oh - do we already have errors due to new git? https://zuul.opendev.org/t/openstack/build/309ad42f7ec748d3bb4ea0b5d8de96a9/log/job-output.txt#2467-2478 | 08:41 |
| noonedeadpunk | actually I was thinking about that | 08:42 |
| noonedeadpunk | hm, I can't reproduce issue locally with newest git on focal | 08:57 |
| jrosser | i wonder if it is becasue in CI we "do the right thing" and directly try to use the repos on the CI node disk | 08:57 |
| jrosser | yeah 2022-04-13 08:37:11.971533 | ubuntu-focal | Obtaining file:///home/zuul/src/opendev.org/openstack/openstack-ansible | 08:59 |
| jrosser | perhaps the change in devstack to make the repos safe also needs to be applied to the zuul on-disk repos | 09:01 |
| noonedeadpunk | sounds like that indeed | 09:02 |
| noonedeadpunk | jrosser: am I missing smth about how to reproduce? https://paste.openstack.org/show/brw79zjj9l9iXUK4w3zx/ | 09:08 |
| jrosser | do you need to chown the original repo to some not-root user? | 09:09 |
| jrosser | oh you did | 09:10 |
| noonedeadpunk | I did chown -R | 09:11 |
| noonedeadpunk | maybe my issue that it's not bare repo.... | 09:11 |
| jrosser | hold on | 09:11 |
| jrosser | you have Processing /root/openstack-ansible..... | 09:11 |
| jrosser | and stat /home/ubuntu/openstack-ansible/ | 09:12 |
| noonedeadpunk | mmm.... | 09:12 |
| noonedeadpunk | yeah as I cloned from /home/ubuntu/openstack-ansible/ to /root/openstack-ansible | 09:12 |
| noonedeadpunk | but indeed - we not clone - we symlink I guess | 09:12 |
| jrosser | perhaps - i was just looking at the zuul log for where actually the working dir is | 09:13 |
| noonedeadpunk | ok, yes, I catched that :) | 09:13 |
| jrosser | 2022-04-13 08:37:13.530478 | ubuntu-focal | cwd: /home/zuul/src/opendev.org/openstack/openstack-ansible/ | 09:13 |
| jrosser | if you can reproduce it i was guessing at something like this https://paste.opendev.org/show/bbvALvPnjYbkcv06ZT5W/ | 09:15 |
| jrosser | likley using '.' there is wrong | 09:15 |
| noonedeadpunk | let me check options | 09:18 |
| noonedeadpunk | I think we can do a bit different hook.... | 09:20 |
| noonedeadpunk | not sure how much better it is though | 09:20 |
| opendevreview | Dmitriy Rabotyagov proposed openstack/openstack-ansible master: Mark OSA repository as safe in git.config in CI https://review.opendev.org/c/openstack/openstack-ansible/+/837706 | 09:29 |
| noonedeadpunk | I wonder if that would help ^ | 09:31 |
| noonedeadpunk | no idea what collections are installed in built-in ansible for zuul... | 09:38 |
| noonedeadpunk | ok, there's no such collection :( | 09:45 |
| opendevreview | Dmitriy Rabotyagov proposed openstack/openstack-ansible master: Mark OSA repository as safe in git.config in CI https://review.opendev.org/c/openstack/openstack-ansible/+/837706 | 09:47 |
| opendevreview | Dmitriy Rabotyagov proposed openstack/openstack-ansible master: Switch services to Yoga https://review.opendev.org/c/openstack/openstack-ansible/+/837701 | 10:09 |
| noonedeadpunk | yep, it works now, just ansible 2.9 is used for zuul | 10:18 |
| opendevreview | Dmitriy Rabotyagov proposed openstack/openstack-ansible-plugins master: Pass ssh_keypairs role in check mode https://review.opendev.org/c/openstack/openstack-ansible-plugins/+/837713 | 11:00 |
| opendevreview | Dmitriy Rabotyagov proposed openstack/openstack-ansible-plugins master: Pass ssh_keypairs role in check mode https://review.opendev.org/c/openstack/openstack-ansible-plugins/+/837713 | 11:04 |
| *** dviroel|out is now known as dviroel | 11:11 | |
| opendevreview | Dmitriy Rabotyagov proposed openstack/ansible-role-systemd_service master: Pass systemd_service role in check mode https://review.opendev.org/c/openstack/ansible-role-systemd_service/+/837716 | 11:12 |
| opendevreview | Dmitriy Rabotyagov proposed openstack/openstack-ansible-rabbitmq_server master: Pass rabbitmq role in check mode https://review.opendev.org/c/openstack/openstack-ansible-rabbitmq_server/+/837718 | 11:28 |
| opendevreview | Siavash Sardari proposed openstack/openstack-ansible-os_cinder master: Add the ability to disable send_actions option in cinder-volume-usage-audit service. https://review.opendev.org/c/openstack/openstack-ansible-os_cinder/+/837570 | 11:36 |
| opendevreview | Dmitriy Rabotyagov proposed openstack/openstack-ansible-rabbitmq_server master: Pass rabbitmq role in check mode https://review.opendev.org/c/openstack/openstack-ansible-rabbitmq_server/+/837718 | 11:37 |
| opendevreview | Dmitriy Rabotyagov proposed openstack/openstack-ansible-rabbitmq_server master: Replace shell with simple command for version verification https://review.opendev.org/c/openstack/openstack-ansible-rabbitmq_server/+/837719 | 11:42 |
| opendevreview | Dmitriy Rabotyagov proposed openstack/openstack-ansible-rabbitmq_server master: Replace shell with simple command for version verification https://review.opendev.org/c/openstack/openstack-ansible-rabbitmq_server/+/837719 | 11:43 |
| opendevreview | amolkahat proposed openstack/openstack-ansible-os_tempest master: Remove tripleo-ci-centos-7-standalone-os-tempest-train https://review.opendev.org/c/openstack/openstack-ansible-os_tempest/+/831373 | 11:47 |
| opendevreview | Dmitriy Rabotyagov proposed openstack/ansible-role-pki master: Allow to run PKI role in check mode https://review.opendev.org/c/openstack/ansible-role-pki/+/837722 | 11:59 |
| opendevreview | Dmitriy Rabotyagov proposed openstack/openstack-ansible master: Allow to run utility playbook in check mode https://review.opendev.org/c/openstack/openstack-ansible/+/837724 | 12:26 |
| NeilHanlon | noonedeadpunk: i had started experimenting with 3.8 a little bit on Rocky. I'll probably start a review soon with what i'd been looking at. should be easy enough to install the py3.8 module stream and use that instead | 12:30 |
| opendevreview | Dmitriy Rabotyagov proposed openstack/openstack-ansible master: Allow to run utility playbook in check mode https://review.opendev.org/c/openstack/openstack-ansible/+/837724 | 13:14 |
| opendevreview | Jonathan Rosser proposed openstack/openstack-ansible-repo_server master: Add facility to store repo contents on a remote mount https://review.opendev.org/c/openstack/openstack-ansible-repo_server/+/837587 | 14:07 |
| SiavashSardari | hey guys, I have a question regarding using OVS for lxc bridges in controllers. before I begin, I need to say, I tested using ovs for br-mgmt, br-storage, etc and it worked, everything is up and running. | 14:32 |
| SiavashSardari | we had a major issue that after restarts containers didn't start | 14:33 |
| SiavashSardari | finally we found out the issue was in veth-cleanup | 14:34 |
| SiavashSardari | I changed the lxc-container-create role and made it work with ovs and I was thinking about contributing upstream | 14:35 |
| jrosser | does that script need adjusting to be ovs-aware somehow? | 14:35 |
| SiavashSardari | but I can not get my head around how it is working in the first place. | 14:36 |
| jrosser | :) | 14:36 |
| SiavashSardari | yeah actually I checked if it is ovs and delete the port on the ovs bridge | 14:37 |
| SiavashSardari | how this (https://opendev.org/openstack/openstack-ansible-lxc_container_create/src/branch/master/files/lxc-veth-wiring.sh#L55) can add port to an ovs switch?? | 14:38 |
| jamesdenton | it wouldn't AFAIK | 14:39 |
| SiavashSardari | I know, but try it. create br-mgmt bridge with ovs, and everything gets created and works like a charm | 14:40 |
| jrosser | brctl sets the bridge config in the kernel? | 14:41 |
| jrosser | which sounds like it's a set abstracted from the actual implementation of the bridge | 14:42 |
| * jrosser guesses | 14:42 | |
| jamesdenton | SiavashSardari do you have some output to validate that? | 14:42 |
| jamesdenton | if i try in the CLI i get an 'operation not supported' error | 14:43 |
| jamesdenton | that's creating OVS bridge with ovs-vsctl add-br and putting an interface in it with brctl addif | 14:43 |
| SiavashSardari | I know I tried and get the same error too. let me share a little of my config and deployment with you | 14:44 |
| opendevreview | Jonathan Rosser proposed openstack/openstack-ansible-plugins master: Add role for creating simple glusterfs clients/servers https://review.opendev.org/c/openstack/openstack-ansible-plugins/+/837582 | 14:45 |
| opendevreview | Jonathan Rosser proposed openstack/openstack-ansible master: Use glusterfs to synchronise repo server contents https://review.opendev.org/c/openstack/openstack-ansible/+/837589 | 14:46 |
| SiavashSardari | this is my openstack_user_config | 14:46 |
| SiavashSardari | https://paste.opendev.org/show/bYSwrrPSXOkur9hc4sKu/ | 14:46 |
| jamesdenton | k | 14:47 |
| SiavashSardari | this is on one of the controllers | 14:49 |
| SiavashSardari | https://paste.opendev.org/show/bHAM6j5QgUP3kMFBQ8H6/ | 14:49 |
| jamesdenton | you mentioned making changes to support OVS w/ LXC? | 14:49 |
| SiavashSardari | it was deployed without my changes | 14:50 |
| jamesdenton | did you create those OVS bridges manually? the br-int, br-lbaas, etc? | 14:51 |
| SiavashSardari | my change just deletes the ovs port in veth-cleanup scripts | 14:51 |
| SiavashSardari | yeah I have a script for that with usual ovs-vsctl commands | 14:52 |
| jamesdenton | gotcha | 14:52 |
| SiavashSardari | just in case, let me upload my patch too | 14:54 |
| jamesdenton | please | 14:55 |
| noonedeadpunk | Just in case https://review.opendev.org/c/openstack/openstack-ansible/+/837706 blocking gates:) | 14:56 |
| jamesdenton | you need a review on that? | 14:59 |
| noonedeadpunk | Yes, please :) | 15:03 |
| jamesdenton | done :D | 15:03 |
| opendevreview | Siavash Sardari proposed openstack/openstack-ansible-lxc_container_create master: Add ovs support for lxc containers. The main problem this commit is resolving is that with openvswitch, the ports were not get cleaned up, and after restart (hard restart or even with soft restart lxc.service gets time-out) containers couldn't start properly, due to existing port on the bridges. https://review.opendev.org/c/openstack/openstack-ans | 15:07 |
| admin1 | SiavashSardari, native ovs support for controllers will be welcome change :) | 15:08 |
| admin1 | right now, mgmt and storage is lb and vlan and vxlan are ovs | 15:08 |
| SiavashSardari | in my case all are ovs | 15:09 |
| SiavashSardari | well except lxcbr | 15:09 |
| SiavashSardari | actually when I started, I didn't know about lxcbr :D | 15:10 |
| admin1 | SiavashSardari, when you put vxlan range like range: "1:16777215" .. it adds all of those ranges in a table and its not indexed | 15:12 |
| opendevreview | Siavash Sardari proposed openstack/openstack-ansible-lxc_container_create master: Add ovs support for lxc containers. https://review.opendev.org/c/openstack/openstack-ansible-lxc_container_create/+/837734 | 15:13 |
| admin1 | i learnt it the hard way .. when neutron was taking ages to load | 15:13 |
| admin1 | so these days i use 11111,99999 as range .. | 15:13 |
| admin1 | i even complained to the neutron guys and i think opened a bug report as well | 15:13 |
| SiavashSardari | I feel you. it took me a long time to find network_segment_range plugin after upgrading to V. | 15:15 |
| SiavashSardari | I believe there was a patch in neutron for the performance issue, I've forgot to check if it got merged or not. | 15:16 |
| jamesdenton | you can always adjust the range later if you want to be a little more conservative at the getgo | 15:17 |
| SiavashSardari | in my case, after upgrade to V, network create used to take about 70 secs, after using network_segment_range plugin, it got to ~25 secs. then I set the range to 167772 and the responce time got reasonable | 15:22 |
| SiavashSardari | about the lxc ovs support patch, it works fine, tested and everything in stage and prod env. I would appreciate any comment on the patch. | 15:25 |
| opendevreview | Jonathan Rosser proposed openstack/openstack-ansible master: Use glusterfs to synchronise repo server contents https://review.opendev.org/c/openstack/openstack-ansible/+/837589 | 15:26 |
| *** dviroel is now known as dviroel|lunch | 15:28 | |
| SiavashSardari | and also PLEASE someone explain how the *** ovs bridges are and was working with lxc_container_create role? i'm literally pulling my hair on this | 15:28 |
| opendevreview | Jonathan Rosser proposed openstack/openstack-ansible-repo_server master: Remove all code for lsync, rsync and ssh https://review.opendev.org/c/openstack/openstack-ansible-repo_server/+/837588 | 15:28 |
| jrosser | SiavashSardari: maybe i would try strace the brctl command and see what it actually does? | 15:29 |
| jamesdenton | SiavashSardari i am trying to replicate in an AIO | 15:31 |
| SiavashSardari | Thank you both, if you need any kinda help or info about my setup, hit me up please. | 15:32 |
| jrosser | SiavashSardari: which OS is this? | 15:34 |
| SiavashSardari | ubuntu 20.04 | 15:36 |
| opendevreview | Dmitriy Rabotyagov proposed openstack/openstack-ansible master: [doc] Add doc how to pass SR-IOV inside containers https://review.opendev.org/c/openstack/openstack-ansible/+/833775 | 15:38 |
| jamesdenton | neat | 15:39 |
| opendevreview | Dmitriy Rabotyagov proposed openstack/openstack-ansible-os_glance master: Allow os_glance to run in check mode https://review.opendev.org/c/openstack/openstack-ansible-os_glance/+/837739 | 15:41 |
| noonedeadpunk | I kind of wonder if we can leverage bridge type somewhere in provider_networks? | 15:42 |
| noonedeadpunk | why we can't jsut add bridge_type there as keyword? | 15:43 |
| noonedeadpunk | gimme several mins | 15:48 |
| jamesdenton | i can confirm that if br-mgmt (and the others) exist as ovs bridges, setup-hosts.yml will shove the veths in there: https://paste.opendev.org/show/b6zYGuSU9lvv9wDU3PTp/ | 15:50 |
| SiavashSardari | I like your Idea. I've taken another approach and let the role finds out about the bridge type by itself. | 15:50 |
| spatel | jrosser are you trying to use gluster to replace lsyncd? that would be good. hope we can use same module to create glance storage when ceph not available :)_ | 15:55 |
| spatel | SiavashSardari i would like to have LXB on controller and computes on OVS (that is easy and simple deployment). | 15:57 |
| SiavashSardari | that was my thinking too, tried the idea back in 2019, it worked but last year we faced the restart issue. while we are on this matter, I user ovs as firewall driver too | 16:01 |
| spatel | faced restart issue? | 16:01 |
| SiavashSardari | but there is a down side about that I think ~2-3k security group rules will cause ovs to crash | 16:02 |
| spatel | That is different story :) I heard there are some issue related connection tracking also. | 16:03 |
| noonedeadpunk | oh, wait, there's already such thing I bleieve | 16:03 |
| SiavashSardari | actually stopping containers with ovs ports won't delete ports and when you want to start the containers it fails with some error like 'port already exists.' | 16:05 |
| SiavashSardari | @spatel I said the firewall driver thing about the idea of replacing lb with ovs in all the deployment. :) | 16:07 |
| jrosser | spatel: i am making patches for two things for the repo server containers - first so they can take some shared storage mount from anywhere (NFS / cephfs / whatever), and second so that there is a default deployment of glusterfs for the repo server when the deployer doesnt choose anything else | 16:08 |
| opendevreview | Dmitriy Rabotyagov proposed openstack/openstack-ansible master: Add ability to define bridge type for containers https://review.opendev.org/c/openstack/openstack-ansible/+/837742 | 16:08 |
| noonedeadpunk | SiavashSardari: won't that make more sense? ^ | 16:08 |
| spatel | jrosser I like that, i would also like if there is a option to move your glance folder over glusterfs. | 16:09 |
| noonedeadpunk | SiavashSardari: as alternative to https://review.opendev.org/c/openstack/openstack-ansible-lxc_container_create/+/837734/2/tasks/lxc_container_ovs_overrides.yml ? | 16:10 |
| jrosser | spatel: right - though that involves a deployment-wide glusterfs which might be a different thing than something specific to the repo containers | 16:10 |
| opendevreview | Dmitriy Rabotyagov proposed openstack/openstack-ansible master: Add ability to define bridge type for containers https://review.opendev.org/c/openstack/openstack-ansible/+/837742 | 16:11 |
| SiavashSardari | @noonedeadpunk TBH I've never thought about that :D | 16:11 |
| opendevreview | Dmitriy Rabotyagov proposed openstack/openstack-ansible master: Add ability to define bridge type for containers https://review.opendev.org/c/openstack/openstack-ansible/+/837742 | 16:11 |
| spatel | currently i have deployed gluster by hand to move my /var/lib/glance over shared filesystem. (in future i would like to use your module to do that job also :) ) | 16:12 |
| jrosser | spatel: did you see this which makes the glance mount support more generic? https://review.opendev.org/c/openstack/openstack-ansible-os_glance/+/837550/1/defaults/main.yml | 16:12 |
| jrosser | though we do have a challenge with where to put the code which ensures that the fuse config is correct and the glusterfs-client package is installed - not really too obvious where that would go for more generic support of glusterfs | 16:13 |
| noonedeadpunk | SiavashSardari: then you will have `bridge_type` as part of `container_networks` dict | 16:13 |
| noonedeadpunk | should be super easy to work with that | 16:13 |
| noonedeadpunk | and likely more logic could be built for better managing ovs for lxc containers | 16:14 |
| noonedeadpunk | but just as an idea how things could be done :) | 16:15 |
| spatel | jrosser assuming we are planning to install gluster on host and bind mount inside repo container right? | 16:16 |
| jrosser | currently my patches run the gluster server inside the repo container and make it totally self contained | 16:16 |
| spatel | hmm | 16:16 |
| jrosser | it is a simple / trivial case and like i say there is a new hook for external storage | 16:16 |
| jrosser | if you have already gluster somwhere else then this would be what you want https://review.opendev.org/c/openstack/openstack-ansible-repo_server/+/837587 | 16:17 |
| SiavashSardari | @noonedeadpunk I like the idea, I was trying to not get too apart from the upstream. but here is the upstream and I'm ok with what works best. | 16:17 |
| jrosser | and do not have OSA deploy any shared storage itself | 16:17 |
| spatel | got it.. | 16:18 |
| jrosser | noonedeadpunk: do you think i should add gluster client support to systemd_mount role a bit like we did for s3fs? | 16:19 |
| jrosser | otherwise for spatel use case with glance we don't have anything to install the client and the mount wont work | 16:20 |
| SiavashSardari | @noonedeadpunk also I'm a bit worried if the final way of handling this issue, breaks our deployments. so kindly, let me keep my job :)) | 16:20 |
| spatel | jrosser why don't we make it default deployment of repo with glusterfs. (its very easy and simple, i don't think anyone trying to mess with repo anymore to mount on external storage) | 16:20 |
| jrosser | spatel: thats how it will work if we merge my patches, with the option to use something else if necessary | 16:21 |
| jrosser | but it won't help you with glance | 16:21 |
| spatel | that is ok.. i will deal with glance my way :) | 16:21 |
| spatel | I thought we are installing gluster on host so thought i can take advantage of binding glance volume | 16:22 |
| noonedeadpunk | jrosser: yeah, I'd say that systemd_mount should be kind of self-contained so make sense to add clients setup there | 16:26 |
| jrosser | thats nice becasue it means the client/server division in my glusterfs playbook goes away | 16:27 |
| jrosser | and it's onlt concerned with deploying the server | 16:27 |
| noonedeadpunk | SiavashSardari: sorry didn't really understand what breaking could be there? | 16:28 |
| jamesdenton | noonedeadpunk LXC appears to have had some logic to determine bridge type (lxc/ovs) for some time: https://discuss.linuxcontainers.org/t/lxc-2-1-has-been-released/487 | 16:30 |
| jamesdenton | which explains why the lxc plays got things plugged in | 16:30 |
| opendevreview | Jonathan Rosser proposed openstack/openstack-ansible-plugins master: Add role for creating simple glusterfs clients/servers https://review.opendev.org/c/openstack/openstack-ansible-plugins/+/837582 | 16:31 |
| noonedeadpunk | yeah, it's surprising how it works out of the box :) | 16:33 |
| SiavashSardari | @noonedeadpunk the way we used ovs for bridges should not work based on the scripts in lxc_container_create role and was not supported by OSA, therefore adding the bridge_type variable and taking advantage of that is not bound to backward compatibility | 16:33 |
| SiavashSardari | all I'm saying is I would be happy to be involved in this change as much as I could possibly can | 16:34 |
| *** dviroel|lunch is now known as dviroel | 16:35 | |
| noonedeadpunk | But bridge_type is optional. | 16:35 |
| noonedeadpunk | It's not smth that must be specified for all networks | 16:35 |
| noonedeadpunk | which means it's quite backwards compatibe? | 16:35 |
| noonedeadpunk | probably I just misunderstood your concern :) | 16:37 |
| SiavashSardari | yeah your right, maybe I cannot properly communicate in English. I will comment on your patch to get updated about this change. also I think, this would fall under a new feature section and it's all new to me. I'd appreciate it if you point me to the right direction of where to start to work on this change | 16:43 |
| SiavashSardari | @jamesdenton Thanks, finally some explanation. they did a really great job on this one and how it works right out of the box. | 16:45 |
| opendevreview | Jonathan Rosser proposed openstack/ansible-role-systemd_mount master: Add support for glusterfs mounts https://review.opendev.org/c/openstack/ansible-role-systemd_mount/+/837750 | 16:47 |
| opendevreview | Jonathan Rosser proposed openstack/openstack-ansible-plugins master: Add role for creating simple glusterfs clients/servers https://review.opendev.org/c/openstack/openstack-ansible-plugins/+/837582 | 16:49 |
| opendevreview | Jonathan Rosser proposed openstack/openstack-ansible master: Use glusterfs to synchronise repo server contents https://review.opendev.org/c/openstack/openstack-ansible/+/837589 | 16:49 |
| jamesdenton | if the bridges are configured by the operator as part of the pre-deployment process, and there's a way to determine bridge type with 'ip link show' (openvswitch vs bridge), then i don't see why the playbooks can't determine that automatically | 16:50 |
| jrosser | jamesdenton: is there already something in the ansible facts - there is ansible_<bridge>.features dict | 16:56 |
| jamesdenton | eh, sorta | 17:06 |
| jamesdenton | i did not see anything network related in ansible_facts. However, if you run the ansible.builtin.setup module, lxc bridges show up as type: bridge while ovs bridges show up as type: ether | 17:06 |
| jamesdenton | but in another environment i do have network info, and nothing pertaining to type is in the features dict. the type key i described is there, though | 17:11 |
| SiavashSardari | i didn't find anything to determine the bridge_type properly, all i found was 'ip link show dev {{ item.1.bridge }} type openvswitch | grep {{ item.1.bridge }}' command | 17:14 |
| opendevreview | Jonathan Rosser proposed openstack/openstack-ansible-repo_server master: Add facility to store repo contents on a remote mount https://review.opendev.org/c/openstack/openstack-ansible-repo_server/+/837587 | 17:15 |
| opendevreview | Jonathan Rosser proposed openstack/openstack-ansible-repo_server master: Remove all code for lsync, rsync and ssh https://review.opendev.org/c/openstack/openstack-ansible-repo_server/+/837588 | 17:15 |
| opendevreview | Jonathan Rosser proposed openstack/openstack-ansible-repo_server master: Remove all code for lsync, rsync and ssh https://review.opendev.org/c/openstack/openstack-ansible-repo_server/+/837588 | 17:15 |
| SiavashSardari | I couldn't find anything else on ansible facts or even os dev files. | 17:15 |
| opendevreview | Jonathan Rosser proposed openstack/ansible-role-systemd_mount master: Add support for glusterfs mounts https://review.opendev.org/c/openstack/ansible-role-systemd_mount/+/837750 | 17:15 |
| opendevreview | Merged openstack/openstack-ansible master: Mark OSA repository as safe in git.config in CI https://review.opendev.org/c/openstack/openstack-ansible/+/837706 | 17:32 |
| opendevreview | Dmitriy Rabotyagov proposed openstack/openstack-ansible stable/xena: Mark OSA repository as safe in git.config in CI https://review.opendev.org/c/openstack/openstack-ansible/+/837686 | 19:38 |
| opendevreview | Dmitriy Rabotyagov proposed openstack/openstack-ansible stable/wallaby: Mark OSA repository as safe in git.config in CI https://review.opendev.org/c/openstack/openstack-ansible/+/837687 | 19:38 |
| opendevreview | Dmitriy Rabotyagov proposed openstack/openstack-ansible stable/victoria: Mark OSA repository as safe in git.config in CI https://review.opendev.org/c/openstack/openstack-ansible/+/837688 | 19:38 |
| *** dviroel is now known as dviroel|afk | 20:22 | |
| *** arxcruz is now known as arxcruz|out | 20:28 | |
Generated by irclog2html.py 2.17.3 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!