| *** ysandeep|out is now known as ysandeep | 01:38 | |
| *** ysandeep is now known as ysandeep|lunch | 07:34 | |
| kleini | I am on V on top of Bionic. I need to a compute node. Would it be possible to deploy a V compute node on top of Focal without having infra nodes yet upgraded to Focal? | 08:44 |
|---|---|---|
| noonedeadpunk | kleini: well, you would need to disable wheels built for this node then | 09:09 |
| noonedeadpunk | kleini: as a side effect with decent amount of computes this might result in dosing opendev gitea as each compute will pull from git independently | 09:10 |
| noonedeadpunk | *DOS-ing | 09:10 |
| noonedeadpunk | actually nice thing about bionic/focal is that live migrations between such computes working perfectly... | 09:12 |
| kleini | as I have different CPUs in compute nodes, I have to configure some common CPU model first before I can live migrate... my cluster is small: only 7 computes then. this should not stress any Git repository too much. | 09:16 |
| noonedeadpunk | yep, it should not. so for host_vars of that new compute, until you get repo server on focal, you can define `venv_wheel_build_enable: false` | 09:18 |
| jrosser | this logging doesn't seem to work https://github.com/openstack/openstack-ansible/blob/master/scripts/log-collect.sh#L156-L167 | 09:23 |
| jrosser | I wonder what this really means for a wildcarded path https://github.com/openstack/openstack-ansible/blob/master/scripts/log-collect.sh#L114 | 09:23 |
| noonedeadpunk | well, sqlite is not generated by ara nowadays, that's true | 09:25 |
| noonedeadpunk | for netstat, we collect ss nowadays so that is replaced | 09:26 |
| noonedeadpunk | oh, yeah, wildcard can't work there | 09:27 |
| noonedeadpunk | unless there's just singe match | 09:27 |
| noonedeadpunk | which I kind of doubt | 09:27 |
| jrosser | I think also we may change the paths to the .txt files in the repo server too | 09:28 |
| noonedeadpunk | so instead we should do smth like `for file in $(ls -1 /openstack/*repo*/repo/os-releases/*/*/*.txt); do store_artifacts ${file} "${WORKING_DIR}/repo"; done | 09:29 |
| jrosser | maybe - store_artifacts is an rsync so I see why passing the wildcard and letting rsync do it all in one go is a good idea | 09:30 |
| noonedeadpunk | well ,then test doesn't make sense.... | 09:31 |
| jrosser | yeah | 09:31 |
| jrosser | anyway, about https://review.opendev.org/c/openstack/openstack-ansible/+/821245 | 09:31 |
| jrosser | are we ready with that? | 09:32 |
| noonedeadpunk | well, I'd rather backport config_template and galera ssl, and just release from rc1, but we can merge it and backport afterwards, and do rc2 next week | 09:34 |
| jrosser | how long do we have? | 09:35 |
| noonedeadpunk | 16 december is deadline | 09:35 |
| noonedeadpunk | week? | 09:35 |
| jrosser | ok - good | 09:35 |
| noonedeadpunk | oh, well, and glance ofc | 09:36 |
| jrosser | we don't have Andrew today to review | 09:38 |
| *** ysandeep|lunch is now known as ysandeep | 09:41 | |
| opendevreview | Merged openstack/openstack-ansible-os_neutron stable/xena: Database connection pooling improvements https://review.opendev.org/c/openstack/openstack-ansible-os_neutron/+/821251 | 09:41 |
| masterpe[m] | I'm using osa 21.2.6 that is running pyroute2 version 0.5.11 I think we are facing https://github.com/svinota/pyroute2/issues/712 issue, As a read the change log we should bump this at least to version 0.5.14 or version 0.5.15. How can I do this? | 10:33 |
| *** ysandeep is now known as ysandeep|afk | 10:35 | |
| noonedeadpunk | masterpe[m]: so eventually this comes from https://opendev.org/openstack/requirements/src/branch/stable/ussuri/upper-constraints.txt#L263 | 10:40 |
| masterpe[m] | yes, so I can change that. And then rerun the neutron playbook? | 10:40 |
| noonedeadpunk | but you can override this by adjusting https://opendev.org/openstack/openstack-ansible/src/branch/master/global-requirement-pins.txt | 10:41 |
| masterpe[m] | ah ok, nice to know thanks. But how to get the new version installed, Do I only need to run the neutron playbook? | 10:42 |
| noonedeadpunk | yep. might be you will need to provide `-e venv_rebuild=true` | 10:43 |
| jrosser | make sure you have this patch if you rebuild the venvs https://github.com/openstack/ansible-role-python_venv_build/commit/73ac14dfaa5b7df9927190b0a5e66de11d5baac5 | 10:45 |
| noonedeadpunk | oh, yes | 10:45 |
| masterpe[m] | Thanks | 10:53 |
| opendevreview | Dmitriy Rabotyagov proposed openstack/openstack-ansible master: Ensure prettytable is present in the ansible-runtime virtualenv https://review.opendev.org/c/openstack/openstack-ansible/+/777061 | 10:54 |
| *** ysandeep|afk is now known as ysandeep | 11:04 | |
| opendevreview | Merged openstack/openstack-ansible-os_tempest master: Use config_template as a collection https://review.opendev.org/c/openstack/openstack-ansible-os_tempest/+/819876 | 11:12 |
| noonedeadpunk | jamesdenton: spotz if you around and have some time for reviews, would be great to get some votes on https://review.opendev.org/c/openstack/openstack-ansible/+/807880, https://review.opendev.org/c/openstack/openstack-ansible-os_glance/+/821097 https://review.opendev.org/c/openstack/openstack-ansible/+/821096 https://review.opendev.org/c/openstack/openstack-ansible/+/821001 | 11:24 |
| *** dviroel|out is now known as dviroel | 11:26 | |
| opendevreview | Dmitriy Rabotyagov proposed openstack/ansible-role-pki master: Allow to provide conditions for certificates https://review.opendev.org/c/openstack/ansible-role-pki/+/821023 | 11:47 |
| opendevreview | Dmitriy Rabotyagov proposed openstack/openstack-ansible-os_neutron master: Implement uWSGI for neutron-api https://review.opendev.org/c/openstack/openstack-ansible-os_neutron/+/486156 | 12:11 |
| opendevreview | Dmitriy Rabotyagov proposed openstack/openstack-ansible-os_neutron master: Implement uWSGI for neutron-api https://review.opendev.org/c/openstack/openstack-ansible-os_neutron/+/486156 | 12:11 |
| noonedeadpunk | just in case we've figured weird behaviour of new lxc, so you might be interested https://github.com/lxc/lxc/issues/4046 | 12:23 |
| noonedeadpunk | this could be workarounded with piping just in case | 12:24 |
| jrosser | well that's a totally unexpected side effect of using < | 12:34 |
| jrosser | surprised that's even possible tbh | 12:35 |
| noonedeadpunk | well yeah... | 12:36 |
| noonedeadpunk | I was surprised as well to see this way in scripts that stopepd working after ubuntu upgrade ( | 12:36 |
| noonedeadpunk | but honestly I think was pretty nice until it worked :) | 12:42 |
| noonedeadpunk | In terms that you don't need to call cat for pipe, just clean feed of data | 12:43 |
| opendevreview | Dmitriy Rabotyagov proposed openstack/openstack-ansible-os_tempest stable/xena: Use config_template as a collection https://review.opendev.org/c/openstack/openstack-ansible-os_tempest/+/821263 | 12:57 |
| spotz | noonedeadpunk: opened will go through them in a bit | 13:01 |
| noonedeadpunk | thanks! much appreciated! | 13:02 |
| opendevreview | Dmitriy Rabotyagov proposed openstack/openstack-ansible master: Add Vault role support https://review.opendev.org/c/openstack/openstack-ansible/+/800787 | 13:34 |
| admin1 | quick question guys .. does a dist-upgrade from 18 -> 20 hamper osa environment ? | 13:47 |
| admin1 | or do we need to format one controller and use focal first .. and then once the hosts-infra is done, can do in the rest ? | 13:48 |
| noonedeadpunk | I think you can do both ways | 13:49 |
| noonedeadpunk | the only thing with upgrade, is that you would need to re-create containers as well, since they will remain running bionic | 13:49 |
| opendevreview | Merged openstack/openstack-ansible-os_tempest stable/xena: Use config_template as a collection https://review.opendev.org/c/openstack/openstack-ansible-os_tempest/+/821263 | 13:53 |
| opendevreview | Merged openstack/ansible-role-uwsgi master: Add support for TLS to UWSGI https://review.opendev.org/c/openstack/ansible-role-uwsgi/+/820532 | 13:55 |
| jrosser | noonedeadpunk: did you try the rejectattr code wit the pki role - I can never remember if it returns a generator and needs "| list" | 14:06 |
| jrosser | it looks like it does the right thing in the tests | 14:10 |
| noonedeadpunk | I tried it with some dummy code and it returned list. at least smth printable with debug:msg | 14:22 |
| noonedeadpunk | I can recall close to only map that requires list afterwards | 14:23 |
| noonedeadpunk | and we used rejectattr a lot with map, so that's why we have list there | 14:23 |
| noonedeadpunk | (well, selectattr) | 14:23 |
| noonedeadpunk | but also - we use these variables only loop, so even if it's generator - it's even better as it should save up us some memory :) | 14:25 |
| noonedeadpunk | *only for loops | 14:25 |
| opendevreview | Dmitriy Rabotyagov proposed openstack/ansible-role-python_venv_build master: Verify that wheels are normally built https://review.opendev.org/c/openstack/ansible-role-python_venv_build/+/821415 | 15:00 |
| noonedeadpunk | jrosser: wdyt? ^ | 15:00 |
| noonedeadpunk | except "security by obscurity" principle for venv_wheel_build_skip_check haha | 15:01 |
| cardiff-dan | hey @jamesdenton I was just wondering whether the bug with network traffic of compute instances using an IP in the range 10.0.3.0/24 getting lost in the nether has been able to be logged yet? | 15:07 |
| *** ysandeep is now known as ysandeep|out | 16:04 | |
| opendevreview | Merged openstack/openstack-ansible master: Add nfs deployment scenario https://review.opendev.org/c/openstack/openstack-ansible/+/821096 | 16:16 |
| admin1 | hi all .. my seutp hosts and setup infra runs just fine .. all good . but setup openstack stalls at /var/www/repo/os-releases/23.1.2/ubuntu-20.04-x86_64/requirements/keystone-23.1.2-constraints.tx | 16:46 |
| admin1 | do i have to delete and redo the repo containers ? | 16:47 |
| admin1 | there is keystone global-constraints . requirements and source-constraints | 16:48 |
| admin1 | 3 fils | 16:48 |
| admin1 | files* | 16:48 |
| admin1 | but not one named just keystone-constraints | 16:48 |
| admin1 | keystone-23.1.2-global-constraints.txt keystone-23.1.2-requirements.txt keystone-23.1.2-source-constraints.txt == there .. keystone-23.1.2-constraints.txt = not there | 16:49 |
| *** dviroel is now known as dviroel|afk | 16:50 | |
| jrosser | admin1: something like that generally means you've previously had an error which you've tried to move past rather than resolve | 16:54 |
| jrosser | you can remove the .txt files for keystone from the repo server and re-run the keystone playbook | 16:54 |
| admin1 | jrosser . the playbook is looking for keystone-23.1.2-constraints.txt .. while its not there .. | 17:02 |
| admin1 | only the other 3 files are there , but not the one that the playbook is checking | 17:02 |
| admin1 | the only error i got was during galera .. but that i have fixed | 17:03 |
| admin1 | the error is in the step: python_venv_build : Slurp up the constraints file for later re-deployment .... question is .. if something might have happened . how do i fix it .. i have re-ran bootstrap and setup hosts-infra many times .. .. what i have not tried is to nuke the repo containers and rebuild them | 17:05 |
| admin1 | nuking the repo containers and retrying .. | 17:10 |
| noonedeadpunk | So eventually, when one of the files is missing, this should trigger wheels build that should result in venv re-installation https://opendev.org/openstack/ansible-role-python_venv_build/src/branch/master/tasks/python_venv_wheel_build.yml#L128 | 17:15 |
| noonedeadpunk | ah, well, you failed even before that... | 17:18 |
| admin1 | i found the error .. but no solution .. https://paste.opendev.org/show/811612/ | 17:18 |
| admin1 | something to do with the ssl cert .. | 17:19 |
| admin1 | shouldn't a bootstrap and setup-hosts fix this ? | 17:19 |
| noonedeadpunk | only if you already have https://opendev.org/openstack/openstack-ansible-openstack_hosts/commit/6fcba602b634bd505c45afdb5e6ad2983df72cf0 | 17:20 |
| noonedeadpunk | or, your lxc image is fresh enough | 17:20 |
| noonedeadpunk | (as it should get latest packages I believe | 17:20 |
| admin1 | i am going to rm -rf /etc/ansible and bootstrap it again | 17:22 |
| admin1 | being on the 23.1.2 tag | 17:22 |
| -opendevstatus- NOTICE: The Gerrit service on review.opendev.org is being quickly restarted for a configuration adjustment, and should return momentarily | 17:25 | |
| admin1 | noonedeadpunk, its not there | 17:28 |
| admin1 | was it fixed after 23.1.2 tag ? | 17:28 |
| noonedeadpunk | yep, I think it;'s not released yet | 17:29 |
| admin1 | ok | 17:29 |
| noonedeadpunk | eventually you can jsut run simple ad-hoc | 17:29 |
| noonedeadpunk | to update ca-certificates everywhere | 17:29 |
| admin1 | in that specific container which has errored: ca-certificates is already the newest version (20210119~20.04.2 | 17:30 |
| admin1 | the lxc-container is focal | 17:30 |
| noonedeadpunk | as patch says - it's always good to have latest ca-certificates :) | 17:38 |
| noonedeadpunk | um, I think this task is delegated to localhost? or one of the repo containers | 17:39 |
| noonedeadpunk | it's not executed on keystone I believe | 17:39 |
| admin1 | i am still under maintenance window .. so will nuke all containers except mariadb and redo it :) | 17:50 |
| jrosser | admin1: the constraints file would be missing because the wheel build delegated from keystone to the repo container failed | 17:52 |
| jrosser | and it would have failed because the ca-certificates package on the repo container was out of date | 17:52 |
| jrosser | this would have been findable in the wheel build log on the repo container | 17:53 |
| jrosser | it would be worth learning how to use ansible ad-hoc commands for things like this https://docs.openstack.org/openstack-ansible/latest/admin/maintenance-tasks/ansible-modules.html | 17:54 |
| admin1 | i already tried ansible -m shell -a "apt update && apt -y upgrade ca-certificates" | 17:55 |
| admin1 | i am going to go through the logs and retry and fix 1 by 1 | 17:56 |
| noonedeadpunk | well, you should consider using native things ;) like `ansible -m package -a "name=ca-certificates update_cache=true state=latest only_upgrade=yes` | 18:28 |
| noonedeadpunk | admin1: also I don't think apt upgrade accepts arguments or package names? | 18:29 |
| noonedeadpunk | dnf/yum does, not apt | 18:29 |
| carlosmss | Hi guys. I've a bug with Cinder in Openstack-Ansible (Stein), the Unity EMC driver randomly delivers a wrong iscsi path as reported here https://bugs.launchpad.net/cinder/+bug/1940859. | 18:45 |
| noonedeadpunk | hey | 18:46 |
| opendevreview | Merged openstack/openstack-ansible-os_glance master: Replace NFS test with integrated one https://review.opendev.org/c/openstack/openstack-ansible-os_glance/+/821097 | 18:46 |
| noonedeadpunk | tbh I never used EMC and has pretty distinct understanding of what is that | 18:47 |
| opendevreview | Dmitriy Rabotyagov proposed openstack/openstack-ansible-os_glance stable/xena: Replace NFS test with integrated one https://review.opendev.org/c/openstack/openstack-ansible-os_glance/+/821451 | 18:47 |
| jrosser | carlosmss: I think the reply from the cinder team means that a lot of issues related to multipath will have been fixed in more recent code | 19:17 |
| jrosser | unless there's a really obvious bug report and fix that has not been backported to stein it's going to be tricky | 19:18 |
| *** dviroel|afk is now known as dviroel | 19:19 | |
| carlosmss | Thanks jrosser. How Can I ask for a revision? I understand that the solution to this bug isn't in the code yet. | 19:21 |
| noonedeadpunk | carlosmss: well, I think that nobody will patch stein, but what you describe sounds like security issue to me tbh, which can be covered with EM. | 19:23 |
| noonedeadpunk | However it's worth to talk with #openstack-cinder team | 19:23 |
| noonedeadpunk | Because eventually nobody care about stein nowadays. Even Ussuri now is under EM which means it will recieve super limited amount of support | 19:24 |
| noonedeadpunk | So my best advice here would probably be test in more recent releases and do upgrade your deployment | 19:24 |
| jrosser | carlosmss: can you test the same thing with an openstack-ansible AIO against your EMC storage? | 19:25 |
| noonedeadpunk | Eventually, you can also try upgrading _onnly_ cinder but that's not smth recommended | 19:25 |
| admin1 | my upgrade worked :) | 19:41 |
| admin1 | thanks for all the inputs and i wish a great weekend | 19:41 |
| jrosser | admin1: great news! nice one | 19:43 |
| noonedeadpunk | so you're on W now? | 19:44 |
| jrosser | would be good to get any feedback on how it worked out with supplying your own cert/key for haproxy as well | 19:45 |
| opendevreview | Dmitriy Rabotyagov proposed openstack/openstack-ansible-os_glance stable/xena: Replace NFS test with integrated one https://review.opendev.org/c/openstack/openstack-ansible-os_glance/+/821451 | 19:46 |
| opendevreview | Dmitriy Rabotyagov proposed openstack/openstack-ansible-os_glance stable/xena: Update TOX_CONSTRAINTS_FILE for stable/xena https://review.opendev.org/c/openstack/openstack-ansible-os_glance/+/820762 | 19:47 |
| carlosmss | Thanks for your advice noonedeadpunk. I will talk with cinder team. | 19:48 |
| carlosmss | jrosser I can't test this with my storage 'cause it is a production environment. And it hasn't available pools. | 19:49 |
| noonedeadpunk | well, these proprietary solutions.... | 19:50 |
| noonedeadpunk | jrosser: so except glance we left only 2 patches to integrated repo that are already in gates I believe? | 19:51 |
| jrosser | this isin the gate https://review.opendev.org/c/openstack/openstack-ansible/+/821001/ | 19:55 |
| admin1 | noonedeadpunk, this platform is on W now .. yes | 19:56 |
| jrosser | this is in check https://review.opendev.org/807880 but has votes for +W | 19:56 |
| noonedeadpunk | well, yeah... it failed to had to recheck. I was more kind of double checking if I might forgot anything | 19:59 |
| noonedeadpunk | oh, woul be great to have docs James wrote https://review.opendev.org/c/openstack/openstack-ansible/+/820370 | 20:00 |
| noonedeadpunk | but we can backportt them for sure | 20:00 |
| jrosser | same for https://review.opendev.org/c/openstack/openstack-ansible/+/820586 to keep the os_neutron ci happy | 20:01 |
| noonedeadpunk | admin1: and have you managed to upgrade the one from Ocata or smth? | 20:01 |
| noonedeadpunk | jrosser: oh, well, it's more needed if we merged uwsgi... I guess I need to -W it... | 20:02 |
| jrosser | maybe thats ok to leave for Y, the ovn situation is a mess there | 20:03 |
| noonedeadpunk | https://review.opendev.org/c/openstack/openstack-ansible-os_neutron/+/486156 is passing, but calico and ovn are not uwsgi | 20:03 |
| noonedeadpunk | if only for Y.... | 20:03 |
| jrosser | there was some ML discussion about things which dont work with uwsgi, an i think ovn was not mentioned there | 20:04 |
| noonedeadpunk | having uwsgi would be so neat with regards to tls topic | 20:04 |
| jrosser | agreed | 20:05 |
| noonedeadpunk | probably spatel would be happy to have https://review.opendev.org/c/openstack/openstack-ansible-os_neutron/+/814825 :D | 20:06 |
| spatel | noonedeadpunk yes i will :) | 20:06 |
| jrosser | spatel: have you tested it? | 20:07 |
| noonedeadpunk | have you tested this final version btw? | 20:07 |
| jrosser | :) | 20:07 |
| spatel | noonedeadpunk no because my lab not setup for DPDK so my plan is to setup DPDK compute node next week and test this new patch | 20:08 |
| noonedeadpunk | spotz jrosser: except xena branching would be awesome to land updates to version bumps https://review.opendev.org/q/topic:"uca_release"+(status:open) | 20:12 |
| jrosser | jamesdenton: have you tried https://review.opendev.org/c/openstack/openstack-ansible-os_neutron/+/814825 ? | 20:13 |
| spotz | looking | 20:15 |
| noonedeadpunk | sorry for pinging too often today but we _really_ need you :) | 20:15 |
| -opendevstatus- NOTICE: The Gerrit service on review.opendev.org is being restarted again for a plugin change, and should be back shortly | 20:16 | |
| spotz | noonedeadpunk: No please do! I have so many channels I get lost between here, Libera, multiple slacks, and discord | 20:17 |
| opendevreview | Merged openstack/openstack-ansible-os_glance stable/xena: Replace NFS test with integrated one https://review.opendev.org/c/openstack/openstack-ansible-os_glance/+/821451 | 20:17 |
| opendevreview | Merged openstack/openstack-ansible-os_glance stable/xena: Update TOX_CONSTRAINTS_FILE for stable/xena https://review.opendev.org/c/openstack/openstack-ansible-os_glance/+/820762 | 20:17 |
| spotz | Need Gerrit back up:( | 20:17 |
| noonedeadpunk | yeah, that's unfortunate( | 20:17 |
| jrosser | massive impact for ubuntu distro installs | 20:18 |
| noonedeadpunk | and might be not limited to it.... | 20:19 |
| noonedeadpunk | as it also has qemu/ovs I believe? | 20:19 |
| jrosser | but i think that anyone who really wanted that would already notice they are running the wrong version | 20:19 |
| noonedeadpunk | I verified that qemu is same vesion though | 20:19 |
| jrosser | and ceph | 20:19 |
| noonedeadpunk | but ceph I think still be installed from external repo? unless there're some race conditions on dependncies and version numbers ofc | 20:20 |
| noonedeadpunk | I will also do minor release to underline that one more time | 20:20 |
| jrosser | i think we have an apt priority pin to prefer upstream ceph rather than uca | 20:20 |
| noonedeadpunk | ah, yes, that's why this stopped happening :) | 20:21 |
| jrosser | this happened before where an upgrade turned into a ceph downgrade, or something | 20:21 |
| noonedeadpunk | yeah, I can recall fighting with that | 20:21 |
| admin1 | noonedeadpunk the one from ocata is upgraded .. there is one more with rocky | 20:29 |
| noonedeadpunk | oh, cool | 20:34 |
| opendevreview | Merged openstack/openstack-ansible master: Remove retrievement of config_tempalte as a module https://review.opendev.org/c/openstack/openstack-ansible/+/821001 | 20:41 |
| spotz | gerrit is back:) | 20:43 |
| opendevreview | Dmitriy Rabotyagov proposed openstack/openstack-ansible master: Freeze roles for Xena RC1 https://review.opendev.org/c/openstack/openstack-ansible/+/821245 | 21:01 |
| opendevreview | Dmitriy Rabotyagov proposed openstack/openstack-ansible master: Freeze roles for Xena RC1 https://review.opendev.org/c/openstack/openstack-ansible/+/821245 | 21:12 |
| opendevreview | Dmitriy Rabotyagov proposed openstack/openstack-ansible master: Start Yoga development cycle https://review.opendev.org/c/openstack/openstack-ansible/+/821476 | 21:15 |
| opendevreview | Dmitriy Rabotyagov proposed openstack/openstack-ansible master: Start Yoga development cycle https://review.opendev.org/c/openstack/openstack-ansible/+/821476 | 21:19 |
| opendevreview | Dmitriy Rabotyagov proposed openstack/openstack-ansible master: Start Yoga development cycle https://review.opendev.org/c/openstack/openstack-ansible/+/821476 | 21:21 |
| opendevreview | Dmitriy Rabotyagov proposed openstack/openstack-ansible master: Start Yoga development cycle https://review.opendev.org/c/openstack/openstack-ansible/+/821476 | 21:22 |
| opendevreview | Dmitriy Rabotyagov proposed openstack/openstack-ansible master: Start Yoga development cycle https://review.opendev.org/c/openstack/openstack-ansible/+/821476 | 21:22 |
| opendevreview | Dmitriy Rabotyagov proposed openstack/openstack-ansible master: Freeze roles for Xena RC1 https://review.opendev.org/c/openstack/openstack-ansible/+/821245 | 21:23 |
| noonedeadpunk | god damn it | 21:23 |
| opendevreview | Dmitriy Rabotyagov proposed openstack/openstack-ansible master: Start Yoga development cycle https://review.opendev.org/c/openstack/openstack-ansible/+/821476 | 21:26 |
| noonedeadpunk | spotz: last 2 things for today:) https://review.opendev.org/c/openstack/openstack-ansible/+/821245 and https://review.opendev.org/c/openstack/openstack-ansible/+/821476/6 | 21:28 |
| noonedeadpunk | jrosser: when around ^ | 21:28 |
| spotz | noonedeadpunk: looking | 21:32 |
| spotz | noonedeadpunk: Any reason why you're calling it Xena series? | 21:34 |
| noonedeadpunk | Um, just following the "template" | 21:35 |
| noonedeadpunk | or let's put it other way - copypasting | 21:35 |
| noonedeadpunk | might be because for Xena we will have not only 24.0.0 but series of releases | 21:35 |
| spotz | ok | 21:36 |
| noonedeadpunk | (as well as for all other releases) | 21:36 |
| noonedeadpunk | if that's what you was asking ) | 21:36 |
| spotz | yep:) | 21:37 |
| spatel | Folk, what options you guys use with image upload, this is what i do - glance image-create --name ubuntu --disk-format raw --container-format bare --property hw_vif_multiqueue_enabled=true --property hw_scsi_model=virtio-scsi --property hw_disk_bus=scsi --visibility public --file focal-server-cloudimg-amd64-disk-kvm.img --progress | 21:38 |
| noonedeadpunk | Looks pretty fair except I use `openstack image create` isntead of `glance image-create` :D | 21:39 |
| spatel | +1 | 21:39 |
| noonedeadpunk | also depends on if you want to have numa pinning or pcpu instead of vcpu or not | 21:39 |
| spatel | hmm why numa comes in picture for image ? | 21:40 |
| noonedeadpunk | iirc it can as well as for flavor | 21:40 |
| spatel | oh! mostly i have flavor for that | 21:40 |
| noonedeadpunk | you have `hw_cpu_policy`, `hw_cpu_sockets`, etc | 21:41 |
| noonedeadpunk | and also can provide traits for image, so yeah | 21:42 |
| spatel | general='true', hw:cpu_policy='dedicated', hw:cpu_sockets='2', hw:cpu_threads='2', hw:mem_page_size='large' | 21:42 |
| spatel | all my flavor has this setting | 21:43 |
| spatel | let me know if i am miss out anything? | 21:43 |
| noonedeadpunk | looks fair enough | 21:47 |
| spatel | I am not using ceph in that case should i use .raw image or .qcow2 ? | 21:47 |
| noonedeadpunk | um, you use nfs? | 21:47 |
| noonedeadpunk | or? | 21:47 |
| spatel | This is small deployment and we have local disk for glance | 21:47 |
| noonedeadpunk | I'd say its worth to be qcow then as it will have smaller size anyway | 21:48 |
| spatel | but what about performance? | 21:48 |
| noonedeadpunk | raw is preffered for ceph as it's amazingly fast to create a drive from raw image even if it's in another pool | 21:48 |
| noonedeadpunk | and what cinder is using? | 21:49 |
| noonedeadpunk | and nova? | 21:49 |
| noonedeadpunk | as if they also store in files, then qcow would be even faster | 21:49 |
| spatel | This is very small deployment we only need CPU/memory for Audio traffic processing. | 21:49 |
| *** dviroel is now known as dviroel}out | 21:49 | |
| spatel | I don't have cinder or any kind of storage related stuff | 21:50 |
| *** dviroel}out is now known as dviroel|out | 21:50 | |
| spatel | vms will run on local-disk on compute (no live migration required) | 21:50 |
| noonedeadpunk | ok, then it totally should be qcow imo | 21:50 |
| noonedeadpunk | as nova will store in qcow | 21:50 |
| spatel | oh! ok then let me convert raw into qcow2 | 21:50 |
| noonedeadpunk | and having raw images would just add delay for data transfer and conversion to qcow | 21:50 |
| spatel | I heard qcow is slow for vm performance | 21:51 |
| noonedeadpunk | + consume more capacity | 21:51 |
| spatel | i am with you | 21:51 |
| spatel | let me convert image in qcow2 | 21:51 |
| spatel | in that case my glance option should be --disk-format qcow2 correct? | 21:52 |
| noonedeadpunk | yeah | 21:52 |
| noonedeadpunk | I will be signing out - almost midnight here :) | 21:52 |
| spatel | :) have a good weekend | 21:52 |
| opendevreview | Dmitriy Rabotyagov proposed openstack/openstack-ansible master: Set galera to use TLS for connections by default https://review.opendev.org/c/openstack/openstack-ansible/+/807880 | 21:54 |
| noonedeadpunk | thanks, you too! | 21:54 |
| opendevreview | James Denton proposed openstack/openstack-ansible-os_neutron master: Add Support for DPDK Bonding https://review.opendev.org/c/openstack/openstack-ansible-os_neutron/+/814825 | 22:45 |
| jamesdenton | @jrosser yes, i just tested it and implemented the suggestion. thanks noonedeadpunk | 22:45 |
| opendevreview | James Denton proposed openstack/openstack-ansible-os_neutron master: Do not set Open vSwitch hostname https://review.opendev.org/c/openstack/openstack-ansible-os_neutron/+/793009 | 22:51 |
| opendevreview | Merged openstack/openstack-ansible stable/wallaby: Update openstack-hosts version https://review.opendev.org/c/openstack/openstack-ansible/+/821236 | 23:05 |
Generated by irclog2html.py 2.17.2 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!