| opendevreview | Merged openstack/openstack-ansible master: Add tempest tests for Blazar https://review.opendev.org/c/openstack/openstack-ansible/+/904786 | 02:24 |
|---|---|---|
| gokhan | noonedeadpunk, I achieved to upgrade from victoria (22.1.0) to Antelope(27.3.0). I only encountered SSL errors in RabbitMQ and Galera because of my configuration in user_variables. I did not encounter any other critical errors. I have also get mount state errors when installing glusterfs but it doesn't affect installation. https://paste.openstack.org/show/bwJxePc8S7Xp7Jt2wAJU/ | 06:27 |
| jrosser | gokhan: I think that task is expected to fail, it’s in an ansible block/rescue so the following task does the needed step | 07:59 |
| jrosser | it is unfortunate when using block/rescue that ansible reports a failure when the whole purpose is to provide a way to permit one task to fail and have a way to handle it properly | 08:02 |
| noonedeadpunk | +1 ^ | 08:57 |
| noonedeadpunk | gokhan: that's really nice to hear | 08:59 |
| admin1 | nice | 09:40 |
| admin1 | gokhan did you do every release, or skipped some ? | 09:40 |
| noonedeadpunk | if anybody around - would be great to have another review on https://review.opendev.org/c/openstack/openstack-ansible-os_tempest/+/907180 | 09:56 |
| noonedeadpunk | to ublock ironic/magnum CI | 09:57 |
| andrewbonney | Done | 10:00 |
| opendevreview | Dmitriy Rabotyagov proposed openstack/openstack-ansible master: [doc] Update documentation for galera cluster recovery https://review.opendev.org/c/openstack/openstack-ansible/+/907525 | 10:27 |
| opendevreview | Dmitriy Rabotyagov proposed openstack/ansible-role-systemd_networkd master: Use OriginalName instead of Name in systemd.link https://review.opendev.org/c/openstack/ansible-role-systemd_networkd/+/907527 | 11:16 |
| opendevreview | Dmitriy Rabotyagov proposed openstack/ansible-role-systemd_networkd master: Use OriginalName instead of Name in systemd.link https://review.opendev.org/c/openstack/ansible-role-systemd_networkd/+/907527 | 11:19 |
| opendevreview | Merged openstack/openstack-ansible-os_blazar master: Create required freepool aggregate https://review.opendev.org/c/openstack/openstack-ansible-os_blazar/+/904876 | 13:04 |
| gokhan | noonedeadpunk, jrosser thanks for your help. I will start distribution upgrade from focal to jammy on antelope. | 13:12 |
| noonedeadpunk | gokhan: you should totally check our patches to upgrade documentation | 13:12 |
| noonedeadpunk | as we're in process of upgrade right now as well and there're some findings | 13:13 |
| andrewbonney | Also worth being aware of https://review.opendev.org/c/openstack/openstack-ansible-os_keystone/+/907397 when you get to your keystone primary | 13:13 |
| noonedeadpunk | and also andrewbonney proposed good looking at a glance patch for fernet rotation | 13:13 |
| gokhan | admin1, ı have only skipped zed. I followed this order victoria>wallaby>xena>yoga>antelope | 13:13 |
| gokhan | noonedeadpunk, I have alsoo seen https://review.opendev.org/c/openstack/openstack-ansible/+/906832 | 13:14 |
| noonedeadpunk | gokhan: there was also https://review.opendev.org/c/openstack/openstack-ansible/+/906750/2/doc/source/admin/upgrades/distribution-upgrades.rst | 13:16 |
| noonedeadpunk | and another one was merged on master already | 13:16 |
| noonedeadpunk | https://review.opendev.org/c/openstack/openstack-ansible/+/906360 | 13:16 |
| noonedeadpunk | btw... should we backport these to 2023.1? I guess - would make sense as ppl will look at docs of respected releases? | 13:17 |
| opendevreview | Dmitriy Rabotyagov proposed openstack/openstack-ansible stable/2023.2: [doc] Reffer need of haproxy backend configuration in upgrade guide https://review.opendev.org/c/openstack/openstack-ansible/+/907561 | 13:17 |
| andrewbonney | I think that makes sense | 13:18 |
| gokhan | thanks noonedeadpunk , I will also share my findings. | 13:18 |
| opendevreview | Dmitriy Rabotyagov proposed openstack/openstack-ansible stable/2023.1: [doc] Reffer need of haproxy backend configuration in upgrade guide https://review.opendev.org/c/openstack/openstack-ansible/+/907562 | 13:18 |
| noonedeadpunk | especially given your patch states that it's tested for 2023.1 | 13:18 |
| TheCompWiz | Good morning everone... Quick question. If I created an new LVM volume for cinder, do I need to pre-format it? or will ansible format it for me? | 15:29 |
| * noonedeadpunk never used lvm for cinder in production | 15:44 | |
| admin1 | TheCompWiz, its just a volume .. when its mounted via cinder to an instance, that instance owner decides what filesystem they want to mount it in | 16:07 |
| TheCompWiz | admin1: I thought cinder would setup multiple "block stores" on that volume for each of the instances... am I wrong? | 16:15 |
| jrosser | a volume is a block device, like a disk drive | 16:17 |
| jrosser | cinder will make as many volumes as are needed in the LVM you supply and export them to instances by (for example) iscsi, but that’s invisible to the user | 16:18 |
| opendevreview | Merged openstack/openstack-ansible-os_tempest master: Fix condition of excludelist file removal https://review.opendev.org/c/openstack/openstack-ansible-os_tempest/+/907180 | 16:19 |
| jrosser | in the instance you see a block device get attached which can then be formatted however needed and mounted | 16:19 |
| TheCompWiz | jrosser: I thought you created the "logical volume" (on the vg) Was I wrong in my assumption? | 16:20 |
| jrosser | check the documentation I gave you yesterday | 16:20 |
| jrosser | the cinder config wants the name of the VG, not an LV | 16:21 |
| jrosser | cinder will create volumes on demand in the VG | 16:21 |
| TheCompWiz | jrosser:. Ok. That makes sense. Much of my brain was distorting the documentation to mean what I assumed. | 16:23 |
| admin1 | cinder will present it as a block device | 16:31 |
| admin1 | leaving the user to do whatever he wants with that | 16:31 |
| jrosser | i think we are missing this https://codesearch.opendev.org/?q=ovn-chassis-mac-mappings&i=nope&literal=nope&files=&excludeFiles=&repos= | 17:06 |
| noonedeadpunk | Um, why it founds nothing in neutron? | 17:22 |
| noonedeadpunk | Also I think vlans working nicely in OVN setup | 17:22 |
| noonedeadpunk | Like right now I'm testing octavia on OVN multinode setup | 17:23 |
| noonedeadpunk | Where vlan is used to connect containers with amphoras | 17:23 |
| noonedeadpunk | And that works out of the box | 17:23 |
| noonedeadpunk | jrosser: | 17:23 |
| noonedeadpunk | https://paste.openstack.org/show/bSaHY2AnbBUKgFUJzGRh/ | 17:25 |
| jrosser | noonedeadpunk: I saw it here https://bugs.launchpad.net/neutron/+bug/1995078 | 17:25 |
| noonedeadpunk | yeah, dunno, might be... based on https://www.ovn.org/support/dist-docs/ovn-architecture.7.html there seems to be different designs.... | 17:27 |
| noonedeadpunk | but might be indeed worth checking why we want it... Potentially I jsut haven't faced the wall with OVN deployment I'm playing now... | 17:29 |
| noonedeadpunk | But I can obviously reach amphoras from lxc containers so far | 17:29 |
| noonedeadpunk | but again - gateways in my case are standalone hosts, not computes | 17:29 |
| noonedeadpunk | I guess it's really needed when you do ovn-bridge-mappings | 17:30 |
| noonedeadpunk | ah, we do that as well though... | 17:31 |
| noonedeadpunk | dunno then... | 17:31 |
| noonedeadpunk | I'm apparently very rusty in networking these days :( | 17:34 |
| noonedeadpunk | But Octavia kinda works in a multinode "sandbox" I'm playing with now | 17:36 |
| spatel | Folks... I am running OSA Xena release.. what is the path to upgrade to zed ? | 17:37 |
| spatel | should I upgrade to yoga first or just go to Zed directly? | 17:38 |
| noonedeadpunk | spatel: frankly speaking I never jumped to Zed. We did Xena -> Antelope lately though | 17:43 |
| spatel | can I do Xena to antelop? | 17:43 |
| noonedeadpunk | it was semi-okeyish, basically same as you can expect with such upgrade (or if you was jumping before, like Victoria->Xena or whatever) | 17:44 |
| noonedeadpunk | So what is supported - Yoga -> Antelope. Xena -> Antelope mostly work. There are couple of things you should know though | 17:44 |
| spatel | tell me what are those :) | 17:45 |
| noonedeadpunk | 1. There's a bug in keystone, so implied roles for application credentials do not work | 17:45 |
| noonedeadpunk | 2. They have fixed roles scope in Keystone, so _member_ role does not work anymore. All users must have `member` role, including application credentials and trusts | 17:45 |
| spatel | hmm do I need to make any change for that or just be aware? | 17:47 |
| noonedeadpunk | 3. Based on these 2 notes - you totally need to "hack" db and explicitly assign reader role to all application credentials (and member if it's absent). Since some nova policies requires `reader`. And `reader` implied by `member` so it should have worked, if not keystone bug | 17:47 |
| noonedeadpunk | there's a ML what we did.... | 17:47 |
| noonedeadpunk | let me try to find | 17:47 |
| noonedeadpunk | https://lists.openstack.org/archives/list/openstack-discuss@lists.openstack.org/thread/5STDOMQZMFXTRE7JNE4QFUYFRIS6W5BJ/#QJCNYFXDVPEXXJQOOL33NXVLADHFMCIK | 17:48 |
| noonedeadpunk | and 4. You'll need to update RPC version in mysql to... 64? for computes/scheduler or otherwise they will get into chiken-egg loop | 17:49 |
| noonedeadpunk | first 3 will hit you regardless once you decide to upgrade to antelope | 17:50 |
| noonedeadpunk | 4th can be avoided with Xena -> Yoga -> Antelope | 17:50 |
| noonedeadpunk | `mysql -e "update nova.services set version = 66 where deleted = 0;"` | 17:52 |
| noonedeadpunk | or well. probably now it's not 66, but indeed 4 or smth, once bugfix was backported on nova side for one small bug | 17:53 |
| noonedeadpunk | *64 | 17:53 |
| noonedeadpunk | I was actually going to play with OVN driver for octavia shortly | 18:05 |
| noonedeadpunk | Hopefully land the patch after some polishing.... | 18:06 |
| noonedeadpunk | (or better say edit to https://review.opendev.org/c/openstack/openstack-ansible-os_octavia/+/868462) | 18:06 |
| opendevreview | Merged openstack/openstack-ansible-os_keystone master: Re-distribute fernet keys when re-building the primary https://review.opendev.org/c/openstack/openstack-ansible-os_keystone/+/907397 | 18:58 |
| spatel | noonedeadpunk This is too much :O | 19:34 |
| spatel | noonedeadpunk thanks for the head up.. | 19:35 |
| spatel | I am thinking I can take intermediate path Xena -> Yoga | 19:36 |
| opendevreview | Jonathan Rosser proposed openstack/openstack-ansible master: Allow zuul pre playbook bootstrap to be skipped https://review.opendev.org/c/openstack/openstack-ansible/+/905290 | 19:42 |
| TheCompWiz | anyone know why neutron would only load "flat" and "vlan" drivers? (missing vxlan) | 20:00 |
| TheCompWiz | the container keeps dying with "Feb 02 19:43:12 charlie-neutron-server-container-009d78f2 neutron-server[3642]: 2024-02-02 19:43:12.735 3642 ERROR neutron.plugins.ml2.managers [-] No type driver for tenant network_type: vxlan. Service terminated!" | 20:01 |
| TheCompWiz | I think I found it. Apparently "neutron_ml2_drivers_type" has a default value of "geneve,vlan,flat"... instead of dynamically building that variable based on what is in the conf. ... so you have to manually override it. | 20:25 |
| * TheCompWiz goes back to grumbling and waiting for a full re-install /sigh. | 20:26 | |
| jrosser | TheCompWiz: there is documentation for the different ways you might use neutron here https://docs.openstack.org/openstack-ansible-os_neutron/latest/ | 20:38 |
| jrosser | also if you want to find where/how any of these variables are used you can check here https://codesearch.opendev.org/?q=neutron_ml2_drivers_type | 20:40 |
| spatel | Does cephadm support RockyLinux 9.x ? | 21:01 |
| spatel | I am planning to use RockyLinux for new ceph deployment instead Ubuntu | 21:01 |
| opendevreview | Merged openstack/openstack-ansible-os_trove master: Adopt for usage openstack_resources role https://review.opendev.org/c/openstack/openstack-ansible-os_trove/+/899284 | 21:26 |
| *** priteau_ is now known as priteau | 21:44 | |
| opendevreview | Merged openstack/openstack-ansible master: Adopt magnum test variables for openstack_resources https://review.opendev.org/c/openstack/openstack-ansible/+/901184 | 21:56 |
| opendevreview | Merged openstack/openstack-ansible master: Add openstack-resources playbook https://review.opendev.org/c/openstack/openstack-ansible/+/901460 | 22:07 |
| opendevreview | Jonathan Rosser proposed openstack/openstack-ansible-os_tempest master: Adopt for usage openstack_resources role https://review.opendev.org/c/openstack/openstack-ansible-os_tempest/+/889741 | 22:28 |
Generated by irclog2html.py 2.17.3 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!