| opendevreview | Pierre Riteau proposed openstack/kayobe master: Document seed_enable_snat https://review.opendev.org/c/openstack/kayobe/+/820653 | 06:02 |
|---|---|---|
| opendevreview | Piotr Parczewski proposed openstack/kayobe master: Enable rtcsync in chrony by default https://review.opendev.org/c/openstack/kayobe/+/820657 | 07:27 |
| opendevreview | MargaritaShakhova proposed openstack/kolla-ansible master: Add ironic-inspector policy configuration https://review.opendev.org/c/openstack/kolla-ansible/+/820063 | 07:31 |
| sorin-mihai | o/ | 07:56 |
| sorin-mihai | I'm planning a production deployment that I should complete in a few weeks and should not suffer any major changes for about 3 years at least, and I wonder what is the current 'stable' situation, should I go with xena-source-centos on top of rocky or maybe other combination? | 07:59 |
| sorin-mihai | i converted most of the centos 8 servers rocky and haven't seen any issue with kolla running on top of that. are there some known issues that I might've missed? | 08:01 |
| frickler | yoctozepto: mnasiadka: mgoddard: please have a look at https://review.opendev.org/c/openstack/kolla-ansible/+/820400 , I'm hitting that issue regularly in local deployments, also some hits in CI | 08:02 |
| mnasiadka | sorin-mihai: we don't support Rocky | 08:31 |
| mnasiadka | well, Rocky Linux | 08:31 |
| mnasiadka | Rocky was an OpenStack Release name and it's EOL | 08:32 |
| sorin-mihai | I know, but RHEL is supported, right? | 08:32 |
| yoctozepto | frickler: replied | 08:33 |
| mnasiadka | sorin-mihai: https://docs.openstack.org/kolla-ansible/xena/user/support-matrix.html#supported-operating-systems - deprecated | 08:35 |
| sorin-mihai | rocky linux/alma linux are supposed to be the replacement for previous centos model, if rhel is supported, I don't see why one of these couldn't be at least un-officially supported, since they are drop in replacement for rhel | 08:35 |
| mnasiadka | There is no such thing as unofficial support - if you run Kolla-Ansible on Rocky Linux 8 today - you're on your own. | 08:36 |
| yoctozepto | sorin-mihai: we don't have the manpower nor upstream compute power to support every combination | 08:37 |
| mnasiadka | And with Python3.6 drop in Openstack Z release - I wouldn't be thinking about running something for 3 years on RHEL8 equivalent. | 08:37 |
| sorin-mihai | but nobody is going to use centos stream 8 in production as host os... this pushes out every rpm user | 08:37 |
| opendevreview | Pierre Riteau proposed openstack/kayobe stable/train: Disable failing Python 2 jobs https://review.opendev.org/c/openstack/kayobe/+/820601 | 08:37 |
| mnasiadka | Unless you're not planning to do any upgrades ;-) | 08:37 |
| yoctozepto | mnasiadka: especially considering that ansible has already dropped support for old pythons ;-) | 08:37 |
| mnasiadka | yoctozepto: at least in Z we should have RHEL9/RockyLinux9/whatever - I have no clue what Ansible folks think about their users. | 08:38 |
| yoctozepto | sorin-mihai: you heard our advice, it's simply how it looks now, it's not that we don't support it on exact purpose, we follow the world the best we can | 08:39 |
| yoctozepto | mnasiadka: they think users should stop using old centos (-: | 08:39 |
| sorin-mihai | yoctozepto, i know, just trying to get a better understanding of how things are seen here | 08:40 |
| mnasiadka | and start using CentOS that has been announced days ago? :) | 08:40 |
| yoctozepto | mnasiadka: yeah, that's my understanding of new redhat policy with its software - annoy folks not buying support | 08:40 |
| sorin-mihai | I wonder if this can be a reason to start a discussion with the Rocky Linux team... | 08:43 |
| mnasiadka | sorin-mihai: discuss what? | 08:44 |
| sorin-mihai | maybe consider a SIG focused on kolla, otherwise there'l be only tripleo deployments, if those would be 'official'... | 08:48 |
| yoctozepto | sorin-mihai: oh, now that you mention tripleo, they went rolling release too, no upkeeping of stable branches | 08:50 |
| yoctozepto | thus why I said what I said ;-) | 08:50 |
| sorin-mihai | so then a SIG might be a better option, there are enough users that want to keep things stable | 08:51 |
| hrw | morning | 09:05 |
| hrw | sorin-mihai: you are free to deploy on whatever you want. If you want to keep 3 years without doing any major changes then build own images and keep them in own registry for those 3 years. And every 4-6 months check which fixes went into stable/BRANCH-YOU-USED and consider building newer images and upgrading. | 09:07 |
| hrw | sorin-mihai: worth checking for newer branches too to look for potential fixes which were not backported. | 09:08 |
| hrw | sorin-mihai: up to Xena we support CentOS (Stream 8), Debian (11 aka bullseye) and Ubuntu (20.04) | 09:09 |
| hrw | sorin-mihai: if you deploy on something else then we may or may not support any support claims | 09:10 |
| jingvar | holy wars :) | 09:13 |
| hrw | jingvar: I would rather say 'far too small team' | 09:14 |
| jingvar | I think main point what is supported by OpenStack | 09:14 |
| hrw | whatever with python 3.6+ and working virtualization? | 09:15 |
| hrw | jingvar: note that you can take any host OS and run kolla containers with other OS inside | 09:15 |
| jingvar | I do | 09:15 |
| hrw | Debian on top of Rocky Linux? not supported but probably works | 09:16 |
| hrw | Ubuntu on top of Arch? again, not supported, probably works | 09:16 |
| jingvar | Centos on top Rocky | 09:16 |
| hrw | Rocky is EOL | 09:16 |
| jingvar | Rocky linux | 09:17 |
| hrw | we do not talk Rocky here anymore ;d | 09:17 |
| jingvar | :) | 09:17 |
| rockey | hrw: luckily there's holidays coming up, so you guys can eat up and become a bit bigger as a team :) | 09:17 |
| hrw | rockey: my doc said I should drop some | 09:17 |
| hrw | ;P | 09:17 |
| rockey | :( | 09:17 |
| jingvar | I like Ubuntu but must work with RPM based ... | 09:18 |
| *** hrww is now known as hrw | 09:40 | |
| opendevreview | Piotr Parczewski proposed openstack/kayobe master: Enable rtcsync in chrony by default https://review.opendev.org/c/openstack/kayobe/+/820657 | 09:49 |
| frickler | yoctozepto: mnasiadka: found an old bug for the grafana issue, now we only need to agree on whether and how to make the delay configurable https://bugs.launchpad.net/kolla-ansible/+bug/1769962 | 09:49 |
| *** hrww is now known as hrw | 09:52 | |
| hrw | mnasiadka, mgoddard: trivial reno https://review.opendev.org/c/openstack/kolla/+/820045 | 10:01 |
| hrw | https://review.opendev.org/c/openstack/kolla/+/818095 is also trivial one mnasiadka yoctozepto mgoddard | 10:02 |
| opendevreview | Marcin Juszkiewicz proposed openstack/kolla master: qdrouterd: add release note with deprecation notice https://review.opendev.org/c/openstack/kolla/+/820045 | 10:22 |
| hrw | mgoddard: reno cut | 10:22 |
| mgoddard | hrw: merci | 10:23 |
| hrw | mgoddard: so https://review.opendev.org/c/openstack/kolla/+/818095 next? | 10:34 |
| hrw | kevko: hi | 10:40 |
| hrw | kevko: https://review.opendev.org/c/openstack/kolla/+/795054 - please read my comment and vote | 10:52 |
| hrw | s/vote/consider to change your vote/ | 10:53 |
| admin1 | hi all .. kolla-ansible boostrap runs fine (multiple times) .. prechecks fail on Checking docker SDK version .. ModuleNotFoundError: No module named 'docker' .. any hints ? | 11:04 |
| admin1 | print(docker.__version__) returns 4.4.4 | 11:11 |
| admin1 | issue in "/usr/bin/python3 -c \"import docker; print(docker.__version__)\" ? | 11:12 |
| hrw | https://review.opendev.org/c/openstack/kolla/+/770647 mgoddard mnasiadka | 11:16 |
| hrw | admin1: ah. I thought that we sorted that long time ago ;( | 11:16 |
| hrw | admin1: does 'dnf/apt install python3-docker' solves problem? | 11:17 |
| hrw | on target system | 11:17 |
| admin1 | it did :) | 11:17 |
| admin1 | target system is ubuntu-focal | 11:17 |
| admin1 | just set it up this morning | 11:18 |
| hrw | yoctozepto: https://review.opendev.org/c/openstack/kolla-ansible/+/792082 is still WIP? | 11:18 |
| hrw | admin1: thx | 11:19 |
| admin1 | thanks hrw . | 11:20 |
| admin1 | hrw, can kolla be created with existing database ? | 11:40 |
| admin1 | use case .. to test for scenarios where the dc is gone and we only have backups of the mysql database .. | 11:41 |
| admin1 | dc gone = controllers down .. not the computes | 11:41 |
| admin1 | to simulate a complete controllers failure and attempted recovery from existing backup databases | 11:42 |
| hrw | sorry, but I do not deploy so cannot answer | 11:45 |
| admin1 | hrw, horizon fails continously with /usr/local/bin/kolla_extend_start: line 121: ENABLE_KARBOR: unbound variable .. also another known issue ? | 11:59 |
| admin1 | https://bugs.launchpad.net/kolla/+bug/1911141 - says not to use master branch .. .. https://docs.openstack.org/kolla-ansible/latest/user/quickstart.html uses master branch but uses the globals.yml for the branch name .. which one is correct ? | 12:03 |
| admin1 | i am not using master in the globals.yml . but i did ip install git+https://opendev.org/openstack/kolla-ansible@master ... should the @master at the end be a branch ? | 12:04 |
| priteau | admin1: This is because you are looking at the "latest" doc which maps to the in-development source tree | 12:06 |
| priteau | If you want to use Xena for example, here's the doc: https://docs.openstack.org/kolla-ansible/xena/user/quickstart.html | 12:07 |
| admin1 | got it | 12:07 |
| admin1 | is there a command to destroy it all and start again (without having to reinstall the os) | 12:07 |
| admin1 | i found it | 12:08 |
| admin1 | thanks | 12:08 |
| opendevreview | Marcin Juszkiewicz proposed openstack/kolla-ansible master: [WIP] Python-Docker from distro https://review.opendev.org/c/openstack/kolla-ansible/+/792082 | 12:14 |
| hrw | yoctozepto: added centos-release-openstack-xena to have RDO repo for python3-docker | 12:16 |
| opendevreview | Marcin Juszkiewicz proposed openstack/kolla master: qdrouterd: add release note with deprecation notice https://review.opendev.org/c/openstack/kolla/+/820045 | 12:28 |
| hrw | pep8 | 12:28 |
| hrw | zuul acts weird today | 13:06 |
| admin1 | what is the best method to upgrade ? create another venv ,checkout a higher version and do the kolla-ansible upgrade ? | 13:18 |
| admin1 | i read pull and upgrade .. the system does not need to be re-bootstrapped for the new version ? | 13:19 |
| opendevreview | Piotr Parczewski proposed openstack/kayobe master: CI: Use correct TD agent repository version https://review.opendev.org/c/openstack/kayobe/+/820862 | 13:35 |
| priteau | admin1: occasionally you may need to rerun bootstrap to apply an updated docker configuration for example | 13:49 |
| opendevreview | Merged openstack/kolla stable/xena: Add Swift lock path in Swift containers https://review.opendev.org/c/openstack/kolla/+/818892 | 14:07 |
| opendevreview | Merged openstack/kolla master: Upgrade pip and friends in source images https://review.opendev.org/c/openstack/kolla/+/770647 | 14:07 |
| kevko | some kolla-swift expert ? :P | 14:44 |
| kevko | i've deployed swift from wallaby (and also secondly redeployed with this patch https://review.opendev.org/c/openstack/kolla-ansible/+/797498) and it's not working ..in proxyserver i have errors here -> https://paste.opendev.org/show/811514/ | 14:46 |
| kevko | hrw: is it really needed to touch binary branch of that Dockerfile ? | 14:48 |
| kevko | hrw: you said "Note that this is cleanup and unification of support for two distributions." but that packages will be switched from binary to source ..so that block just dissapear | 14:49 |
| hrw | kevko: we still support binary builds | 14:49 |
| hrw | kevko: once we drop binary images such patches will make no sense. until then they are fine | 14:50 |
| opendevreview | Pierre Riteau proposed openstack/kolla master: Move nvme-cli from binary to source dependencies in nova https://review.opendev.org/c/openstack/kolla/+/820892 | 14:55 |
| opendevreview | Pierre Riteau proposed openstack/kolla master: Ensure nvme-cli is present in nova-compute images https://review.opendev.org/c/openstack/kolla/+/820892 | 15:02 |
| opendevreview | Mark Goddard proposed openstack/ansible-collection-kolla master: Initialise Ansible collection https://review.opendev.org/c/openstack/ansible-collection-kolla/+/820166 | 15:20 |
| opendevreview | Mark Goddard proposed openstack/ansible-collection-kolla master: Sync basics with kolla-ansible https://review.opendev.org/c/openstack/ansible-collection-kolla/+/820167 | 15:20 |
| opendevreview | Mark Goddard proposed openstack/ansible-collection-kolla master: docs: remove most boilerplate, initialise contributor guide https://review.opendev.org/c/openstack/ansible-collection-kolla/+/820560 | 15:20 |
| opendevreview | Mark Goddard proposed openstack/ansible-collection-kolla master: Import baremetal role from kolla-ansible https://review.opendev.org/c/openstack/ansible-collection-kolla/+/820168 | 15:20 |
| opendevreview | Mark Goddard proposed openstack/ansible-collection-kolla master: CI: run ansible-test sanity checks https://review.opendev.org/c/openstack/ansible-collection-kolla/+/820902 | 15:20 |
| opendevreview | Michal Nasiadka proposed openstack/kayobe master: Add dnf versionlock support https://review.opendev.org/c/openstack/kayobe/+/820906 | 15:49 |
| opendevreview | Mark Goddard proposed openstack/ansible-collection-kolla master: CI: add initial deploy jobs https://review.opendev.org/c/openstack/ansible-collection-kolla/+/820907 | 15:49 |
| holtgrewe | Hello, I have a nicely running openstack cluster ... up to the point where all nodes went down at the same time. | 16:51 |
| holtgrewe | Now the mysql servers don't start up again... | 16:52 |
| holtgrewe | (Cluster setup with kolla/kayobe). What is the appropriate way of getting the mariadb database to run again? | 16:53 |
| mgoddard | holtgrewe: kayobe overcloud database recover | 16:58 |
| holtgrewe | mgoddard, thanks | 17:00 |
| guesswhat | jingvar: ping, did you see my last messsage? here is error message from all cinder containers .... ERROR oslo_service.service requests.exceptions.SSLError: HTTPSConnectionPool(host='172.25.254.164', port=2379): Max retries exceeded with url: /v3alpha/lease/grant (Caused by SSLError(SSLCertVerificationError(1, '[SSL: CERTIFICATE_VERIFY_FAILED] certificate verify failed: unable to get | 17:12 |
| guesswhat | local issuer certificate (_ssl.c:1131)'))) | 17:12 |
| guesswhat | not sure why oslo service can not validate certificates in etcd.. | 17:13 |
| holtgrewe | mgoddard, somehow it does not recover properly -- https://paste.openstack.org/show/bF9YvgSotZT69RcTFUUZ/ | 17:30 |
| holtgrewe | ah, it looks like I'm having network issues | 17:33 |
| opendevreview | Merged openstack/kolla stable/victoria: nova-compute: trim image a bit on CentOS https://review.opendev.org/c/openstack/kolla/+/820347 | 18:39 |
| opendevreview | Radosław Piliszek proposed openstack/kolla-ansible master: [docs] Mark init-runonce properly https://review.opendev.org/c/openstack/kolla-ansible/+/820940 | 19:19 |
| admin1 | Inventory's group loadbalancer does not exist or it is empty. Please update inventory, as haproxy group was renamed to loadbalancer in the Xena release. -- where is the inventory file ? and is a manual edit enough ? | 19:51 |
| admin1 | figured it out | 20:00 |
| guesswhat | Guys? Do i need to configure some additional flag to enable TLS cinder->etcd ? theres oslo service calling etcd, but fails to TLSVerificationFailed.. Thanks | 20:04 |
Generated by irclog2html.py 2.17.2 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!