opendevreview | Merged openstack/kolla-ansible master: CI: improve get_logs.sh https://review.opendev.org/c/openstack/kolla-ansible/+/870521 | 06:56 |
---|---|---|
sylvr | Hello ! So yesterday I finally completed a deployment using Kayobe, but now I need to tailor it for the datacenter needs. If I remember correctly, it's the kolla-ansible configuration I need to tweak to add openstack component/use Ceph and such ? | 07:06 |
mnasiadka | sylvr: yes | 07:40 |
sylvr | mnasiadka : thanks ! | 07:41 |
SvenKieske | o/ | 07:56 |
SvenKieske | maybe someone can check these trivial refactorings: https://review.opendev.org/c/openstack/kolla-ansible/+/922618 | 07:56 |
SvenKieske | https://review.opendev.org/c/openstack/kolla-ansible/+/922617 | 07:56 |
SvenKieske | and a missing import: https://review.opendev.org/c/openstack/kolla-ansible/+/922619 | 07:56 |
SvenKieske | mhm, the last one I probably need to fix first :) | 07:57 |
SvenKieske | ok, let me try to get rid of distutils | 08:09 |
opendevreview | Matúš Jenča proposed openstack/kolla-ansible master: CI: Fix upgrade jobs failing on adding new certificates https://review.opendev.org/c/openstack/kolla-ansible/+/926284 | 08:19 |
opendevreview | Sven Kieske proposed openstack/kolla-ansible master: py312 support: remove distutils usage https://review.opendev.org/c/openstack/kolla-ansible/+/926615 | 08:23 |
SvenKieske | while I have your attention: as we have missed imports in the past to add to requirements.txt I was thinking about a small linter that checks all imports against our requirements.txt file, what do you think? | 08:36 |
SvenKieske | I found https://github.com/bndr/pipreqs which should be able to print a diff of found imported modules and the actual requirements.txt file | 08:37 |
SvenKieske | it also can autogenerate the file, but I guess we don't want to do _that_ :) | 08:37 |
opendevreview | Sven Kieske proposed openstack/kolla-ansible master: py312 support: remove distutils usage https://review.opendev.org/c/openstack/kolla-ansible/+/926615 | 08:50 |
opendevreview | Magnus Lööf proposed openstack/kolla-ansible master: Enable TLS backend for designate https://review.opendev.org/c/openstack/kolla-ansible/+/866524 | 08:54 |
kevko | SvenKieske: did you see the answer from legal-discuss ? | 08:59 |
kevko | SvenKieske: "After consulting foundation legal counsel, this type of integration should not be a problem from a licensing perspective as long as no BSL licensed code is added into OpenStack code and as long as use of the HashiCorp product remains optional." | 09:00 |
SvenKieske | kevko: yes I did see it, good to hear the clarification :) | 09:01 |
SvenKieske | once question that arises is the definition of "optional" I guess, but afaik we don't have anywhere hard requirements that are only possible via BSL licenced code? | 09:02 |
SvenKieske | one* | 09:02 |
SvenKieske | so, like some functionality that only works with consul, or vault? Idk, but I also don't really care that much currently. | 09:03 |
SvenKieske | anybody in here ever used https://github.com/ansible/ansible-builder ? | 09:04 |
yusufgungor_ | Hi kolla team. Is there any work to integrate designate with neutron? | 09:55 |
yusufgungor_ | It has been said that : Designate <-> Nova-compute relation is obsolete starting from Queens | 09:55 |
yusufgungor_ | https://bugs.launchpad.net/charm-nova-compute/+bug/1738100#:~:text=Designate%20%3C%2D%3E%20Nova%2Dcompute%20relation%20is%20obsolete%20starting%20from%20Queens%3A | 09:55 |
yusufgungor_ | We have dns record deletion problems when instance deleted and got no support because we are using notifications method | 09:55 |
kevko | SvenKieske: well - it's same as masakari has an option to use consul ...or barbican to use vault ...they are supporting it ... i don't why kolla can't provide an option to install this dependency if other projects can ... What we can add to documentation is note where we declare that this piece of SW is optional and licences under BSL license and | 10:01 |
kevko | is up to deployer to solve law questions ... | 10:01 |
kevko | i don't why kolla can't provide an option to install this dependency if other projects can support them | 10:02 |
SvenKieske | sure, I guess we can. | 10:02 |
kevko | yusufgungor_: https://docs.openstack.org/kolla-ansible/latest/reference/networking/designate-guide.html | 10:03 |
kevko | SvenKieske: I am going to implement it ...as our new customer wants it and I want to then share it with community (because i really don't like to backport everything from version to version :) ) ....as corosync-pacemaker is only for 32 hosts ..and they will have more hypervisors ... | 10:04 |
kevko | SvenKieske: masakari works with corosync-pacemaker but up to 32 hosts ... unfortunatelly ... ^^ this is answer to your question above ... | 10:05 |
opendevreview | Michal Arbet proposed openstack/kolla-ansible master: Add ansible-core as a dependency https://review.opendev.org/c/openstack/kolla-ansible/+/922369 | 10:09 |
opendevreview | Alex Welsh proposed openstack/kolla-ansible stable/2024.1: Fix prometheus.yml templating https://review.opendev.org/c/openstack/kolla-ansible/+/926618 | 10:09 |
opendevreview | Merged openstack/kayobe stable/2023.1: Bump stackhpc.systemd_networkd role to v1.0.6 https://review.opendev.org/c/openstack/kayobe/+/926122 | 10:24 |
opendevreview | Matúš Jenča proposed openstack/kolla-ansible master: Fix upgrade jobs failing on adding new certificates https://review.opendev.org/c/openstack/kolla-ansible/+/926284 | 10:26 |
opendevreview | Mark Goddard proposed openstack/kayobe master: Prevent running from a different Kayobe configuration repository https://review.opendev.org/c/openstack/kayobe/+/903139 | 10:28 |
opendevreview | Mark Goddard proposed openstack/kayobe master: CI: Add openstack-tox-pep8 job, fix pep8 failures https://review.opendev.org/c/openstack/kayobe/+/926619 | 10:28 |
yusufgungor_ | @kevko thanks for the lightning. i have not been aware of neutron_dns_integration parameter because of not exist on Zed documentation. | 10:41 |
kevko | yusufgungor_: well, check the code if there is such parameter ..this was merged to master ... and the logic is right there ...but you can have same config with overrides ... | 11:00 |
yusufgungor_ | @kevko thanks you 🙏 | 11:01 |
kevko | yusufgungor_: i know it because i rewrote it :D | 11:01 |
yusufgungor_ | :D | 11:01 |
yusufgungor_ | Have you ever encountered a situation where DNS records were not automatically deleted when using designate sink and nova notification? | 11:02 |
kevko | yusufgungor_: do you really need sink ? | 11:07 |
kevko | yusufgungor_: but yeah, this can happen if your notifications not work correctly ... rabbitmq issues or ..something ... | 11:08 |
kevko | yusufgungor_: because sink works like that ... | 11:08 |
PrzemekK | Hi | 11:09 |
PrzemekK | related to notification and rabbit You talking about Queue versioned_notifications.info for masakari ? Just joined chat | 11:10 |
PrzemekK | we have same issue https://access.redhat.com/solutions/3139721 | 11:10 |
kevko | PrzemekK: which version of kolla-ansible ? | 11:15 |
PrzemekK | 2024.1 | 11:16 |
kevko | PrzemekK: using kolla-ansible for deployment ? | 11:16 |
kevko | PrzemekK: and using designate-sink ? | 11:17 |
PrzemekK | yes but rabbit is external | 11:17 |
kevko | PrzemekK: did you check the docs for designate ? i mean kolla-ansible's ? | 11:17 |
kevko | PrzemekK: https://docs.openstack.org/kolla-ansible/latest/reference/networking/designate-guide.html | 11:17 |
PrzemekK | Queue versioned_notifications.info have 0 consumers and messages growing like 3000 for week | 11:18 |
kevko | PrzemekK: Service designate-sink in kolla deployments is disabled by default and can be enabled by designate_enable_notifications_sink: yes. <<< | 11:18 |
yusufgungor_ | @kevko in zed version, | 11:21 |
yusufgungor_ | designate_enable_notifications_sink: "{{ enable_designate | bool }}" | 11:21 |
yusufgungor_ | this is the default value in all.yml | 11:21 |
PrzemekK | but we dont use dns/designated and stuck messages are from masakari {"oslo.version": "2.0", "oslo.message": "{\"message_id\": \"ab84916c-7889-40d6-8be7-ace772c33cd9\", \"publisher_id\": \"masakari-engine:xxx\", \"event_type\": \"notification.process.start\", \"priority\": \"INFO\", \"payload\": {\"masakari_object.name\": \"NotificationApiPayload\", \"masakari_object.namespace\": \"masakari\", \"masakari_object.version\": \ | 11:22 |
PrzemekK | https://docs.openstack.org/masakari/latest/configuration/sample_config.html look for notification | 11:23 |
PrzemekK | masakari config [oslo_messaging_notifications] driver = messagingv2 topics = notifications [oslo_messaging_rabbit] heartbeat_in_pthread = False rabbit_quorum_queue = true [oslo_middleware] enable_proxy_headers_parsing = True | 11:24 |
opendevreview | Merged openstack/kolla-ansible master: Fix prometheus.yml templating https://review.opendev.org/c/openstack/kolla-ansible/+/926198 | 11:26 |
kevko | PrzemekK: I know how it works , there are several services or external softwares can read and consume notifications ...designate, CADF, masakari, prometheus ...etc etc ..etc | 11:27 |
kevko | PrzemekK: the issue will be quorum queues or related to this i think | 11:28 |
kevko | maybe .. | 11:28 |
PrzemekK | but workaround from redhat is not good idea - how You think ? Set notification_format in the nova configuration to unversioned on all controllers and compute nodes to not send messages to the versioned queue with no consumers. | 11:29 |
kevko | PrzemekK: what driver do you have in oslo_messaging_notifications ? | 11:32 |
kevko | ah, ok | 11:35 |
kevko | i see | 11:35 |
PrzemekK | its from template kolla-ansible | 11:35 |
PrzemekK | same issue on all eviroments | 11:35 |
kevko | PrzemekK: well - that's the reason why I asked ... https://github.com/openstack/kolla-ansible/blob/4615f247fcf50bcdb5f7329f0bd88b58dac57302/ansible/roles/masakari/templates/masakari.conf.j2#L44-L51 << this is in master ...so if you don't have ceilometer ..there is noop in driver conf option | 11:37 |
kevko | PrzemekK: i don't know masakari in detail ..i've just checked kolla-ansible code | 11:37 |
kevko | PrzemekK: so, i was wondering if it's not bug in kolla-ansible to just set messagingv2 | 11:37 |
PrzemekK | we use ceilometer also | 11:38 |
kevko | PrzemekK: ah, ok | 11:38 |
kevko | that i am afraid i can't help you as i currently don't have similar installation under my hands | 11:39 |
PrzemekK | its ok | 11:39 |
PrzemekK | TotalReadyUnackedIn memory readyDead-lettered Messages 4,9104,910000 Message body bytes 9.0 MiB9.0 MiB0 B0 B0 B Process memory 1.7 MiB | 11:39 |
kevko | PrzemekK: but from kolla-ansible code - you have probably problem with ceilometer no ? | 11:41 |
kevko | PrzemekK: because normally you don't need notifications from masakari no ? | 11:41 |
opendevreview | Mark Goddard proposed openstack/kolla-ansible stable/2024.1: Fix prometheus.yml templating https://review.opendev.org/c/openstack/kolla-ansible/+/926618 | 11:43 |
opendevreview | Mark Goddard proposed openstack/kolla-ansible stable/2023.2: Fix prometheus.yml templating https://review.opendev.org/c/openstack/kolla-ansible/+/926624 | 11:43 |
opendevreview | Alex Welsh proposed openstack/kolla-ansible stable/2024.1: Fix prometheus.yml templating https://review.opendev.org/c/openstack/kolla-ansible/+/926618 | 11:45 |
PrzemekK | We didnt seen issues with ceilometer. We can read performance from gnocchi - we dont need notifications from masakari i think . In ceilometer logs i can see only error like 2024-08-20 01:36:39.972 748 ERROR ceilometer.publisher.gnocchi [-] Error creating resource dac2efec-672f-4b18-a371-cda964ab686a: {'cause': 'Invalid input', 'reason': "required key not provided @ data['name']", 'detail': ['name']} (HTTP 400): gnocchicli | 11:49 |
kevko | PrzemekK: so turn off the notifications from masakari if nobody reads it ... | 11:53 |
PrzemekK | ok thx | 11:55 |
kevko | PrzemekK: In kolla-ansible, if Ceilometer is not enabled, notifications are turned off. This obviously means that for the masakari itself, you don't need notifications at all. However, if it is enabled in kolla and there is a condition that it's enabled when enable_ceilometer is set... someone might have assumed that Ceilometer can/will consume | 11:56 |
kevko | these notifications... but evidently not. I didn't find anything in the Ceilometer documentation suggesting that it processes these messages, so it could easily be a bug in kolla-ansible. | 11:56 |
kevko | in template | 11:56 |
opendevreview | Mark Goddard proposed openstack/kayobe master: CI: Return to openstack-python3-jobs, fix pep8 failures https://review.opendev.org/c/openstack/kayobe/+/926619 | 12:01 |
opendevreview | Michal Nasiadka proposed openstack/kolla-ansible master: [WIP] Enable ML2/OVN and distributed FIP by default. https://review.opendev.org/c/openstack/kolla-ansible/+/904959 | 12:11 |
opendevreview | Merged openstack/kolla stable/2023.2: Trivial fix letsencrypt base image https://review.opendev.org/c/openstack/kolla/+/925081 | 12:22 |
opendevreview | Michal Arbet proposed openstack/kolla-ansible master: Add REQUESTS_CA_BUNDLE to kolla-toolbox container https://review.opendev.org/c/openstack/kolla-ansible/+/923105 | 12:26 |
chembervint | mnasiadka: I've commented at my patch https://review.opendev.org/c/openstack/kolla-ansible/+/920377 - I'm not really sure, that flat\vlan network will increase reliability of the amphoras healthchecking | 12:31 |
opendevreview | Michal Nasiadka proposed openstack/kolla-ansible master: horizon: Fetch venv python path instead of static definition https://review.opendev.org/c/openstack/kolla-ansible/+/926632 | 12:37 |
opendevreview | Michal Arbet proposed openstack/kolla-ansible master: Add REQUESTS_CA_BUNDLE to kolla-toolbox container https://review.opendev.org/c/openstack/kolla-ansible/+/923105 | 12:37 |
opendevreview | Sven Kieske proposed openstack/kolla-ansible master: reorder requirements.txt alphabetically https://review.opendev.org/c/openstack/kolla-ansible/+/922617 | 12:42 |
opendevreview | Michal Nasiadka proposed openstack/ansible-collection-kolla master: docker/podman_sdk: Install sdk from deb when python externally managed https://review.opendev.org/c/openstack/ansible-collection-kolla/+/925994 | 12:50 |
opendevreview | Merged openstack/kolla stable/2024.1: Trivial fix letsencrypt base image https://review.opendev.org/c/openstack/kolla/+/925080 | 12:50 |
opendevreview | Merged openstack/kolla master: Fix build of prometheus-ovn-exporter https://review.opendev.org/c/openstack/kolla/+/925886 | 12:50 |
opendevreview | Will Szumski proposed openstack/kayobe master: Aligns Kayobe Ansible version with Kolla Ansible https://review.opendev.org/c/openstack/kayobe/+/926639 | 13:10 |
opendevreview | Simon Dodsley proposed openstack/kolla-ansible master: Add NVMe-TCP Cinder driver support for Pure Storage https://review.opendev.org/c/openstack/kolla-ansible/+/919977 | 13:10 |
SvenKieske | anyone has any idea why I get "ModuleNotFoundError: No module named 'packaging'" on debian bookworm with python 3.11.2 in CI (I added the package to requirements.txt)? https://zuul.opendev.org/t/openstack/build/6b3345b8040c4139ac8e67a56274332c/log/primary/logs/ansible/deploy | 13:14 |
SvenKieske | do we use system packages for bookworm? I tend to forget because it's always different, depending on distribution. | 13:15 |
SvenKieske | but still stuff from requirements.txt should get installed, no? | 13:16 |
SvenKieske | can't replicate this in a local bookworm container at least | 13:20 |
opendevreview | Will Szumski proposed openstack/kayobe master: Bump Ansible version installed in Kolla Ansible virtalenv https://review.opendev.org/c/openstack/kayobe/+/923556 | 13:24 |
opendevreview | Will Szumski proposed openstack/kayobe master: Aligns Kayobe Ansible version with Kolla Ansible https://review.opendev.org/c/openstack/kayobe/+/926639 | 13:24 |
opendevreview | Will Szumski proposed openstack/kayobe master: Aligns Kayobe Ansible version with Kolla Ansible https://review.opendev.org/c/openstack/kayobe/+/926639 | 13:31 |
opendevreview | Michal Nasiadka proposed openstack/kolla-ansible master: service-python-path: Introduce a role to get python_path of an image https://review.opendev.org/c/openstack/kolla-ansible/+/926632 | 13:36 |
opendevreview | Merged openstack/kolla-ansible master: Add support for docker_image_name_prefix https://review.opendev.org/c/openstack/kolla-ansible/+/924450 | 13:37 |
opendevreview | Merged openstack/kolla-ansible master: Add fluentd_enable_watch_timer https://review.opendev.org/c/openstack/kolla-ansible/+/785309 | 13:41 |
opendevreview | Mark Goddard proposed openstack/kolla-ansible master: Remove "distutils" library https://review.opendev.org/c/openstack/kolla-ansible/+/857003 | 13:47 |
kevko | SvenKieske: change ? | 13:52 |
PrzemekK | Is cpu_allocation_ratio works without any modification in /etc/kolla/config/nova/myhost/nova.conf or there need to add to scheduler filter AggregateCoreFilter https://github.com/openstack/kolla-ansible/blob/master/ansible/roles/nova/templates/nova.conf.j2 | 13:52 |
SvenKieske | kevko: nvm; I see my change is a duplicate. beside that, you can always scroll just to the top of the deploy log and there is the change id with a hyperlink which directs you to the appropriate gerrit link | 13:54 |
kevko | SvenKieske: yeah, i have this type of patch in gerrit for 2 months :D https://review.opendev.org/c/openstack/kolla-ansible/+/922621 | 13:55 |
SvenKieske | PrzemekK: afaik AggregateCoreFilter is not a think since ocata release, but I'm just a time traveller here, because I was not around in openstack land around this time: https://docs.openstack.org/releasenotes/nova/ocata.html#relnotes-15-1-1-origin-stable-ocata | 13:57 |
SvenKieske | thing* | 13:57 |
SvenKieske | it's also not mentioned in latest docs here: https://docs.openstack.org/nova/latest/admin/scheduling.html | 13:59 |
PrzemekK | Ok. Latest nova.conf example show #enabled_filters = ComputeFilter,ComputeCapabilitiesFilter,ImagePropertiesFilter,ServerGroupAntiAffinityFilter,ServerGroupAffinityFilter | 14:08 |
opendevreview | Matt Crees proposed openstack/kolla-ansible master: Add a precheck to catch RMQ SLURP upgrades https://review.opendev.org/c/openstack/kolla-ansible/+/926146 | 14:09 |
opendevreview | Michal Arbet proposed openstack/kolla-ansible master: Add ansible-core as a dependency https://review.opendev.org/c/openstack/kolla-ansible/+/922369 | 14:13 |
SvenKieske | I'm not sure I would take an example as a reference, it's an example, not a reference. Then again openstack projects are very inconsistent with this. Some projects don't have a reference, but only an example config. :( | 14:31 |
kevko | PrzemekK: what about to check the filters for your version ? :) https://github.com/openstack/nova/tree/master/nova/scheduler/filters | 14:34 |
PrzemekK | just was trying to help someone on reddit :D https://www.reddit.com/r/openstack/comments/1ewrg4w/kolla_config_does_not_affected/ | 14:37 |
opendevreview | Will Szumski proposed openstack/kayobe master: Bump Ansible version installed in Kolla Ansible virtalenv https://review.opendev.org/c/openstack/kayobe/+/923556 | 14:41 |
opendevreview | Will Szumski proposed openstack/kayobe master: Bump Ansible version installed in Kolla Ansible virtalenv https://review.opendev.org/c/openstack/kayobe/+/923556 | 14:42 |
opendevreview | Will Szumski proposed openstack/kayobe master: Bump Ansible version installed in Kolla Ansible virtalenv https://review.opendev.org/c/openstack/kayobe/+/923556 | 14:43 |
opendevreview | Will Szumski proposed openstack/kayobe master: Bump Ansible version installed in Kolla Ansible virtalenv https://review.opendev.org/c/openstack/kayobe/+/923556 | 14:44 |
opendevreview | Will Szumski proposed openstack/kayobe master: Aligns Kayobe Ansible version with Kolla Ansible https://review.opendev.org/c/openstack/kayobe/+/926639 | 14:44 |
kevko | PrzemekK: NUMATopologyFilter | 14:50 |
kevko | or maybe it is defined in placement resource ? | 14:53 |
SvenKieske | kevko: btw did you read fungi's reply on openstack-discuss regarding consul? I'm not sure if we are producing any artifacts during CI runs which include consul et al? maybe you can answer those questions? | 14:53 |
kevko | SvenKieske: let me check | 14:54 |
fungi | yes, apologies but it's unclear to me the extent to which consul integration in kolla means that you might, e.g., embed consul itself in images the kolla team distributes rather than expecting the user to install consul from some other source (hashicorp official container images or whatever) | 14:55 |
kevko | SvenKieske: can you point me to email ? i can't see in my inbox | 14:55 |
fungi | i'll get you a link | 14:55 |
fungi | kevko: it's the last message in this thread: https://lists.openstack.org/archives/list/openstack-discuss@lists.openstack.org/thread/4JHNYSMXFVPW5S3MGI4OCNTELHT7YS2O/ | 14:56 |
opendevreview | Mark Goddard proposed openstack/kolla-ansible master: Remove "distutils" library https://review.opendev.org/c/openstack/kolla-ansible/+/857003 | 14:56 |
fungi | kevko: a related earlier question on the legal-discuss thread which went unanswered was whether kolla would be installing consul in ci/cd jobs in order to test its integration | 14:57 |
fungi | these are the sorts of subtopics likely to be raised in the tc meeting today when they discuss the situation | 14:58 |
kevko | fungi: regarding CI - i didn't try to resolve the question IF kolla team CAN install or whatever in CI ..i just asked for add an option to build image (kolla) and orchestrate those images (kolla-ansible) ...as this is of course possible according to the BSL license | 15:01 |
kevko | fungi: from BSL >> "But, similar to open source licenses, BSL-licensed source code is publicly available, and anyone is free to use, modify, and/or copy it for non-production purposes." | 15:02 |
fungi | kevko: yes, the tc won't be solely concerned with what is legal (though that is a primary concern), but also what makes sense for openstack as a whole | 15:03 |
fungi | e.g. whether binary artifacts published by project teams (which it already puts some limits on) should be allowed to include software with non-open-source/non-osi-approved licenses | 15:04 |
kevko | fungi: from my perspective of view it makes sense to support BSL stuff ... from my point of view openstack is not only opensource for couple of folks installing cloud in garage ...but also for big companies which for example can pay for piece of sotware ..but they know openstack support it | 15:04 |
fungi | your opinion definitely matters, but i'm not the one you need to present it to | 15:05 |
kevko | and they don't need also pay for it ... because BSL means ... for example for vault -- OKAY - download the vault ...but don't sell as your own product and get a money from that ... | 15:05 |
fungi | yes, i get that you're advocating for "free as in beer" being good enough, and that open source ideals aren't all that important. it's just worth keeping in mind that not all of the project leadership will agree with you | 15:06 |
kevko | fungi: BSL also means that if you have secrets stored in vault in production ....BUT it's not sold as a product ..you also can use it ...but you can't provide an API to end users ...it can be used as backend dependency | 15:06 |
fungi | i fully understand the license, there's no need to re-explain it | 15:07 |
fungi | i'm not even a tc member, i just wanted to make sure you know that the tc members are planning to discuss it in their meeting today | 15:08 |
kevko | fungi: So you agree that there are no obstacles to offering BSL options for Kolla ... and by extension for OpenStack as a whole. | 15:10 |
kevko | fungi: i didn't know this ..but I am really curious about it ... can u please share the channel name and time ? i will probably attend | 15:10 |
fungi | i agree that the foundation's lawyers have said that the license doesn't present new legal risks, but that's not the only factor the tc members will be concerned with | 15:11 |
fungi | "it's not illegal" is just one part of this | 15:12 |
kevko | fungi: By the way, the question about CI didn't even occur to me because I assumed this was already resolved and in order ... since Kolla already does this for Hashicorp Vault here https://github.com/openstack/kolla-ansible/blob/master/tests/test-hashicorp-vault-passwords.sh | 15:12 |
fungi | kevko: did you read the message i linked? it includes the url for the tc meeting agenda which has the relevant meeting details | 15:13 |
fungi | kevko: also, that kolla hasn't done anything about continuing to install vault after its license ceased to be open source is an important point to bring up, yes. i don't know whether that will be a convincing argument in favor of adding a new test dependency which is known to not be open source | 15:15 |
kevko | fungi: yeah - totally understand. Yes, but we are an open-source community, and we build it together. I considered it useful and necessary to raise this question and address it, especially because it seems almost like a trend and could happen elsewhere. Moreover, I never understood why Kolla can't provide the installation of Hashicorp Vault when | 15:16 |
kevko | the OpenStack community as a whole supports Hashicorp Vault. | 15:16 |
fungi | yes, supported it while it was still open source | 15:16 |
fungi | hasn't removed support for it yet, nor stopped installing it in tests | 15:16 |
fungi | again, please hear me, you don't need to litigate this case to me, i just wanted to let you know the tc is planning to discuss whether they're comfortable with the plan you've proposed | 15:17 |
kevko | fungi: thank you very much | 15:17 |
kevko | fungi: i mean - it's nice to present arguments in a discussion to someone, so yes, I understand it's not valid for you, but as a discussion exercise—fine. | 15:19 |
fungi | i'm perfectly happy to be a sounding board, sure, and i'm at least familiar with what some of the tc members' concerns are likely to be | 15:19 |
SvenKieske | kevko: channel should be #openstack-tc time is 18:00 UTC, if the wiki is correct | 15:20 |
fungi | the main points are 1. foundation legal counsel has said that openstack software interfacing with bsl software doesn't present license incompatibilities, 2. openstack added support for hashicorp vault when it was still open source and hasn't been told to stop installing it in tests or distributing copies of it after hashicorp switched it to bsl, 3. the kolla team is asking to add a new | 15:23 |
fungi | test dependency on additional non-open-source software under the same license and distribute copies of that | 15:23 |
fungi | likely policy questions for the tc members to tackle are: 1. should the current guidance on distribution of binary artifacts be extended to require that they contain only open source software? 2. is it okay for an openstack project to add a new test dependency which is not open source? 3. if a test dependency switches to a non-open-source license, what actions is openstack expected to | 15:27 |
fungi | take? (pin to the last open source version, switch to an open source fork, remove upstream testing of the integration...) | 15:27 |
fungi | frickler: since you're chairing the tc meeting today, possibly worth noting this ^ is a quick summary of the "Implications of testing/shipping non-OSI licensed software within OpenStack" topic on the agenda | 15:30 |
frickler | thx for the pointer, seems goutham has timed his absence well :-S | 15:34 |
kevko | fungi frickler: But the BSL license changes in older versions of the software and becomes completely open source... In other words, OpenStack users can use a fully open-source version if they use an older version of the BSL software... because at that point, it will already be completely open source. | 15:35 |
fungi | yes, assuming users are comfortable not getting security fixes for it (or waiting two years for each security fix anyway) | 15:37 |
kevko | fungi: so still, in overall - i think this is question mainly for end users ..not for openstack as a community ... | 15:38 |
opendevreview | Will Szumski proposed openstack/kayobe master: Support running without root privileges https://review.opendev.org/c/openstack/kayobe/+/923430 | 15:40 |
fungi | well, whether openstack installs non-open-source software in its upstream testing and whether openstack distributes copies of non-open-source software on the internet are certainly questions for openstack as a community. whether to install such software in production is a question for the end user, yes | 15:40 |
opendevreview | Mark Goddard proposed openstack/kayobe master: ironic: revert to ironic's default drivers & interfaces https://review.opendev.org/c/openstack/kayobe/+/836999 | 15:40 |
kevko | fungi: yes, and lawyer said it's not problem :D ...but let's wait for TC meeting | 15:41 |
opendevreview | Mark Goddard proposed openstack/kolla master: Pin OpenSearch Dashboards to 2.15 https://review.opendev.org/c/openstack/kolla/+/926528 | 15:44 |
PrzemekK | @kevko i chcecked with just add to file vi /etc/kolla-pilot/config/nova.conf cpu_allocation_ratio and it works on 2024.1 without any changes in filters | 15:52 |
SvenKieske | PrzemekK: yes that should work, used that at least in the past myself. and thanks for helping users on reddit! :) | 15:53 |
opendevreview | Will Szumski proposed openstack/kayobe master: Support running without root privileges https://review.opendev.org/c/openstack/kayobe/+/923430 | 15:55 |
opendevreview | Will Szumski proposed openstack/kayobe master: Support running without root privileges https://review.opendev.org/c/openstack/kayobe/+/923430 | 15:55 |
PrzemekK | yeah but for isolate Windows VM i think there is missing filter AggregateImagePropertiesIsolation | 15:58 |
opendevreview | Will Szumski proposed openstack/kayobe master: Bump Ansible version installed in Kolla Ansible virtalenv https://review.opendev.org/c/openstack/kayobe/+/923556 | 17:05 |
opendevreview | Will Szumski proposed openstack/kayobe master: Aligns Kayobe Ansible version with Kolla Ansible https://review.opendev.org/c/openstack/kayobe/+/926639 | 17:09 |
opendevreview | Will Szumski proposed openstack/kayobe master: Aligns Kayobe Ansible version with Kolla Ansible https://review.opendev.org/c/openstack/kayobe/+/926639 | 17:15 |
opendevreview | Will Szumski proposed openstack/kayobe master: DNM: Test simplified deps https://review.opendev.org/c/openstack/kayobe/+/926664 | 17:45 |
opendevreview | Will Szumski proposed openstack/kayobe master: Bump Ansible version installed in Kolla Ansible virtualenv https://review.opendev.org/c/openstack/kayobe/+/923556 | 17:46 |
opendevreview | Will Szumski proposed openstack/kayobe master: Aligns Kayobe Ansible version with Kolla Ansible https://review.opendev.org/c/openstack/kayobe/+/926639 | 17:46 |
opendevreview | Will Szumski proposed openstack/kayobe master: DNM: Test simplified deps https://review.opendev.org/c/openstack/kayobe/+/926664 | 17:46 |
Generated by irclog2html.py 2.17.3 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!