opendevreview | Michal Arbet proposed openstack/kolla-ansible master: Bash replacement https://review.opendev.org/c/openstack/kolla-ansible/+/920368 | 02:12 |
---|---|---|
opendevreview | Michal Arbet proposed openstack/kolla-ansible master: Bash replacement https://review.opendev.org/c/openstack/kolla-ansible/+/920368 | 02:45 |
frickler | fyi the failures are due to a bug in zuul. there is a fix for that already and it will hopefully be rolled out by tomorrow | 06:44 |
mnasiadka | morning | 07:04 |
mnasiadka | frickler: the codespell ones here? https://zuul.opendev.org/t/openstack/build/707d770fdbd6484db7d41c6d1f9a675f | 07:05 |
frickler | mnasiadka: no, the failures from https://review.opendev.org/c/openstack/kolla-ansible/+/920294 that were mentioned yesterday evening | 07:33 |
mnasiadka | ok, wonder what happened that we get codespell errors in master now :) | 07:34 |
mnasiadka | ah, codespell 2.3.0 released yesterday | 07:34 |
mnasiadka | let me pin | 07:34 |
frickler | codespell got less woke it seems :-/ | 07:36 |
frickler | though the error it complains about is a valid one I'd say | 07:36 |
opendevreview | Michal Nasiadka proposed openstack/kolla-ansible master: CI: Pin codespell to <2.3.0 https://review.opendev.org/c/openstack/kolla-ansible/+/920373 | 07:37 |
mnasiadka | yeah, but I'd prefer to pin now, and then update in a later change | 07:37 |
frickler | if there were like 50 changes to make, I could agree, but just for a single typo? | 07:39 |
mnasiadka | well, the same problem will be in stable branches | 07:40 |
mnasiadka | and updating renos is probably not something we should do again | 07:40 |
frickler | ok, we should pin codespell for stable branches probably like other linters, agreed | 07:40 |
sylvr | Hello ! kayobe can gather information using PXEboot (with custom RAMdisk) AND by connecting to the node's oob controller is that right ? | 07:46 |
sylvr | is "inspection" for using their oob controller and "introspection" for using the IPA RAMdisk ? | 07:47 |
opendevreview | Michal Nasiadka proposed openstack/kolla-ansible master: CI: Pin requests to <2.32 for docker sdk https://review.opendev.org/c/openstack/kolla-ansible/+/920169 | 07:48 |
mnasiadka | sylvr: by default kayobe (actually bifrost deployed by kolla-ansible by kayobe) uses IPA ramdisk for inspection - but if your hardware is redfish or something similar and there's a inspection driver that allows out of band inspection - then you can use that, but that will require customising bifrost.yml | 07:52 |
sylvr | mnasiadka: I put "idrac" in "bifrost enabled hardware types" but running `kayobe overcloud inventory discover` does nothing (no inventory) on the bifrost container I can see my nodes, but their oob isn't managed by the seed | 07:56 |
mnasiadka | you mean you can see the nodes in "openstack baremetal node list" in the bifrost container? | 07:57 |
sylvr | I think I might have misconfigured my networks... but I'm having trouble figuring out why | 07:57 |
sylvr | mnasiadka: yes! | 07:57 |
mnasiadka | then overcloud inventory discover should add then to inventory | 07:59 |
sylvr | they're all in the "enroll" provisioning state | 08:00 |
sylvr | they're unnamed, no UUID and no power state | 08:00 |
sylvr | running "baremetal node show" I don't see any mention of idrac, but only ipmi | 08:01 |
opendevreview | Matt Crees proposed openstack/kayobe stable/2023.1: Fix permissions on dev/rabbitmq-migrate-queues.sh https://review.opendev.org/c/openstack/kayobe/+/920375 | 08:02 |
mnasiadka | sylvr: they should be in available, no enroll | 08:02 |
mnasiadka | so you booted them via network? | 08:03 |
mnasiadka | or added manually? | 08:03 |
sylvr | well, I've turned the machines on and they should try PXEboot on the mgnt network (which the seed has access to) | 08:04 |
sylvr | I can see them using baremetal node list, and I get information about their PXE interface | 08:05 |
mnasiadka | So, usually we use name from port description to name the nodes (from lldp data) see https://github.com/openstack/kayobe/blob/27ece2622b934eb6f4d9712d1879ac9bab8d236e/ansible/inventory/group_vars/all/inspector#L214 (inspector_rule_var_lldp_switch_port_interface variable) | 08:06 |
mnasiadka | to change the default enroll node driver - you can set this variable in kolla/config/bifrost/bifrost.yml - https://github.com/openstack/bifrost/blob/e0a5a7216d74821aee8cc32a8767c790ed731483/playbooks/roles/bifrost-ironic-install/templates/ironic-inspector.conf.j2#L95 | 08:07 |
mnasiadka | frickler: https://review.opendev.org/c/openstack/kolla-ansible/+/920373 - passed CI | 08:07 |
sylvr | mnasiadka: okay so in kayobe-config/.../inspector.yml ? | 08:09 |
sylvr | "inspector_discovery_enroll_node_driver" | 08:10 |
mnasiadka | kayobe-config/etc/kayobe/kolla/config/bifrost/bifrost.yml if I remember correct | 08:11 |
mnasiadka | kayobe uses kolla-ansible to deploy bifrost, that's why kolla/config | 08:11 |
mnasiadka | https://github.com/openstack/kayobe/blob/27ece2622b934eb6f4d9712d1879ac9bab8d236e/doc/source/configuration/reference/bifrost.rst#L32 | 08:12 |
frickler | mnasiadka: oh, that went fast, +2, do you want to fast-approve, too? | 08:14 |
opendevreview | Victor Chembaev proposed openstack/kolla-ansible master: Fix octavia-interface timeout https://review.opendev.org/c/openstack/kolla-ansible/+/920377 | 08:15 |
mnasiadka | frickler: added +w, that's only running pep8 and linters mainly, so gate should be fast as well | 08:16 |
sylvr | mnasiadka: okay, so now my file kayobe-config/etc/kayobe/kolla/config/bifrost/bifrost.yml contain the two lines following : enable_inspector_discovery: "true" inspector_default_node_driver: "idrac" | 08:16 |
sylvr | and running kayobe seed service upgrade/deploy should do the trick ? | 08:17 |
mnasiadka | deploy, not upgrade | 08:17 |
mnasiadka | upgrade is only between releases | 08:17 |
sylvr | yes okay | 08:17 |
sylvr | I'll try this out right now, I'll let you know ! thank you very much | 08:17 |
mnasiadka | frickler: It's a bit irritating that ansible-galaxy connects to internet for installing a git based collection - I think I'll try to get rid of galaxy.ansible.com dependency in CI | 08:19 |
mnasiadka | ah, for ansible-core we're not going to skip ansible-galaxy | 08:35 |
mnasiadka | frickler: do you think we could pass that traffic through some caching proxy in opendev? | 08:35 |
frickler | mnasiadka: iirc this didn't work with the weird setup galaxy is using | 08:42 |
opendevreview | Merged openstack/kolla-ansible master: CI: Pin codespell to <2.3.0 https://review.opendev.org/c/openstack/kolla-ansible/+/920373 | 08:44 |
sylvr | mnasiadka: still no inventory... baremetal node show still doesn't contain any mention of idrac... I might have misconfigured a lot of things | 08:47 |
opendevreview | Michal Nasiadka proposed openstack/kolla-ansible master: cli: add ansible-galaxy retries https://review.opendev.org/c/openstack/kolla-ansible/+/920378 | 08:49 |
mnasiadka | frickler: so let's try retries | 08:49 |
frickler | mnasiadka: ack, that would match what we do for toolbox build. your code looks incomplete, though | 08:53 |
SvenKieske | here is a python implementation which seems to be able to build a ansible-galaxy cache: https://github.com/jctanner/galaxy-mirror | 08:55 |
frickler | SvenKieske: I saw that, but it seems incomplete in terms of implementing a TTL for cached items. also not sure how we could integrate that with our AFS setup | 08:57 |
SvenKieske | implement a TTL via cron? :D | 08:57 |
SvenKieske | I don't know anything about the AFS though | 08:58 |
SvenKieske | there was also an attempt in triple-o, but it was abandoned together with it, it seems: https://review.opendev.org/c/openstack/tripleo-ansible/+/866031 | 08:59 |
frickler | the actual mirror servers only have access to a read-only copy of the mirror fs. so a cache-on-demand setup would be not as effective as a real mirror which would create some kind of data that could then be replicated | 09:00 |
SvenKieske | okay, was enhanced security the thought about that design decision (read only copy) or what else played into it? | 09:03 |
frickler | I think the main reason is that coordinating multiple writers to avoid conflicting updates would be difficult, so there is a single mirror-update node that does the update jobs | 09:12 |
SvenKieske | seems reasonable | 09:21 |
SvenKieske | we have quite some jobs with POST_FAILURE currently in kolla: https://zuul.opendev.org/t/openstack/builds?project=openstack/kolla | 09:22 |
sylvr | modified kayobe-config/etc/kayobe/inspector.yml to reflect the change in the other file (kolla/bifrost.yml) still, baremetal node show doesn't list idrac as driver... | 09:22 |
SvenKieske | just wanted to figure out if I can recheck https://review.opendev.org/c/openstack/kolla/+/918974 but I _think_ we may need https://review.opendev.org/c/openstack/kolla/+/920279 first? looking into the builds now.. | 09:22 |
SvenKieske | or is that some quay.io issue? mhm | 09:23 |
frickler | that looks like the publishing jobs are also affected by the docker-py issue | 09:23 |
SvenKieske | ah no, it's not INFO:kolla.common.utils:Exception caught: Error while fetching server API version: Not supported URL scheme http+docker | 09:23 |
SvenKieske | so it seems we need https://review.opendev.org/c/openstack/kolla/+/920279 (or some variant of it) | 09:25 |
opendevreview | Will Szumski proposed openstack/kolla master: Add packaging as an explicit dependency https://review.opendev.org/c/openstack/kolla/+/920381 | 09:30 |
opendevreview | Will Szumski proposed openstack/kayobe master: Bump up Ansible supported versions to 8.x/9.x https://review.opendev.org/c/openstack/kayobe/+/910513 | 09:32 |
jovial[m] | slyvr: You are going to need to enable the idrac hardware type and all the other interfaces like in this: https://github.com/openstack/bifrost/commit/d8751563e74750ce2c7bd0e7b076604a7641d74a | 09:36 |
jovial[m] | I wouldn't recommend using the idrac driver | 09:37 |
jovial[m] | and then rediscover the node | 09:39 |
sylvr | jovial: I did, I only want to use the idrac driver, but it doesn't seems to work... my oob network contain all my idracs configured to do DHCP | 09:56 |
sylvr | My plan was to have a switch with the oob network with all my machines idrac with DHCP to be controlled by the seed/bifrost for inspection/introspection | 09:57 |
sylvr | I'm guessing provisioning can be done with PXEboot or through the idrac, and it is a requirement that my nodes can be baremetal, so PXEboot might not be an option in this case | 09:58 |
opendevreview | Michal Arbet proposed openstack/kolla-ansible master: Bash replacement https://review.opendev.org/c/openstack/kolla-ansible/+/920368 | 11:45 |
opendevreview | Michal Arbet proposed openstack/kolla-ansible master: Bash replacement https://review.opendev.org/c/openstack/kolla-ansible/+/920368 | 11:47 |
opendevreview | Michal Arbet proposed openstack/kolla-ansible master: Bash replacement https://review.opendev.org/c/openstack/kolla-ansible/+/920368 | 11:49 |
opendevreview | Merged openstack/kayobe stable/2023.1: Fix permissions on dev/rabbitmq-migrate-queues.sh https://review.opendev.org/c/openstack/kayobe/+/920375 | 12:03 |
opendevreview | Michal Arbet proposed openstack/kolla-ansible master: Bash replacement https://review.opendev.org/c/openstack/kolla-ansible/+/920368 | 12:20 |
opendevreview | Michal Nasiadka proposed openstack/kolla-ansible master: cli: add ansible-galaxy retries https://review.opendev.org/c/openstack/kolla-ansible/+/920378 | 12:28 |
opendevreview | Michal Nasiadka proposed openstack/kolla-ansible master: cli: add ansible-galaxy retries https://review.opendev.org/c/openstack/kolla-ansible/+/920378 | 12:29 |
opendevreview | Michal Arbet proposed openstack/kolla-ansible master: Bash replacement https://review.opendev.org/c/openstack/kolla-ansible/+/920368 | 12:45 |
opendevreview | Michal Nasiadka proposed openstack/kolla-ansible stable/2024.1: CI: Pin codespell to <2.3.0 https://review.opendev.org/c/openstack/kolla-ansible/+/920352 | 13:02 |
opendevreview | Michal Nasiadka proposed openstack/kolla-ansible stable/2023.2: CI: Pin codespell to <2.3.0 https://review.opendev.org/c/openstack/kolla-ansible/+/920354 | 14:26 |
opendevreview | Michal Nasiadka proposed openstack/kolla-ansible stable/2023.1: CI: Pin codespell to <2.3.0 https://review.opendev.org/c/openstack/kolla-ansible/+/920355 | 14:26 |
opendevreview | Merged openstack/kolla-ansible master: CI: Pin requests to <2.32 for docker sdk https://review.opendev.org/c/openstack/kolla-ansible/+/920169 | 14:34 |
opendevreview | Mariusz Karpiarz proposed openstack/kolla-ansible master: Allow to run post-deploy without privilege escalation https://review.opendev.org/c/openstack/kolla-ansible/+/920294 | 14:43 |
opendevreview | Michal Nasiadka proposed openstack/kolla-ansible master: cli: add ansible-galaxy retries https://review.opendev.org/c/openstack/kolla-ansible/+/920378 | 15:32 |
opendevreview | Michal Arbet proposed openstack/kolla-ansible master: Bash replacement https://review.opendev.org/c/openstack/kolla-ansible/+/920368 | 16:24 |
opendevreview | Doug Szumski proposed openstack/kolla master: WIP: Restore Zookeeper https://review.opendev.org/c/openstack/kolla/+/920405 | 16:40 |
Generated by irclog2html.py 2.17.3 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!