| opendevreview | Merged openstack/openstack-ansible-os_neutron stable/victoria: Remove legacy centos-8 jobs https://review.opendev.org/c/openstack/openstack-ansible-os_neutron/+/827483 | 00:40 |
|---|---|---|
| *** Guest2 is now known as prometheanfire | 01:28 | |
| *** akahat|PTO is now known as akahat | 05:22 | |
| evrardjp[m] | damiandabrowski: balance source was for federation purposes | 07:18 |
| evrardjp[m] | Not sure it still applies as fédération is now different | 07:18 |
| evrardjp[m] | mgariepy: ask noonedeadpunk he has seen that ;) | 07:23 |
| evrardjp[m] | At the end I am wondering why we are using a list of memcached and not simply rely on the one local on the controller. It’s simple operations wise | 07:24 |
| jrosser | you have to configure it for all the services though? | 07:36 |
| jrosser | like https://github.com/openstack/openstack-ansible-os_glance/blob/master/defaults/main.yml#L88 | 07:36 |
| noonedeadpunk | balance source must be set for glance for sure at least | 07:38 |
| noonedeadpunk | mgariepy: oh yes, I have seen, and used haproxy balancing to cover that. | 07:39 |
| noonedeadpunk | evrardjp[m]: one point against just local memcache was - what if that process dies or get killed? then whole controller is kind of useless? | 07:39 |
| noonedeadpunk | from other side I guess damiandabrowski[m] was testing that if the only memcached goes down, so no real pool used, then everything just continues working? I recall things correctly? | 07:40 |
| noonedeadpunk | but we can likely add a variable to be able to easily control how memcached will be configured (ie reflect memcached_servers generation) - should it be only "local" server or pool as an example | 07:42 |
| jrosser | would be worth checking with andrewbonney as we did some quite deep digging into how the memcached client worked with dead backends | 07:42 |
| jrosser | it was odd / complicated iirc | 07:43 |
| noonedeadpunk | oh yes, it is | 07:43 |
| noonedeadpunk | Damian was digging there as well, so yeah. | 07:43 |
| jrosser | OIDC federation apache uses memcached too | 07:44 |
| jrosser | to store session state | 07:44 |
| noonedeadpunk | eventually mcrouter is only really working solution imo, but it's facebook stuff so it's not smth I'd rely on in prod | 07:44 |
| evrardjp_ | jrosser: not all python libs are using the same memcache client. | 07:56 |
| jrosser | i think there was recently an effort to clean up keystone about that | 07:56 |
| evrardjp_ | if we rely just on protocol by default, the sharding is handled on the client easily. | 07:56 |
| jrosser | as it seemed to have its own | 07:56 |
| evrardjp_ | For me it's all about risk management on partition | 07:58 |
| evrardjp_ | But I will leave the experts at play :) | 07:59 |
| evrardjp_ | habve a good day everyone! | 07:59 |
| *** arxcruz is now known as arxcruz|ruck | 09:13 | |
| damiandabrowski[m] | hey everyone! thanks JP for Your input | 09:24 |
| damiandabrowski[m] | I'm off today but in fact, I've spent some time testing oslo_cache.memcache_pool lately | 09:25 |
| damiandabrowski[m] | here's the outcome: https://paste.openstack.org/show/bm9hnzbwiaG9wWXWQRiT/ | 09:25 |
| damiandabrowski[m] | TLDR: when all memcached instances are down, it marks them as up which makes a mess | 09:26 |
| damiandabrowski[m] | proposed a change to fix it but it didn't get any interest: https://review.opendev.org/c/openstack/oslo.cache/+/824716 | 09:26 |
| jrosser | damiandabrowski[m]: you might get more traction by making a bug against oslo.cache | 09:56 |
| damiandabrowski[m] | yeah, You may be right, thanks | 09:57 |
| jrosser | or even post to the ML | 09:58 |
| evrardjp[m] | damiandabrowski: nice find. You’re a bit proving my point however: the python libs do matter , so does our config. In terms of operations, it’s up to the operator. I think it makes sense to have 1 controller as a ‘block’ from operations perspective, but that’s only my personal opinion | 10:14 |
| opendevreview | Merged openstack/openstack-ansible stable/wallaby: Rename RBD cinder backend https://review.opendev.org/c/openstack/openstack-ansible/+/828664 | 10:16 |
| opendevreview | Merged openstack/openstack-ansible stable/victoria: Rename RBD cinder backend https://review.opendev.org/c/openstack/openstack-ansible/+/828665 | 10:16 |
| noonedeadpunk | andrewbonney: mgariepy if/when have time would be great to get review on https://review.opendev.org/c/openstack/openstack-ansible-rabbitmq_server/+/826338 | 10:23 |
| *** dviroel|out is now known as dviroel | 11:15 | |
| admin1 | hi guys .. anyone using ceph-ansible managed ceph and osa and integrating osa object storage to ceph s3, can you please share your configs .. mine is at https://gist.githubusercontent.com/a1git/78761a5346aef2e5db732b0c838b72ff/raw/72fbb5a60ee6a10d3fa9bb0a4bdf2f0ce115725e/gistfile1.txt and i am having an issue of auth failure and unabel to | 11:31 |
| admin1 | understand what is happening and where i went wrong | 11:31 |
| jrosser | admin1: did you enable debug and follow through the logs like i suggested? | 11:34 |
| jrosser | and silly question - you did source the openrc file? | 11:35 |
| noonedeadpunk | admin1: shouldn't `rgw_keystone_admin_domain = default` be `Default`? | 11:43 |
| noonedeadpunk | not sure if id is fine or not... | 11:44 |
| admin1 | i changed default to Default .. still the same | 11:45 |
| admin1 | and i did auth the openrc | 11:45 |
| noonedeadpunk | another thing is that I can recall some flukes when using python/style booleans | 11:45 |
| noonedeadpunk | ie `True` vs `true` | 11:45 |
| noonedeadpunk | I see all docs saying it should be text field to not be converted to True by ansible | 11:46 |
| noonedeadpunk | so we used `rgw_s3_auth_use_keystone: 'true'` explicitly | 11:46 |
| noonedeadpunk | same does tripleo | 11:47 |
| noonedeadpunk | and puppet, and helm... | 11:47 |
| admin1 | in our own , we use True vs true | 11:51 |
| admin1 | i setup keystone to debug, all i see there is 2022-02-11 11:51:09.227 3930 WARNING keystone.server.flask.application [req-54729a8d-f78a-4206-8b09-7c5064aaf386 - - - - -] Authorization failed. The request you have made requires authentication. from 172.29.236.31: keystone.exception.Unauthorized: The request you have made requires authentication .. | 11:51 |
| admin1 | and not why it failed | 11:51 |
| admin1 | i have tried both True/true ( caps, no caps) | 11:52 |
| admin1 | and Default / default | 11:52 |
| admin1 | if you guys have similar config and can share, maybe that will help me pinpoint what my issue is | 11:57 |
| noonedeadpunk | https://paste.opendev.org/show/b8BEvqAcpELZyYeLFXY9/ | 12:03 |
| admin1 | noonedeadpunk | 12:12 |
| admin1 | i have _ in all my configs | 12:12 |
| admin1 | you have none | 12:12 |
| noonedeadpunk | doesn't matter | 12:12 |
| admin1 | oh | 12:12 |
| noonedeadpunk | works both ways | 12:12 |
| admin1 | on the openstack side, the user should be added to Default service as admin role, that is done | 12:13 |
| admin1 | can you also share your endpoint list ( with the domain hidden ) .. i want to know if the endpoint is added correctly | 12:13 |
| admin1 | hmm.no error now .. on swift list .. yours was rgw keystone admin password .. while mine had rgw keystone admin token | 12:17 |
| admin1 | checking via horizon now | 12:17 |
| admin1 | and i also removed the _ | 12:17 |
| noonedeadpunk | yes, sec | 12:17 |
| noonedeadpunk | https://paste.opendev.org/show/bT43SKDjGSqKbJhBjfhN/ | 12:19 |
| admin1 | noonedeadpunk, jrosser .. mine is working now .. \o/ | 12:20 |
| admin1 | many thanks | 12:20 |
| noonedeadpunk | oh, what worked? | 12:20 |
| admin1 | osa -> object storage -> ceph | 12:20 |
| admin1 | creating containers, adding objects | 12:20 |
| admin1 | i was stuck on auth failure before | 12:20 |
| noonedeadpunk | I mean what was the issue :D | 12:21 |
| admin1 | one line was incorrect rgw keystone admin token vs rgw keystone admin password | 12:36 |
| admin1 | mine had token .. yours had password | 12:36 |
| admin1 | noonedeadpunk, one more thing .. when you make a container public and upload a png/jpg in it, it downloads .. shouldn't it open in the browser itself ? | 12:45 |
| noonedeadpunk | I think there's some setting for that... | 12:51 |
| noonedeadpunk | like for static websites | 12:51 |
| jrosser | admin1: the answer to your rgw problem was in the OSA AIO all along https://github.com/openstack/openstack-ansible/blob/master/inventory/group_vars/ceph-rgw.yml | 13:02 |
| jrosser | this is all deployed / tested in our ceph CI jobs so would have been a good place to compare | 13:03 |
| mgariepy | good morning thanks for all the input about memcached | 13:15 |
| mgariepy | noonedeadpunk, do you put memcached behind haproxy to replicate the writes on all of them ? | 13:16 |
| noonedeadpunk | yep, but based on source ip. so each backend goes to one memcached mainly | 14:06 |
| mgariepy | in case one is down. one is getting really slow ? | 14:09 |
| mgariepy | or one third** | 14:09 |
| evrardjp[m] | Haproxy doesn’t replicate writes | 14:18 |
| mgariepy | good point evrardjp[m] :) | 14:20 |
| evrardjp[m] | I have installed irc on my phone. That’s pretty evil. I can now be pedantic when I randomly see messages (as I can’t keep up with you folks!) ;) | 14:27 |
| spatel | evrardjp[m] do they have app for Phone? i would like to have one | 14:30 |
| spatel | I bought Textual7 for macbook pro for OSX | 14:31 |
| mgariepy | evrardjp[m], so we can ping you anytime now that's a good news ;p | 14:37 |
| spatel | lol | 14:51 |
| *** dviroel is now known as dviroel|lunch\ | 14:54 | |
| noonedeadpunk | sure there's irc client for phone | 15:01 |
| noonedeadpunk | though it's still recommended to have bouncer | 15:01 |
| noonedeadpunk | to have history when you're offline:) | 15:02 |
| spatel | I may need bouncer soon to stay connected | 15:03 |
| opendevreview | Dmitriy Rabotyagov proposed openstack/openstack-ansible-os_nova master: Implement nova direct RBD image retrieve https://review.opendev.org/c/openstack/openstack-ansible-os_nova/+/828897 | 15:24 |
| opendevreview | Dmitriy Rabotyagov proposed openstack/openstack-ansible-os_nova master: Implement nova direct RBD image retrieve https://review.opendev.org/c/openstack/openstack-ansible-os_nova/+/828897 | 15:25 |
| opendevreview | Dmitriy Rabotyagov proposed openstack/openstack-ansible-os_nova master: Implement nova direct RBD image retrieve https://review.opendev.org/c/openstack/openstack-ansible-os_nova/+/828897 | 15:31 |
| admin1 | noonedeadpunk, if you create a container and make it public .. it gives a link .. if you click the link, what does it say .. ? mine says NoSuchBucket .. when i upload an image and download it, it downloads fine ( from the browser ) .. but when i curl -I or try to wget, i get 500 error .. | 15:42 |
| admin1 | so i am trying to work out how this works | 15:42 |
| admin1 | my idea/expectation is that if there is a image/jpg/png, it just displays in the browser and not downloads it .. | 15:43 |
| admin1 | noonedeadpunk, please ignore my question | 15:44 |
| admin1 | works on lab .. not on prod . something mistake i did | 15:44 |
| opendevreview | Jonathan Rosser proposed openstack/openstack-ansible-repo_server stable/xena: Ensure insist=true is always set for lsyncd https://review.opendev.org/c/openstack/openstack-ansible-repo_server/+/828835 | 16:03 |
| opendevreview | Jonathan Rosser proposed openstack/openstack-ansible-repo_server stable/wallaby: Ensure insist=true is always set for lsyncd https://review.opendev.org/c/openstack/openstack-ansible-repo_server/+/828836 | 16:03 |
| opendevreview | Jonathan Rosser proposed openstack/openstack-ansible-repo_server stable/victoria: Ensure insist=true is always set for lsyncd https://review.opendev.org/c/openstack/openstack-ansible-repo_server/+/828837 | 16:03 |
| *** dviroel|lunch\ is now known as dviroel | 16:08 | |
| opendevreview | Dmitriy Rabotyagov proposed openstack/openstack-ansible-os_keystone stable/xena: Drop ProxyPass out of VHost https://review.opendev.org/c/openstack/openstack-ansible-os_keystone/+/828838 | 16:11 |
| noonedeadpunk | jrosser: I put some comments on keypairs topic | 16:34 |
| noonedeadpunk | I think main thing is to pass ssh_keypairs_setup_host for each role. As now we can generate ssh-authority on openstack_ssh_keypairs_setup_host but then nothing else aware of that host | 16:35 |
| * jrosser looks | 16:37 | |
| jrosser | ah yes we should probably do it like the pki role | 16:39 |
| jrosser | leave the ssh_keypairs role not knowing about OSA vars, but override through the roles that call it | 16:40 |
| evrardjp[m] | Seems a good pattern ;) | 16:48 |
| evrardjp[m] | noonedeadpunk: https://www.mirantis.com/blog/how-to-use-gpu-virtualization-with-mirantis-openstack-for-kubernetes/ | 16:49 |
| noonedeadpunk | damn ,they've stolen my talk! | 16:50 |
| noonedeadpunk | Fuck, and that folk is from Mirantis office from my city! | 16:51 |
| noonedeadpunk | uhhhh | 16:51 |
| jrosser | i think we finally get our gpu hosts next week | 16:52 |
| jrosser | something like 9months delivery time :/ | 16:52 |
| noonedeadpunk | And their HR's were so nice to me lately... I bet they were wondering about GPU details :D | 16:52 |
| noonedeadpunk | uhhhh | 16:58 |
| noonedeadpunk | I'm quite angry on Mirantis now. They should have released that half year later or just keep silent forever :D | 16:59 |
| noonedeadpunk | Well, still it's a bit outdated now with Ampere :D | 16:59 |
| noonedeadpunk | evrardjp[m]: you shouldn't have shown me this :D | 17:05 |
| noonedeadpunk | *half year earlier | 17:12 |
| noonedeadpunk | anyway | 17:12 |
| noonedeadpunk | jrosser: just in time for updated DLS image :D | 17:13 |
| opendevreview | Jonathan Rosser proposed openstack/openstack-ansible master: Create ssh certificate authority https://review.opendev.org/c/openstack/openstack-ansible/+/825292 | 17:13 |
| opendevreview | Jonathan Rosser proposed openstack/openstack-ansible-os_nova master: Use ssh_keypairs role to generate cold migration ssh keys https://review.opendev.org/c/openstack/openstack-ansible-os_nova/+/825306 | 17:15 |
| opendevreview | Jonathan Rosser proposed openstack/openstack-ansible-os_keystone master: Use ssh_keypairs role to generate fernet sync ssh keys https://review.opendev.org/c/openstack/openstack-ansible-os_keystone/+/827090 | 17:17 |
| opendevreview | Jonathan Rosser proposed openstack/openstack-ansible-repo_server master: Use ssh_keypairs role to generate keys for repo sync https://review.opendev.org/c/openstack/openstack-ansible-repo_server/+/827100 | 17:19 |
| jrosser | ^ i've made these more like the pki role | 17:19 |
| jrosser | i think i will also look at making the pki and ssh keys role keep the private keys in hashicorp vault rather than as files | 17:21 |
| jrosser | but in another patch | 17:21 |
| evrardjp[m] | noonedeadpunk: are you saying you’re interviewing with Mirantis ? Their CTO is a nice guy… too. ;) | 17:22 |
| noonedeadpunk | nah I'm not. | 17:24 |
| evrardjp[m] | Haha | 17:24 |
| noonedeadpunk | Just long history with them :D | 17:24 |
| admin1 | think i met evrardjp[m] and others in osa summit barcelona .. and i recall the mirantis party :) | 17:24 |
| evrardjp[m] | It’s been a while :) | 17:26 |
| evrardjp[m] | Is there a design document for why OSA now carries pki and ssh keygen roles ? | 17:28 |
| jrosser | there is a spec for pki stuff | 17:29 |
| evrardjp[m] | Awesome I will search for that when I am not on my phone | 17:29 |
| jrosser | and keygen is done primarily to get rid of the horror in os_nova with keys for migrations | 17:29 |
| jrosser | it is using signed keys so there is no need for fully populated authrorized_keys on all computes any more | 17:30 |
| evrardjp[m] | I can’t remember that part. I will read | 17:30 |
| evrardjp[m] | Yeah I know the principles of ssh ca ;) | 17:30 |
| evrardjp[m] | Didn’t remember we were using this. | 17:31 |
| jrosser | it is new stuff | 17:31 |
| evrardjp[m] | Need to read nova role | 17:31 |
| evrardjp[m] | :) | 17:31 |
| evrardjp[m] | Technology! Future is now! | 17:31 |
| noonedeadpunk | it's not only nova, but also repo container | 17:32 |
| noonedeadpunk | for lsync | 17:32 |
| noonedeadpunk | and keystone for sync of fernet | 17:33 |
| noonedeadpunk | so quite some places | 17:33 |
| jrosser | we could also use for octavia as well | 17:33 |
| noonedeadpunk | does octavia need ssh keys? I think it's more about pki role, isn't it? | 17:33 |
| jrosser | i think there is some stuff there to generate a debug key for the amphora | 17:34 |
| noonedeadpunk | oh, yes, indeed | 17:34 |
| jrosser | or maybe that got taken out - can't remember | 17:34 |
| noonedeadpunk | it is there | 17:34 |
| noonedeadpunk | I was thinking about more regular SSL for amphoras that also needs to be taken care of) | 17:34 |
| noonedeadpunk | so quite some work to do:) | 17:35 |
| jrosser | oh yes well that too | 17:35 |
| johnsom | Yeah, ssh is optional on the amps. Just there if the operator wants to get into the amp for some reason | 17:35 |
| jrosser | but i am liking the refactoring we are doing | 17:35 |
| noonedeadpunk | yes, totally | 17:35 |
| jrosser | too many places doing the same thing but different | 17:35 |
| noonedeadpunk | it's great clean-up | 17:35 |
| jrosser | downside is it probably increases runtime especially for AIO/CI | 17:36 |
| noonedeadpunk | evrardjp[m]: link to spec https://specs.openstack.org/openstack/openstack-ansible-specs/specs/wallaby/ssl-root-ca.html | 17:36 |
| jrosser | like when we moved to python_venv_build and uwsgi roles | 17:36 |
| jrosser | as these generic things always involve more tasks than something very specific | 17:36 |
| noonedeadpunk | I'm not really sure it does | 17:36 |
| noonedeadpunk | as if we take example with uwsgi and metal tests - same venv get re-used for all services | 17:37 |
| noonedeadpunk | also conditios and includes kind of help | 17:38 |
| jrosser | formatting has gone a bit odd in the ssl spec | 17:41 |
| noonedeadpunk | yeah | 17:41 |
| noonedeadpunk | just saw that | 17:41 |
| noonedeadpunk | I can recall verifying it's renderring before merging... | 17:42 |
| evrardjp[m] | johnsom: hello you :) hope you’re doing well ! | 17:42 |
| evrardjp[m] | noonedeadpunk: specs can be updated | 17:42 |
| evrardjp[m] | :) | 17:42 |
| johnsom | evrardjp[m] Hey there, hope you are well too! | 17:42 |
| evrardjp[m] | :) | 17:43 |
| noonedeadpunk | evrardjp[m]: I have some docs to write already :p | 17:43 |
| noonedeadpunk | will fix a bit later | 17:45 |
| spatel | noonedeadpunk question, can we limit common-mq to not run on all nova-computes? openstack-ansible setup-openstack.yml --tags common-mq | 17:47 |
| spatel | in my case i have 200 compute nodes and some of them are down so ansible takes hell of the time to finish it | 17:47 |
| noonedeadpunk | -- limit '!nova_compute' ? | 17:48 |
| jrosser | https://docs.ansible.com/ansible/latest/user_guide/intro_patterns.html | 17:48 |
| spatel | Oop! :) | 17:48 |
| spatel | I didn't know you can invert with ! | 17:48 |
| spatel | thank you! | 17:48 |
| noonedeadpunk | yup | 17:48 |
| opendevreview | Jonathan Rosser proposed openstack/openstack-ansible-os_keystone master: Use ssh_keypairs role to generate fernet sync ssh keys https://review.opendev.org/c/openstack/openstack-ansible-os_keystone/+/827090 | 17:59 |
| opendevreview | Jonathan Rosser proposed openstack/openstack-ansible-os_nova master: Use ssh_keypairs role to generate cold migration ssh keys https://review.opendev.org/c/openstack/openstack-ansible-os_nova/+/825306 | 18:00 |
| opendevreview | Jonathan Rosser proposed openstack/openstack-ansible master: Create ssh certificate authority https://review.opendev.org/c/openstack/openstack-ansible/+/825292 | 18:02 |
| jrosser | i took out all the things like `when: groups['keystone_all'] | length > 1` | 18:03 |
| jrosser | otherwise it doesnt run in CI which is not helpful | 18:03 |
| noonedeadpunk | well, we kind of hardly test it's functionality anyway except infra job for repo? | 18:04 |
| noonedeadpunk | but yeah, maybe it makes sense | 18:05 |
| noonedeadpunk | should we also return keystone_ssh_keypairs_dir to state of patchset 7 https://review.opendev.org/c/openstack/openstack-ansible-os_keystone/+/827090/7..10/defaults/main.yml#b554 ? | 18:06 |
| noonedeadpunk | tbh I dunno what's better - fallback to default always or get things defined in group_vars... | 18:07 |
| noonedeadpunk | having big group vars was bad for performance iirc | 18:07 |
| jrosser | well, i was looking at PS7 and thinking it was not nice that the directory layout of the deploy host needs to be in an os_<foo> role | 18:09 |
| noonedeadpunk | agree | 18:09 |
| jrosser | and that also breaks completely OSA_CONFIG_DIR | 18:10 |
| noonedeadpunk | some dummy default maybe? like /tmp or smth... | 18:10 |
| jrosser | which i may have already broken in the PKI role too tbh | 18:10 |
| noonedeadpunk | but yeah, maybe it's fine this way... we jsut didn't tend to leave things undefined in roles so catch my eye | 18:12 |
| noonedeadpunk | *catched | 18:12 |
| jrosser | hrrm https://github.com/openstack/openstack-ansible-os_nova/blob/master/defaults/main.yml#L556 | 18:12 |
| noonedeadpunk | oh, ok then :D | 18:13 |
| jrosser | maybe thats wrong too :) | 18:13 |
| noonedeadpunk | I bet I asked the same for PKI tbh | 18:13 |
| jrosser | can always improve this - no problem to change it | 18:13 |
| noonedeadpunk | agree | 18:13 |
| jrosser | not sure i know really what to do | 18:15 |
| noonedeadpunk | let's leave as is and improve if decide so later | 18:15 |
| jrosser | if we default it in the roles then OSA_CONFIG_DIR is broken for sure | 18:15 |
| jrosser | *more broken | 18:15 |
| noonedeadpunk | btw OSA_CONFIG_DIR is great point | 18:16 |
| jrosser | this is probably worth a bug in LP then we can fix it properly | 18:16 |
| noonedeadpunk | we must use `openstack_config_dir` everywhere I bet | 18:17 |
| noonedeadpunk | https://opendev.org/openstack/openstack-ansible/src/branch/master/inventory/group_vars/all/all.yml#L20 | 18:17 |
| opendevreview | Dmitriy Rabotyagov proposed openstack/openstack-ansible master: Fix behaviour of OSA_CONFIG_DIR https://review.opendev.org/c/openstack/openstack-ansible/+/828915 | 18:31 |
| noonedeadpunk | damn, and now https://review.opendev.org/c/openstack/openstack-ansible/+/825292fails :( | 18:32 |
| opendevreview | Merged openstack/openstack-ansible-rabbitmq_server master: Use sysctl ini-like config file https://review.opendev.org/c/openstack/openstack-ansible-rabbitmq_server/+/826338 | 18:38 |
| opendevreview | Merged openstack/openstack-ansible stable/xena: Bump os_zun role SHA to pick up new location for kata repo https://review.opendev.org/c/openstack/openstack-ansible/+/827412 | 18:38 |
| noonedeadpunk | wow | 18:42 |
| noonedeadpunk | now no need to push to gates anymore? | 18:42 |
| noonedeadpunk | https://review.opendev.org/c/openstack/openstack-ansible-rabbitmq_server/+/826463/3 got to zuul once lower patch get +w | 18:42 |
| jrosser | in the same gate queue i think that is automagic | 18:47 |
| jrosser | clarkb commented yesterday in #opendev that we were missing out on that by having gate-queue per repo | 18:47 |
| jrosser | though i'm not sure why we have it that way - there must have been a reason to do it | 18:47 |
| opendevreview | Dmitriy Rabotyagov proposed openstack/openstack-ansible master: Drop linuxcontainers hardcode for CI https://review.opendev.org/c/openstack/openstack-ansible/+/828918 | 19:00 |
| opendevreview | Merged openstack/openstack-ansible-rabbitmq_server master: Use systemd_service role for overrides https://review.opendev.org/c/openstack/openstack-ansible-rabbitmq_server/+/826463 | 19:19 |
| jrosser | this has been ready a while now https://review.opendev.org/c/openstack/openstack-ansible-os_designate/+/824335 | 19:36 |
| admin1 | i wish you all a very good weekend | 19:46 |
| mgariepy | have a nice one too | 19:50 |
| spatel | who has HP hardware here :) | 19:54 |
| mgariepy | i don't why ? | 19:54 |
| spatel | i want one bios firmware | 19:55 |
| spatel | HP asking me for account subscription :( | 19:55 |
| mgariepy | wow. | 19:55 |
| spatel | This is crazy that they lockdown for hardware firmware (it should be free) | 19:56 |
| jrosser | you need some less enterprise hardware | 19:56 |
| mgariepy | hahah | 19:56 |
| spatel | lol | 19:56 |
| spatel | no comments (they are cheaper) | 19:57 |
| spatel | slowly we are moving out from own datacenter to rental datacenter | 19:57 |
| spatel | so we don't need to worry about hardware anymore. | 19:58 |
| mgariepy | you should run in the cloud to skip the hardware :) haha | 19:58 |
| spatel | we are building cloud haha | 19:59 |
| mgariepy | spatel, what kind of server ? | 20:00 |
| spatel | We rent datacenter build openstack and run are service.. | 20:00 |
| mgariepy | which model do you have ? | 20:00 |
| spatel | Dell R 740xd / Dual Gold 6240 | 20:01 |
| mgariepy | the hp ones :) haha | 20:01 |
| spatel | HP C7000 chassis with BL460c Gen9 | 20:02 |
| spatel | https://ibb.co/2tHnjRq | 20:03 |
| spatel | we do run openstack on C7000 | 20:03 |
| spatel | This is one datacenter | 20:04 |
| spatel | we have one more which runs on all pizza box 1U but same HP hardware | 20:05 |
| mgariepy | cool. i do have dell c64xx for computes nodes. | 20:07 |
| spatel | We have dell in Rental Datacenter | 20:08 |
| mgariepy | dell are ok but their stupid license to have access to an html5 console is not fun. | 20:09 |
| spatel | yep!! HP doesn't have that issue and it comes with HTML5 | 20:10 |
| spatel | In Rental datacenter we have web portal which automatically bring up console so i don't need to deal with anything | 20:11 |
| spatel | question: in openstack network agent list output what is Alive vs State ? | 20:13 |
| opendevreview | Jonathan Rosser proposed openstack/openstack-ansible-repo_server stable/victoria: Ensure insist=true is always set for lsyncd https://review.opendev.org/c/openstack/openstack-ansible-repo_server/+/828837 | 20:15 |
| mgariepy | alive is probably when the process/ the agents reports it's status. | 20:16 |
| mgariepy | state would be if enabled or disabled ? | 20:16 |
| mgariepy | or something along those lines. | 20:17 |
| spatel | i have noticed when rabbitMQ was down State was still up but Alive was XXX | 20:20 |
| mgariepy | like: admin_state_up=True, alive=True | 20:20 |
| opendevreview | Jonathan Rosser proposed openstack/openstack-ansible master: Create ssh certificate authority https://review.opendev.org/c/openstack/openstack-ansible/+/825292 | 20:20 |
| mgariepy | i have a couple of compute nodes that are down and the admin_state_up=True, but alive=False | 20:21 |
| spatel | just trying to understand when State will be down | 20:23 |
| *** dviroel is now known as dviroel|out | 20:59 | |
| opendevreview | Merged openstack/openstack-ansible-os_designate master: Refactor use of include_vars https://review.opendev.org/c/openstack/openstack-ansible-os_designate/+/824335 | 21:31 |
| opendevreview | Dmitriy Rabotyagov proposed openstack/openstack-ansible-rabbitmq_server master: Verify if hosts file already managed with OSA https://review.opendev.org/c/openstack/openstack-ansible-rabbitmq_server/+/828929 | 22:22 |
| opendevreview | Dmitriy Rabotyagov proposed openstack/openstack-ansible-rabbitmq_server stable/xena: Unpin Erlang distribution server port https://review.opendev.org/c/openstack/openstack-ansible-rabbitmq_server/+/828932 | 22:39 |
| opendevreview | Merged openstack/openstack-ansible stable/xena: Rename RBD cinder backend https://review.opendev.org/c/openstack/openstack-ansible/+/828663 | 23:05 |
Generated by irclog2html.py 2.17.3 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!