openstackgerrit | Merged openstack/openstack-ansible stable/train: Bump SHAs for stable/train https://review.opendev.org/760728 | 00:30 |
---|---|---|
*** macz_ has quit IRC | 00:35 | |
openstackgerrit | Merged openstack/openstack-ansible stable/ussuri: Bump SHAs for stable/ussuri https://review.opendev.org/760731 | 00:48 |
openstackgerrit | Merged openstack/openstack-ansible master: Improve repo server haproxy healthcheck to verify repo synchronisation https://review.opendev.org/660477 | 00:51 |
*** gyee has quit IRC | 00:53 | |
*** cshen has joined #openstack-ansible | 00:59 | |
*** cshen has quit IRC | 01:03 | |
*** klamath_atx has quit IRC | 01:42 | |
*** pto has joined #openstack-ansible | 01:53 | |
*** pto has quit IRC | 01:58 | |
*** NewJorg has quit IRC | 02:08 | |
*** NewJorg has joined #openstack-ansible | 02:08 | |
*** klamath_atx has joined #openstack-ansible | 02:11 | |
*** brad[] has quit IRC | 02:34 | |
*** brad[] has joined #openstack-ansible | 02:40 | |
*** cshen has joined #openstack-ansible | 02:59 | |
*** cshen has quit IRC | 03:03 | |
*** nurdie has quit IRC | 03:58 | |
*** cshen has joined #openstack-ansible | 04:59 | |
*** macz_ has joined #openstack-ansible | 05:01 | |
*** nurdie has joined #openstack-ansible | 05:03 | |
*** cshen has quit IRC | 05:04 | |
*** macz_ has quit IRC | 05:05 | |
*** nurdie has quit IRC | 05:09 | |
*** NewJorg has quit IRC | 05:10 | |
*** NewJorg has joined #openstack-ansible | 05:11 | |
*** evrardjp has quit IRC | 05:33 | |
*** evrardjp has joined #openstack-ansible | 05:33 | |
*** pto has joined #openstack-ansible | 05:55 | |
*** pto has quit IRC | 05:59 | |
*** rpittau|afk is now known as rpittau | 06:47 | |
*** cshen has joined #openstack-ansible | 06:59 | |
*** miloa has joined #openstack-ansible | 07:03 | |
*** cshen has quit IRC | 07:05 | |
*** nurdie has joined #openstack-ansible | 07:05 | |
*** nurdie has quit IRC | 07:10 | |
*** pto has joined #openstack-ansible | 07:15 | |
*** cshen has joined #openstack-ansible | 07:18 | |
*** pto has quit IRC | 07:20 | |
*** pto has joined #openstack-ansible | 07:29 | |
*** pto_ has joined #openstack-ansible | 07:32 | |
*** pto has quit IRC | 07:35 | |
*** spatel has joined #openstack-ansible | 07:35 | |
*** spatel has quit IRC | 07:39 | |
*** luksky has joined #openstack-ansible | 07:56 | |
openstackgerrit | Dmitriy Rabotyagov (noonedeadpunk) proposed openstack/openstack-ansible stable/stein: Actually use ANSIBLE_FORKS in openstack-ansible.rc https://review.opendev.org/760540 | 08:03 |
*** andrewbonney has joined #openstack-ansible | 08:11 | |
noonedeadpunk | morning | 08:29 |
*** cshen has quit IRC | 08:59 | |
*** cshen has joined #openstack-ansible | 09:00 | |
*** nurdie has joined #openstack-ansible | 09:06 | |
*** nurdie has quit IRC | 09:11 | |
*** shyamb has joined #openstack-ansible | 09:29 | |
openstackgerrit | Merged openstack/openstack-ansible-repo_server master: Use infra jobs for testing https://review.opendev.org/760960 | 09:37 |
Adri2000 | hello jrosser, if I read correctly, this bug report https://bugs.launchpad.net/openstack-ansible/+bug/1882886 can now be set as "fix released"? | 09:54 |
openstack | Launchpad bug 1882886 in openstack-ansible "Horizon dropdown menu is empty in Stein" [Undecided,Fix committed] | 09:54 |
openstackgerrit | Merged openstack/openstack-ansible-os_manila master: Add Ubuntu Focal CI jobs https://review.opendev.org/754653 | 10:55 |
*** nurdie has joined #openstack-ansible | 11:08 | |
*** nurdie has quit IRC | 11:13 | |
*** yann-kaelig has joined #openstack-ansible | 11:14 | |
*** nurdie has joined #openstack-ansible | 11:40 | |
*** nurdie has quit IRC | 11:45 | |
*** yann-kaelig has quit IRC | 11:50 | |
*** pto_ has quit IRC | 11:58 | |
*** pto has joined #openstack-ansible | 11:59 | |
*** NewJorg has quit IRC | 12:02 | |
*** NewJorg has joined #openstack-ansible | 12:04 | |
*** rfolco has joined #openstack-ansible | 12:06 | |
openstackgerrit | Dmitriy Rabotyagov (noonedeadpunk) proposed openstack/openstack-ansible-os_neutron master: Implement uWSGI for neutron-api https://review.opendev.org/486156 | 12:21 |
openstackgerrit | Merged openstack/openstack-ansible stable/stein: Actually use ANSIBLE_FORKS in openstack-ansible.rc https://review.opendev.org/760540 | 12:31 |
*** shyamb has quit IRC | 12:44 | |
*** jamesdenton has quit IRC | 12:48 | |
openstackgerrit | Dmitriy Rabotyagov (noonedeadpunk) proposed openstack/openstack-ansible-os_manila master: Cleanup after repo_build and pip_install retirement https://review.opendev.org/727659 | 12:56 |
openstackgerrit | Dmitriy Rabotyagov (noonedeadpunk) proposed openstack/openstack-ansible-os_manila master: Fix pygments style https://review.opendev.org/729513 | 12:57 |
*** nurdie has joined #openstack-ansible | 12:57 | |
openstackgerrit | Dmitriy Rabotyagov (noonedeadpunk) proposed openstack/openstack-ansible-os_manila master: Start using uWSGI role https://review.opendev.org/704935 | 12:59 |
*** nurdie has quit IRC | 13:03 | |
*** miloa has quit IRC | 13:05 | |
*** tosky has joined #openstack-ansible | 13:07 | |
*** rh-jelabarre has joined #openstack-ansible | 13:09 | |
*** rh-jelabarre has quit IRC | 13:13 | |
*** rh-jelabarre has joined #openstack-ansible | 13:14 | |
openstackgerrit | Dmitriy Rabotyagov (noonedeadpunk) proposed openstack/openstack-ansible-os_glance stable/stein: Do not symlink ceph libraries for distro path https://review.opendev.org/760818 | 13:18 |
openstackgerrit | Dmitriy Rabotyagov (noonedeadpunk) proposed openstack/openstack-ansible-tests stable/stein: Add requirements_git_install_branch to test-vars https://review.opendev.org/732522 | 13:27 |
*** nurdie has joined #openstack-ansible | 13:34 | |
*** nurdie has quit IRC | 13:39 | |
*** d34dh0r53 has quit IRC | 13:43 | |
*** pto has quit IRC | 13:50 | |
*** pto_ has joined #openstack-ansible | 13:50 | |
*** pto_ has quit IRC | 13:50 | |
*** pto has joined #openstack-ansible | 13:51 | |
openstackgerrit | Rafael Folco proposed openstack/openstack-ansible-os_tempest stable/ussuri: Build containers for single consumer job https://review.opendev.org/761019 | 13:55 |
openstackgerrit | Rafael Folco proposed openstack/openstack-ansible-os_tempest stable/train: Build containers for single consumer job https://review.opendev.org/761021 | 13:56 |
*** d34dh0r53 has joined #openstack-ansible | 14:01 | |
odyssey4me | hey folks, I have had an interesting proposal in https://github.com/ansible-collections/community.rabbitmq/issues/45 and I wonder if y'all think that it'd be a good idea to pull the rabbitmq_server role into the collection for broader use. | 14:05 |
odyssey4me | One downer thing is that it'd fragment the development a little because it'd be using the GH development process, but it may be worth that cost to gain the broader audience. | 14:06 |
*** spatel has joined #openstack-ansible | 14:09 | |
noonedeadpunk | odyssey4me: well, we were thinking about adding our infra roles to the galaxy, but I'm not sure what you've ment about GH process? | 14:16 |
noonedeadpunk | like ansible-collections-openstack are in opendev and are good with that? | 14:17 |
noonedeadpunk | and eventually there's one requirement that needs to be done, which is config_template | 14:18 |
noonedeadpunk | we need to make it collection as well I guess to be able to use it for roles afterwards | 14:18 |
odyssey4me | noonedeadpunk: no, this is different - the request in that ticket is to have roles to setup rabbitmq in the community.rabbitmq collection... and the suggestion is perhaps to seed that from the OSA role | 14:18 |
noonedeadpunk | ah | 14:19 |
noonedeadpunk | well it will influence CI as well, since we won't able to use depends on and other fancy stuff | 14:21 |
noonedeadpunk | I'm kind of for making role easier to use outside of the osa, but I don't like idea of losing control, since we depend on it really heavily | 14:22 |
noonedeadpunk | jrosser: what do you think about it? | 14:23 |
noonedeadpunk | (me hates working with github as well :p) | 14:24 |
*** nurdie has joined #openstack-ansible | 14:31 | |
*** jamesdenton has joined #openstack-ansible | 14:32 | |
noonedeadpunk | well, rabbitmq seems to not be using config_template as for now | 14:34 |
noonedeadpunk | uh, seems I start stop liking idea of collections.. So like to get just mongo module now I should pull tons of roles as well? | 14:36 |
noonedeadpunk | It sucks tbh... | 14:36 |
noonedeadpunk | Considering galaxy itself instantly fails it makes things even worse when it comes to retirevment of collections.... | 14:39 |
noonedeadpunk | maybe we should move all openstack project roles to openstack collection, huh?:) | 15:03 |
*** suryasingh has joined #openstack-ansible | 15:07 | |
noonedeadpunk | in the meanwhile vote on https://review.opendev.org/#/c/732522/ would be awesome | 15:07 |
*** nurdie has quit IRC | 15:15 | |
*** nurdie has joined #openstack-ansible | 15:15 | |
ThiagoCMC | Hey guys, I'm curious about something... When deploying OSA with Ceph, the br-storage becomes the Ceph Private, and the br-mgmt the Ceph Public. Thing is, why the "group_binds" has glance_api, cinder_api, and even cinder_volume under it?! I see no reason for those services to reach the Ceph Private networks... Any idea? | 15:15 |
noonedeadpunk | Um, the idea here is that all storage traffic should go through storage network only | 15:16 |
*** jbadiapa has quit IRC | 15:16 | |
noonedeadpunk | as storage network might have another throughput and equipment underneath | 15:16 |
noonedeadpunk | so iteraction of services with ceph monitors should also occur through storage network iirc | 15:17 |
ThiagoCMC | Exactly, my Ceph Private "br-storage" is completely isolated in hardware! | 15:17 |
noonedeadpunk | but anyway, ceph-client needs to reach osds directly, which are available only storage network by design | 15:17 |
ThiagoCMC | But Ceph mon talks via br-mgmt (ceph public) | 15:17 |
ThiagoCMC | with the Ceph OSDs | 15:17 |
*** rpittau is now known as rpittau|afk | 15:18 | |
noonedeadpunk | I'm not sure it's correct honestly | 15:18 |
ThiagoCMC | Ceph OSD as 2 interfaces, br-mgmt (available everywhere) and br-storage (isolated) | 15:18 |
ThiagoCMC | while OSA controllers only has br-mgmt, not br-storage | 15:18 |
noonedeadpunk | well, we have weird aio which doesn not use br-storage at all, but ti's not good example to follow in production | 15:19 |
ThiagoCMC | I see, I'm not using the AIO | 15:19 |
noonedeadpunk | osa controllers should have br-storage as well | 15:19 |
ThiagoCMC | Why? | 15:19 |
spatel | I do have br-storage on controller node | 15:19 |
ThiagoCMC | It belongs to ceph private, for OSDs to talk to each other | 15:19 |
spatel | br-storage connected to cinder | 15:20 |
noonedeadpunk | well, the most traffic goes not between osds but from client to osd | 15:20 |
noonedeadpunk | so you need to have good throughput here | 15:20 |
ThiagoCMC | That happens via ceph public br-mgmt | 15:20 |
noonedeadpunk | you can do without br-storage at all actually | 15:20 |
noonedeadpunk | it's just matter of traffic siolation | 15:21 |
ThiagoCMC | =P | 15:21 |
ThiagoCMC | Exactly | 15:21 |
noonedeadpunk | and by design it was that storage traffic shouldn't go through mgmt | 15:21 |
ThiagoCMC | Sure | 15:21 |
gillesMo | Hello, I've just upgraded my Rocky install to Ubuntu 18. And it's a mess. It seems I have problems with rabbitmq. neutron-l3-agent cannot contact neutron-server, but both are started, and I see them connected in the rabbimq logs | 15:22 |
openstackgerrit | Merged openstack/openstack-ansible-os_neutron stable/stein: Enable notifications when Designate is enabled https://review.opendev.org/761242 | 15:23 |
spatel | ThiagoCMC: If i am not wrong then br-mgmt is just for ansible for deploying stuff and some monitoring of Ceph OSD. (I believe we don't need br-storage interface on controller) | 15:25 |
ThiagoCMC | According to openstack_user_config.yml.prod-ceph.example, br-mgmt becomes Ceph Public and br-storage becomes Ceph Private (exclusively for OSDs to talk among themselves and with the Compute Nodes). | 15:26 |
ThiagoCMC | So, the way all Ceph Clients within an OSA+Ceph deployment talk with Ceph, is via br-mgmt... | 15:28 |
spatel | all compute nodes (client) use br-storage to talk to ceph OSD nodes. (not br-mgmt ) | 15:32 |
ThiagoCMC | yes | 15:33 |
ThiagoCMC | I mean lol | 15:33 |
ThiagoCMC | compute nodes has both br-mgmt and br-storage, but the controllers doesn't have br-storage, they don't need access to the Ceph's storage area | 15:33 |
ThiagoCMC | neither the Network Nodes, they also doesn't have br-storage. | 15:34 |
spatel | Yes (no requirement to have br-storage on controller) | 15:35 |
gillesMo | someone already have errors with cinder playbbok, the "online_data_migrations" fails but I'm on the same version... | 15:35 |
ThiagoCMC | cool | 15:35 |
ThiagoCMC | *Only* Ceph OSD nodes and Compute Nodes participate in br-storage :-D | 15:35 |
spatel | I did create br-storage on controller based on example config, but look like we don't need that. | 15:35 |
ThiagoCMC | So, then, I believe that the provided example: "openstack_user_config.yml.prod-ceph.example" is wrong/misleading. | 15:35 |
spatel | But i believe your deployment node need br-storage (because ansible playbook try to check that interface, i didn't know why but i believe i had that issue) | 15:37 |
ThiagoCMC | It does that because of "group_binds:" of 'container_bridge: "br-storage"' has glance_api, cinder_api, cinder_volume... | 15:37 |
spatel | let me do tcpdump on br-storage to see what kind of traffic flowing | 15:37 |
ThiagoCMC | Nice! :-D | 15:38 |
ThiagoCMC | I'll be away for a couple of hours but IRC will remain open =P | 15:38 |
spatel | I am not seeing any traffic other then STP :) | 15:38 |
spatel | jrosser: are you around? | 15:40 |
*** cshen has quit IRC | 15:47 | |
*** cshen has joined #openstack-ansible | 15:48 | |
*** cshen has quit IRC | 16:05 | |
*** sshnaidm_ has joined #openstack-ansible | 16:12 | |
*** sshnaidm|afk has quit IRC | 16:13 | |
*** spatel has quit IRC | 16:18 | |
*** sshnaidm_ is now known as sshnaidm|afk | 16:21 | |
*** macz_ has joined #openstack-ansible | 16:30 | |
*** gyee has joined #openstack-ansible | 16:33 | |
*** cshen has joined #openstack-ansible | 16:33 | |
*** cshen has quit IRC | 16:37 | |
*** cshen has joined #openstack-ansible | 16:44 | |
*** spatel has joined #openstack-ansible | 16:46 | |
admin0 | I see a lot of traffic in my br-storage | 16:49 |
admin0 | in a cluster using ceph | 16:49 |
admin0 | because my mons are in the controllers | 16:50 |
admin0 | and in a more recent cluster where the mons and ceph is external, i still see traffic on br-storage on the controllers | 16:53 |
admin0 | so I dont think the config is wrong | 16:54 |
jrosser | ThiagoCMC: I think that the ceph example is maybe wrong | 17:43 |
*** suryasingh has quit IRC | 17:45 | |
ThiagoCMC | jrosser, admin0 thanks for the input! I'll keep trying to fully isolate my Ceph OSDs by not exposive br-storage to the controllers. =) | 17:53 |
jrosser | for my deployments br-storage is the ceph public network on computers and infra hosts//containers | 18:00 |
jrosser | there is an entirely separate ceph private network that’s completely out of scope of OSA | 18:00 |
jrosser | the controllers will need access to the ceph public network for external ceph deployments as glance / cinder do rbd operations on the storage from the controller nodes | 18:04 |
spatel | jrosser: hmmm | 18:06 |
spatel | that make sense now | 18:06 |
*** andrewbonney has quit IRC | 18:07 | |
spatel | for glance sure we need br-storage i totally forgot that | 18:07 |
spatel | jrosser: i have question related br-mgmt vs br-host (you said last time you have br-mgmt routable but how do you direct SSH to container IP?) | 18:08 |
spatel | container default gw is lxc bridge interface. | 18:09 |
jrosser | lxcbr0 yes | 18:09 |
spatel | i am trying to create that model but found i can't ssh my container IP directly. | 18:10 |
jrosser | I’m not sure that is necessary? | 18:10 |
jrosser | maybe to put it differently, I ssh to the host then lxc attach | 18:11 |
spatel | my monitoring system is in different subnet so i thought if i have reachability to that network from container then i can directly monitor containers | 18:11 |
jrosser | not ssh to the container | 18:11 |
jrosser | perhaps the easiest thing to do is give the monitoring an IP on be-mgmt | 18:12 |
jrosser | *br-mgmt | 18:12 |
spatel | i want to install monitoring agent on container to send data to my monitoring system | 18:12 |
jrosser | yes I do this, and I extend mgmt network to logstash nodes for example to enable that whilst keeping mgmt network isolated | 18:13 |
spatel | hmm | 18:13 |
jrosser | decide what is routable and what’s not | 18:14 |
jrosser | if something isn’t, then bring extra interfaces into things where needed (like extra vlan tag on trunk for example) | 18:15 |
spatel | let me see how i can handle that | 18:15 |
spatel | that is what i did in my old datacenter, attached br-mgmt to my ELK and Monitoring system | 18:16 |
jrosser | another example, if you make external ceph, radosgw would need mgmt network to talk auth to keystone | 18:16 |
jrosser | so you have to consider all these cases similarly | 18:16 |
spatel | hmm | 18:16 |
spatel | totally | 18:16 |
jrosser | unless of course all the subnets just route to each other through some Nat/firewall | 18:17 |
ThiagoCMC | Interesting points! | 18:17 |
jrosser | but I have been really strict not to do that | 18:17 |
spatel | jrosser: do you know how to add custom repo? | 18:23 |
spatel | can we do that via user_variables? | 18:24 |
jrosser | spatel: i can be helpful sometimes to make extra subnets, like in ceph public there can be a /24 for "clients of ceph in the openstack controllers" "clients of ceph in the compute nodes" "ceph OSD on the storage nodes" | 18:24 |
jrosser | the examples really are just examples, you can cut this up however makes the most sense for you | 18:24 |
spatel | jrosser: currently my ceph just serving glance images so i have very simple storage network but in future i may add rados gateway etc for s3 style services that time i might need something fancy | 18:26 |
jrosser | spatel: maybe this https://github.com/openstack/openstack-ansible-openstack_hosts/blob/master/defaults/main.yml#L60 ? | 18:26 |
spatel | sweet! | 18:28 |
spatel | jrosser: | 18:28 |
jrosser | then as usual codesearch will tell you how thats used http://codesearch.openstack.org/?q=openstack_hosts_package_repos | 18:28 |
spatel | I want to add some NFV repo to test deployment so thought rather than messing ansible role, it would be good to just add in user_variable | 18:29 |
spatel | thanks | 18:31 |
*** dave-mccowan has quit IRC | 18:49 | |
*** cshen has quit IRC | 19:35 | |
*** luksky has quit IRC | 19:47 | |
*** luksky has joined #openstack-ansible | 19:48 | |
*** cshen has joined #openstack-ansible | 20:22 | |
ThiagoCMC | Thank you guys for all the help during this week! I wish an awesome weekend for everybody! | 20:36 |
*** cshen has quit IRC | 20:44 | |
*** cshen has joined #openstack-ansible | 20:50 | |
fanfi | Guys, could somebody help me please. When I trying to create a new image I got following error and I cant find any error in log files. :( ...Image creation failed: Unable to establish connection to http://172.16.1.91:9292/v2/images/b333b5df-29bb-40a7-adba-0fbbcc10d759/file : ('Connection aborted.', BrokenPipeError(32, 'Broken pipe')) | 20:50 |
fanfi | and glace api works | 20:50 |
fanfi | there is commands debug https://pastebin.com/7tVeXWSN | 21:05 |
*** melwitt is now known as jgwentworth | 21:26 | |
*** spatel has quit IRC | 21:32 | |
*** pcaruana has quit IRC | 21:34 | |
*** cshen has quit IRC | 22:03 | |
*** SmearedBeard has joined #openstack-ansible | 22:12 | |
*** rfolco has quit IRC | 22:14 | |
openstackgerrit | Merged openstack/openstack-ansible-os_manila master: Cleanup after repo_build and pip_install retirement https://review.opendev.org/727659 | 22:18 |
*** rh-jelabarre has quit IRC | 22:21 | |
*** cshen has joined #openstack-ansible | 22:24 | |
*** luksky has quit IRC | 22:26 | |
*** pfsmorigo has joined #openstack-ansible | 22:30 | |
*** dave-mccowan has joined #openstack-ansible | 23:07 | |
*** pto has quit IRC | 23:56 |
Generated by irclog2html.py 2.17.2 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!