| *** weifan has joined #openstack-ansible | 00:25 | |
| *** weifan has quit IRC | 00:25 | |
| *** gyee has quit IRC | 00:28 | |
| *** dmsimard4 is now known as dmsimard | 00:32 | |
| *** irclogbot_1 has joined #openstack-ansible | 00:54 | |
| *** irclogbot_1 has quit IRC | 01:00 | |
| *** kplant has quit IRC | 01:15 | |
| *** irclogbot_2 has joined #openstack-ansible | 01:24 | |
| *** irclogbot_2 has quit IRC | 01:34 | |
| *** irclogbot_3 has joined #openstack-ansible | 02:24 | |
| *** irclogbot_3 has quit IRC | 02:30 | |
| *** altlogbot_0 has joined #openstack-ansible | 02:46 | |
| *** altlogbot_0 has quit IRC | 02:52 | |
| *** irclogbot_1 has joined #openstack-ansible | 03:20 | |
| *** BjoernT has joined #openstack-ansible | 03:21 | |
| *** irclogbot_1 has quit IRC | 03:26 | |
| *** BjoernT has quit IRC | 03:26 | |
| *** BjoernT has joined #openstack-ansible | 03:27 | |
| *** irclogbot_1 has joined #openstack-ansible | 03:50 | |
| *** irclogbot_1 has quit IRC | 03:56 | |
| *** BjoernT has quit IRC | 04:02 | |
| *** udesale has joined #openstack-ansible | 04:05 | |
| *** poopcat has quit IRC | 04:52 | |
| *** irclogbot_1 has joined #openstack-ansible | 04:58 | |
| *** pcaruana has joined #openstack-ansible | 05:13 | |
| *** irclogbot_1 has quit IRC | 05:16 | |
| *** aedc has quit IRC | 05:47 | |
| *** jbadiapa has quit IRC | 05:55 | |
| *** cshen has joined #openstack-ansible | 05:59 | |
| *** cshen has quit IRC | 06:03 | |
| *** cshen has joined #openstack-ansible | 06:07 | |
| *** altlogbot_3 has joined #openstack-ansible | 06:08 | |
| *** altlogbot_3 has quit IRC | 06:13 | |
| *** luksky11 has joined #openstack-ansible | 06:14 | |
| *** Namrata has joined #openstack-ansible | 06:32 | |
| *** irclogbot_2 has joined #openstack-ansible | 06:34 | |
| *** irclogbot_2 has quit IRC | 06:40 | |
| *** cshen has quit IRC | 06:41 | |
| *** ivve has joined #openstack-ansible | 06:46 | |
| *** aedc has joined #openstack-ansible | 06:58 | |
| *** miloa has joined #openstack-ansible | 07:02 | |
| miloa | Morning | 07:02 |
|---|---|---|
| *** cshen has joined #openstack-ansible | 07:03 | |
| *** ygk_12345 has joined #openstack-ansible | 07:06 | |
| ygk_12345 | hi all | 07:06 |
| ygk_12345 | i facing an issue with my rocky setup | 07:07 |
| ygk_12345 | the le agent is taking a lot of time to add subnet interfaces to the router | 07:07 |
| ygk_12345 | *l3 agent | 07:07 |
| *** Namrata has quit IRC | 07:13 | |
| *** cshen has quit IRC | 07:14 | |
| *** cshen has joined #openstack-ansible | 07:15 | |
| *** rpittau|afk is now known as rpittau | 07:18 | |
| *** tosky has joined #openstack-ansible | 07:19 | |
| *** jbadiapa has joined #openstack-ansible | 07:22 | |
| *** jbadiapa has quit IRC | 07:27 | |
| *** Namrata has joined #openstack-ansible | 07:29 | |
| *** cshen has quit IRC | 07:30 | |
| *** cshen has joined #openstack-ansible | 07:30 | |
| *** ygk_12345 has quit IRC | 07:31 | |
| *** cshen has quit IRC | 07:43 | |
| *** aedc has quit IRC | 07:48 | |
| *** shyamb has joined #openstack-ansible | 07:49 | |
| *** aedc has joined #openstack-ansible | 07:54 | |
| *** cshen has joined #openstack-ansible | 07:55 | |
| *** shyamb has quit IRC | 07:59 | |
| *** altlogbot_1 has joined #openstack-ansible | 08:00 | |
| openstackgerrit | Dmitriy Rabotyagov (noonedeadpunk) proposed openstack/openstack-ansible-os_neutron master: service_setup: refactor service setup to a single file https://review.opendev.org/670460 | 08:01 |
| openstackgerrit | Dmitriy Rabotyagov (noonedeadpunk) proposed openstack/openstack-ansible-os_neutron master: service_setup: refactor service setup to a single file https://review.opendev.org/670460 | 08:03 |
| *** altlogbot_1 has quit IRC | 08:04 | |
| *** yolanda has quit IRC | 08:08 | |
| *** yolanda has joined #openstack-ansible | 08:09 | |
| *** cshen has quit IRC | 08:10 | |
| *** altlogbot_2 has joined #openstack-ansible | 08:10 | |
| *** gkadam has joined #openstack-ansible | 08:15 | |
| openstackgerrit | Dmitriy Rabotyagov (noonedeadpunk) proposed openstack/openstack-ansible-os_gnocchi master: service_setup: refactor service setup to a single file https://review.opendev.org/670471 | 08:15 |
| *** altlogbot_2 has quit IRC | 08:16 | |
| *** altlogbot_2 has joined #openstack-ansible | 08:16 | |
| *** gkadam has quit IRC | 08:17 | |
| *** cshen has joined #openstack-ansible | 08:20 | |
| *** altlogbot_2 has quit IRC | 08:22 | |
| *** altlogbot_1 has joined #openstack-ansible | 08:22 | |
| Namrata | Hi folks while upgrading from Rocky to Stein the upgrade script fails on openstack-ansible setup-infrastructure.yml -e 'galera_upgrade=true' -e 'rabbitmq_upgrade=true'it fails at http://paste.openstack.org/show/754340/ and what I can see is there is no galera_cluster_name variable in /etc/openstack_deploy/user_*.yml but there is wsrep_cluster_name | 08:24 |
| Namrata | in running galera cluster as the failure message suggests. | 08:24 |
| Namrata | how to solve this failure | 08:24 |
| openstackgerrit | Dmitriy Rabotyagov (noonedeadpunk) proposed openstack/openstack-ansible-os_ceilometer master: service_setup: refactor service setup to a single file https://review.opendev.org/670473 | 08:25 |
| *** altlogbot_1 has quit IRC | 08:28 | |
| *** cshen has quit IRC | 08:33 | |
| *** heikkine has joined #openstack-ansible | 08:39 | |
| *** cshen has joined #openstack-ansible | 08:42 | |
| *** shyamb has joined #openstack-ansible | 08:46 | |
| openstackgerrit | Jonathan Rosser proposed openstack/openstack-ansible master: Differentiate AIO /etc/hosts entries for localhost and aio1 https://review.opendev.org/670392 | 08:46 |
| *** cshen has quit IRC | 08:46 | |
| *** aedc has quit IRC | 08:48 | |
| jrosser | Namrata: going from rocky to stein changes the wsrep method for galera in a way that requires a manual intervention during the upgrade | 08:48 |
| Namrata | jrosser what changes should I make | 08:49 |
| Namrata | and the manual intervention is for time being till the stein release is not prepared ? | 08:50 |
| jrosser | Namrata: i copy this for you off our internal wiki http://paste.openstack.org/show/754342/ | 08:50 |
| jrosser | the issue is that the upgrade tooling does not handle this properly | 08:51 |
| Namrata | jrosser thanks a lot | 08:51 |
| *** altlogbot_0 has joined #openstack-ansible | 08:52 | |
| Namrata | jrosser can you address the other two issues http://paste.openstack.org/show/754202/ http://paste.openstack.org/show/754194/ | 08:54 |
| Namrata | this also needs manual intervention | 08:54 |
| Namrata | while running upgarde scripy | 08:54 |
| Namrata | *script | 08:54 |
| jrosser | as far as i know passing the --skip-tags worked for me for the first one | 08:55 |
| jrosser | and the second one you need to manually checkout the right sha - i've not found the root cause of why the ansible git module does that when the equivalent cli command does not seem to | 08:55 |
| Namrata | --skip-tags works for me too but that should be not the ideal case or not I wanted to confirm | 08:56 |
| Namrata | and second one i also checked out the right sha i am also not able to identify why there is an issue | 08:57 |
| *** altlogbot_0 has quit IRC | 08:58 | |
| *** irclogbot_2 has joined #openstack-ansible | 08:58 | |
| jrosser | Namrata: this git issue always happens with the same set of repos so i imagine it's a sublte bug in ansible which is only triggered by somthing unusual in those repos, like a missing tag or a historical force-push having made the history slightly inconsistent | 09:02 |
| *** irclogbot_2 has quit IRC | 09:04 | |
| Namrata | @jrosser thanks for addressing the issues. thanks for your time. | 09:04 |
| *** luksky11 has quit IRC | 09:09 | |
| CeeMac | morning | 09:18 |
| *** luksky11 has joined #openstack-ansible | 09:32 | |
| *** vnogin has joined #openstack-ansible | 09:44 | |
| *** shyamb has quit IRC | 10:05 | |
| *** irclogbot_0 has joined #openstack-ansible | 10:30 | |
| *** yolanda has quit IRC | 10:31 | |
| *** yolanda has joined #openstack-ansible | 10:32 | |
| *** irclogbot_0 has quit IRC | 10:38 | |
| *** shyamb has joined #openstack-ansible | 10:40 | |
| *** irclogbot_0 has joined #openstack-ansible | 10:41 | |
| *** irclogbot_0 has quit IRC | 10:44 | |
| *** ivve has quit IRC | 11:01 | |
| *** ivve has joined #openstack-ansible | 11:01 | |
| *** altlogbot_1 has joined #openstack-ansible | 11:03 | |
| *** miloa has quit IRC | 11:05 | |
| *** shyam89 has joined #openstack-ansible | 11:08 | |
| *** altlogbot_1 has quit IRC | 11:08 | |
| *** shyamb has quit IRC | 11:08 | |
| *** altlogbot_2 has joined #openstack-ansible | 11:11 | |
| *** kplant has joined #openstack-ansible | 11:14 | |
| *** altlogbot_2 has quit IRC | 11:16 | |
| *** cshen has joined #openstack-ansible | 11:25 | |
| *** cshen has quit IRC | 11:34 | |
| *** cshen has joined #openstack-ansible | 11:39 | |
| *** shyam89 has quit IRC | 11:39 | |
| *** shyam89 has joined #openstack-ansible | 11:48 | |
| *** vnogin has quit IRC | 11:49 | |
| *** cshen has quit IRC | 11:53 | |
| *** aedc has joined #openstack-ansible | 11:54 | |
| *** udesale has quit IRC | 11:59 | |
| *** udesale has joined #openstack-ansible | 12:00 | |
| *** cshen has joined #openstack-ansible | 12:04 | |
| *** altlogbot_3 has joined #openstack-ansible | 12:07 | |
| *** altlogbot_3 has quit IRC | 12:08 | |
| *** goldyfruit has quit IRC | 12:22 | |
| *** cshen has quit IRC | 12:22 | |
| *** electrofelix has joined #openstack-ansible | 12:23 | |
| *** aedc has quit IRC | 12:27 | |
| *** cshen has joined #openstack-ansible | 12:37 | |
| evrardjp | morning CeeMac | 12:37 |
| *** Namrata has quit IRC | 12:42 | |
| *** markvoelker has quit IRC | 12:45 | |
| *** viks___ has quit IRC | 12:46 | |
| *** aedc has joined #openstack-ansible | 13:00 | |
| CeeMac | Afternoon surely :) | 13:10 |
| *** aedc has quit IRC | 13:12 | |
| *** gtarnaras has joined #openstack-ansible | 13:13 | |
| *** cshen has quit IRC | 13:17 | |
| *** luksky11 has quit IRC | 13:22 | |
| *** cshen has joined #openstack-ansible | 13:24 | |
| guilhermesp | mornings! | 13:31 |
| *** BjoernT has joined #openstack-ansible | 13:32 | |
| *** shyam89 has quit IRC | 13:33 | |
| *** cshen has quit IRC | 13:34 | |
| *** irclogbot_3 has joined #openstack-ansible | 13:35 | |
| *** irclogbot_3 has quit IRC | 13:38 | |
| *** goldyfruit has joined #openstack-ansible | 13:49 | |
| *** cshen has joined #openstack-ansible | 13:50 | |
| *** cshen has quit IRC | 13:54 | |
| NobodyCam | Good Morning OSA folks, just wanted to ask a dumb question: | 14:04 |
| NobodyCam | to add a new cinder back end to a existing cloud one would: | 14:06 |
| NobodyCam | modify openstack_user_config.yml adding the new backend info | 14:06 |
| NobodyCam | then run openstack-ansible os-cinder-install.yml ? | 14:06 |
| *** irclogbot_2 has joined #openstack-ansible | 14:09 | |
| *** altlogbot_1 has joined #openstack-ansible | 14:13 | |
| *** ivve has quit IRC | 14:14 | |
| *** chandankumar is now known as raukadah | 14:14 | |
| openstackgerrit | Mohammed Naser proposed openstack/openstack-ansible master: masakari: add missing is_metal field https://review.opendev.org/670568 | 14:15 |
| openstackgerrit | Mohammed Naser proposed openstack/openstack-ansible master: inventory: don't crash with missing global overrides https://review.opendev.org/670569 | 14:16 |
| *** cshen has joined #openstack-ansible | 14:16 | |
| *** spatel has joined #openstack-ansible | 14:17 | |
| mnaser | hmm | 14:20 |
| mnaser | anyone ever ran osa from a non-root user? | 14:20 |
| mnaser | looks like its trying to connect as $USER instead or 'root' | 14:21 |
| *** cshen has quit IRC | 14:38 | |
| *** rpittau is now known as elfosardo | 14:52 | |
| *** markvoelker has joined #openstack-ansible | 14:53 | |
| jrosser | mnaser: have you seen that the utility playbook doesn’t update the sym links to the clip tools correctly? | 14:55 |
| jrosser | clip/cli | 14:55 |
| mnaser | jrosser: this is for updates right? | 14:56 |
| jrosser | Anything that already is linked doesn’t get relinked to point to the binaries in a new venv when we do upgrades | 14:56 |
| mnaser | not fresh deploys | 14:56 |
| *** markvoelker has quit IRC | 14:56 | |
| mnaser | oof | 14:56 |
| mnaser | why is that | 14:56 |
| mnaser | does ansible not relink things | 14:56 |
| jrosser | The task is not smart enough | 14:56 |
| jrosser | It’s using shell/command rather than a more native ansible-ism | 14:57 |
| jrosser | And I could really see a good reason for that | 14:57 |
| jrosser | Couldnt | 14:57 |
| mnaser | jrosser: i think its the black magic around regex stuff with python-whatever | 14:58 |
| mnaser | but yes this seems like something that should be refactored asily | 14:58 |
| mnaser | i think you can just do the clean up in _openstack_client_list | 14:58 |
| *** ansmith has quit IRC | 15:23 | |
| *** cshen has joined #openstack-ansible | 15:44 | |
| goldenfri | Hi everyone. So nova alias requires multiple repeating alias entries for multiple aliases, but OSA will only take the last alias entry. Any ideas how I can work around this? | 15:45 |
| *** cshen has quit IRC | 15:48 | |
| *** gtarnaras has quit IRC | 15:49 | |
| *** ansmith has joined #openstack-ansible | 15:57 | |
| *** adrianreza_ has joined #openstack-ansible | 16:01 | |
| *** gyee has joined #openstack-ansible | 16:01 | |
| evrardjp | jrosser: I am validating this whole chain to be consistent, but else I don't think we should do this. I think we should be using the config file overrides for it. | 16:02 |
| jrosser | evrardjp: is like an example of what you think would acceptable in that case | 16:05 |
| evrardjp | I have started to merge stuff to still be consistent | 16:05 |
| evrardjp | I prefer consistency first | 16:05 |
| * mnaser is trying to get the context for ^ | 16:05 | |
| evrardjp | it's an improvement already to do this | 16:05 |
| evrardjp | I mean we could have in user_variables some kind of override | 16:05 |
| evrardjp | it would mean one per service ofc, but they will probably be all the same, therefore a reuse of yaml anchors could be a win | 16:06 |
| evrardjp | but that's already a good improvement to me | 16:06 |
| jrosser | mnaser: I asked evrardjp to take a look at https://review.opendev.org/#/q/topic:bind-to-mgmt+(status:open+OR+status:merged) | 16:06 |
| evrardjp | mnaser: it's about running services not on 0.0.0.0 | 16:07 |
| jrosser | evrardjp: I see this as not just about the AIO | 16:07 |
| jrosser | I’d like to be binding to the mgmt network by default n all deploys | 16:08 |
| evrardjp | yup I guessed that, because I saw you changed the groupvars | 16:08 |
| mnaser | in terms of security yes, this pretty much values what jrosser needs/wants in terms of network segration | 16:08 |
| mnaser | and also once again, doing a 3 node ctl plane with haproxy+keepalive on metal is broken in our current system | 16:08 |
| mnaser | which imho is a pretty common use case.. 3 controllers, N computes .. it's what tripleo does too | 16:09 |
| evrardjp | yeah it's common | 16:09 |
| jrosser | mnaser: how far off fixing that is my patch set? | 16:09 |
| evrardjp | jrosser: I like the fact we're providing this for everyone in fact | 16:10 |
| evrardjp | and I think your way is probably the most elegant way to provide that for everyone | 16:10 |
| evrardjp | I am maybe too careful in terms of group vars | 16:10 |
| evrardjp | :D | 16:10 |
| evrardjp | I have to run | 16:11 |
| evrardjp | ttyl! | 16:11 |
| mnaser | jrosser: i think its pretty much most of it, i dunno if guilhermesp had a chance to share our configs but thats pretty much mostly it | 16:12 |
| mnaser | the neat thing is just add haproxy and watch it all blow up in flames :P | 16:12 |
| jrosser | mnaser: I think if that goes on .101 as the vip we might get lucky | 16:13 |
| jrosser | Give or take horizon or something | 16:14 |
| jrosser | I need to do a bit more on Galera as some tasks assume 127.0.0.1 right now but that looks fairly ok to fix | 16:15 |
| mnaser | jrosser: deploying an all metal galera cluster as we speak | 16:15 |
| mnaser | ill give the group_vars when done | 16:15 |
| *** elfosardo is now known as rpittau|afk | 16:27 | |
| jrosser | mnaser: i think this is what needs dealing with https://github.com/openstack/openstack-ansible-galera_server/search?q=127.0.0.1&unscoped_q=127.0.0.1 | 16:34 |
| jrosser | we've talked a lot about moving from group vars to config overrides, and it would be good to see a worked example of that | 16:35 |
| jrosser | as you get one chance only to define the overrides for a particular config file, and we need to feed the "default wiring" in as well as anything else the user wants to add | 16:36 |
| openstackgerrit | Mohammed Naser proposed openstack/openstack-ansible-galera_server master: cleanup: refactor into galera_mariadb_service_name https://review.opendev.org/670599 | 16:42 |
| mnaser | jrosser: pushing fixes shortly | 16:52 |
| openstackgerrit | Mohammed Naser proposed openstack/openstack-ansible-galera_server master: setup: don't try to login to create users https://review.opendev.org/670601 | 16:52 |
| openstackgerrit | Mohammed Naser proposed openstack/openstack-ansible-galera_server master: setup: don't try to login to create users and upgrade https://review.opendev.org/670601 | 16:54 |
| guilhermesp | mnaser jrosser evrardjp I think we have an opportunity now to test the bind-br-mgmt | 16:55 |
| guilhermesp | we have a couple of deploy ahead all on metal | 16:55 |
| guilhermesp | I think more things will come up through the process | 16:55 |
| guilhermesp | the vars I have are all like specific for an ipv6 deployment. | 16:55 |
| mnaser | fwiw those patches + http://paste.openstack.org/show/754349/ = a succesful galera deploy here (with bonus of ipv6 too) | 16:56 |
| guilhermesp | that would be useful for us to work in the future btw | 16:56 |
| mnaser | yeah its a lot of |ipwrap-ing for ipv6 | 16:56 |
| guilhermesp | yeah the ipwarp for example | 16:56 |
| mnaser | i just also noticed we create users N times in galera | 16:57 |
| guilhermesp | it is somethign that we can have wihtout affecting ipv4 deplomnts | 16:57 |
| mnaser | ++ | 16:57 |
| jrosser | Can we take a second pass across all this for ipv6? | 16:57 |
| guilhermesp | @mnaser is this using the db-refactor? | 16:57 |
| guilhermesp | no? | 16:57 |
| mnaser | jrosser: i think it would be good after we land this yeah | 16:58 |
| mnaser | guilhermesp: i dont think we applied that logic there | 16:58 |
| guilhermesp | jrosser: I think I can find some time yeah, just need to finish some priorities on our side | 16:58 |
| jrosser | teamwork++ :) | 16:58 |
| * mnaser would be happy to see us run ipv6 only jobs in gate | 16:58 | |
| jrosser | I’d be totally happy with that in my deployments too | 16:59 |
| *** udesale has quit IRC | 16:59 | |
| openstackgerrit | Merged openstack/openstack-ansible-os_ironic master: Fix CentOS installtion https://review.opendev.org/670294 | 17:09 |
| openstackgerrit | Merged openstack/openstack-ansible-os_ironic master: Add integrated coverage https://review.opendev.org/660083 | 17:09 |
| openstackgerrit | Merged openstack/openstack-ansible-rabbitmq_server master: Fix detection of rabbitmq restart https://review.opendev.org/670274 | 17:27 |
| *** nicolasbock has joined #openstack-ansible | 17:29 | |
| jrosser | mnaser: will this https://review.opendev.org/#/c/670599/ leave the service with the old name running at upgrade time? | 17:40 |
| *** poopcat has joined #openstack-ansible | 17:54 | |
| *** kplant has quit IRC | 18:04 | |
| openstackgerrit | Merged openstack/openstack-ansible-os_nova master: Add default settings for nova metadata and service bind addresses https://review.opendev.org/670046 | 18:10 |
| *** electrofelix has quit IRC | 18:18 | |
| mnaser | jrosser: not sure actually.. | 18:20 |
| openstackgerrit | Mohammed Naser proposed openstack/openstack-ansible stable/stein: ceph: stop adding repos https://review.opendev.org/670617 | 18:39 |
| *** irclogbot_2 has quit IRC | 18:49 | |
| *** luksky11 has joined #openstack-ansible | 18:50 | |
| *** irclogbot_1 has joined #openstack-ansible | 18:51 | |
| openstackgerrit | Jonathan Rosser proposed openstack/openstack-ansible master: [WIP] Bind services to mgmt network addresses https://review.opendev.org/670051 | 19:27 |
| *** kplant has joined #openstack-ansible | 20:32 | |
| *** goldyfruit has quit IRC | 20:35 | |
| redkrieg | is there any supported way to ensure openstack service users get the same uid/gid between computes? I've got some computes where nova is 998 and others where it's 999, need them to be the same for network share permissions | 20:44 |
| jrosser | redkrieg: nova_system_user_uid / gid | 20:48 |
| redkrieg | jrosser: thanks much. I've got ceph-mds working btw, I'll probably need some tips on getting a patch together to submit upstream next week | 20:49 |
| redkrieg | now to reinstall all my computes... | 20:49 |
| jrosser | good news on ceph, what are you going to use cephfs for? | 20:49 |
| redkrieg | if you promise not to yell at me... /var/lib/nova/instances/snapshots and /var/lib/nova/instances/_base | 20:50 |
| redkrieg | trust me, I /really/ don't want to be using it for those | 20:50 |
| jrosser | :) | 20:50 |
| spatel | while talking about ceph, is it ok to add different size of OSD in ceph or better to keep same size? | 20:52 |
| redkrieg | I know next to nothing about the inner workings of ceph, I've basically crash-coursed getting it set up this week | 20:54 |
| kplant | redkrieg: you can use ceph as a backend for nova without using cephfs | 20:55 |
| kplant | if that's what you're looking for | 20:55 |
| redkrieg | I know, I'm using it for glance and cinder, using local storage for nova instance disks, but using cephfs for _base and snapshots because they can be very large and I don't have the local storage on every node for them (for instance, if my node is at 90% capacity and someone who takes up 25% of it wants to take a snapshot) | 20:56 |
| kplant | ah, gottcha | 20:57 |
| kplant | i didn't want to deal with that so i just force everyone to use cinder volumes | 20:57 |
| kplant | lazy :] | 20:57 |
| *** pcaruana has quit IRC | 20:59 | |
| redkrieg | my clients will be more performance hungry than cinder volumes can support for me :) | 20:59 |
| spatel | redkrieg: you are saying cephfs provide more performance compare to cinder volumes? | 21:00 |
| *** TxGirlGeek has joined #openstack-ansible | 21:00 | |
| kplant | i think redkrieg is referring to the nova disks for the instances vs. cinder volumes | 21:00 |
| kplant | on drives local to the compute nodes | 21:00 |
| redkrieg | yeah, cephfs is being used for image cache and temporary snapshot data in my setup | 21:01 |
| kplant | redkrieg: i'll probably run into that soon and need to start delivering compute instances with local nvme drives :-| | 21:01 |
| kplant | for now they're more than happy with our ceph cluster which is all nvme | 21:01 |
| kplant | but ceph doesn't have the best latency | 21:01 |
| redkrieg | you're probably going to want my ceph-mds patch then :D | 21:01 |
| kplant | oh? | 21:04 |
| redkrieg | if you plan to offload some things to cephfs, yeah | 21:04 |
| kplant | oh, no | 21:05 |
| kplant | i'll just say "these are you local disks with no redundancy, have fun" | 21:05 |
| kplant | no snapshots | 21:05 |
| redkrieg | hah, that works too | 21:05 |
| kplant | if they care about their data it should be on a cinder volume | 21:05 |
| *** weifan has joined #openstack-ansible | 21:07 | |
| *** weifan has quit IRC | 21:18 | |
| mnaser | fyi ceph for image cache might be problematic because i think with nova the image cache isnt actually a cache but actually where the base images are stored if you're using qcow2 | 21:22 |
| *** weifan has joined #openstack-ansible | 21:22 | |
| jrosser | mnaser: less success now we bind galera address https://review.opendev.org/#/c/670051/ | 21:22 |
| jrosser | the jobs that are working are happy accident because of this https://github.com/openstack/openstack-ansible/blob/dc5729ad6f10b7e6083d60c2270f75090cf3d5f4/inventory/group_vars/all/infra.yml#L35 | 21:23 |
| mnaser | jrosser: nice! | 21:23 |
| mnaser | and oh | 21:24 |
| redkrieg | mnaser: we configured it so _base is only temporarily needed during instance creation. it creates a copy of the image for nova to actually use as the disk. | 21:24 |
| mnaser | yeah that'll do it lol | 21:24 |
| mnaser | oh so you're not doing cow | 21:24 |
| jrosser | i think they fail becasue the mysql_db ansible module fails as it defaults to 'localhost', so the db_setup fails | 21:24 |
| mnaser | well im seeing rabbit fail to start up, not galera jrosser | 21:24 |
| mnaser | unless i looked at the wrong job | 21:24 |
| jrosser | probably multiple errors | 21:24 |
| jrosser | i think i need to nuke my local AIO and start again | 21:25 |
| redkrieg | mnaser: yeah that was it. use_cow_images: False | 21:25 |
| redkrieg | takes about 7 seconds to spin up an ubuntu server instance with this setup, so not too bad | 21:25 |
| jrosser | mnaser: this not too happy either http://logs.openstack.org/51/670051/12/check/openstack-ansible-deploy-aio_metal-opensuse-15/9824a47/logs/openstack/aio1-mysql_logs/galera_server_error.log.txt.gz | 21:26 |
| mnaser | weird | 21:30 |
| jrosser | slightly annoyingly the metal-bionic job passes | 21:30 |
| jrosser | and i'm not sure why there is a distro dependance here | 21:30 |
| *** spatel has quit IRC | 21:35 | |
| *** weifan has quit IRC | 21:38 | |
| *** weifan has joined #openstack-ansible | 21:40 | |
| *** weifan has quit IRC | 21:41 | |
| *** weifan has joined #openstack-ansible | 21:42 | |
| *** weifan has quit IRC | 21:47 | |
| *** BjoernT has quit IRC | 21:47 | |
| *** luksky11 has quit IRC | 21:48 | |
| *** ianychoi has quit IRC | 21:54 | |
| *** KeithMnemonic has quit IRC | 22:02 | |
| openstackgerrit | Merged openstack/ansible-config_template master: Add option to use a remote source https://review.opendev.org/669758 | 22:32 |
| openstackgerrit | Mohammed Naser proposed openstack/openstack-ansible-tests master: sync: add db_setup.yml https://review.opendev.org/670672 | 22:43 |
| mnaser | guilhermesp: when you have some spare time, if you can do fix up the 4 roles so we can finish up that work :D | 22:46 |
| openstackgerrit | Merged openstack/openstack-ansible-os_neutron master: Add neutron_api_bind_address default variable https://review.opendev.org/670285 | 22:52 |
| openstackgerrit | Merged openstack/openstack-ansible-openstack_hosts stable/queens: use ansible_hostname to construct canonical_hostname in /etc/hosts https://review.opendev.org/621688 | 23:17 |
| *** tosky has quit IRC | 23:20 | |
| *** nicolasbock has quit IRC | 23:22 | |
| *** weifan has joined #openstack-ansible | 23:23 | |
| *** weifan has quit IRC | 23:28 | |
Generated by irclog2html.py 2.15.3 by Marius Gedminas - find it at mg.pov.lt!