| *** gregoryo has joined #openstack-ansible | 01:19 | |
| openstackgerrit | lvxianguo proposed openstack/openstack-ansible-os_tacker master: Add source code repository notes link to README https://review.openstack.org/591264 | 01:36 |
|---|---|---|
| *** shyamb has joined #openstack-ansible | 01:37 | |
| cloudnull | Tahvok do you want to set it in the container or on the host ? | 02:19 |
| *** shyamb has quit IRC | 02:21 | |
| openstackgerrit | Kevin Carter (cloudnull) proposed openstack/openstack-ansible-ops master: Move most of the variables into the roles https://review.openstack.org/591218 | 02:50 |
| *** vnogin has joined #openstack-ansible | 02:52 | |
| *** vnogin has quit IRC | 02:56 | |
| *** chandankumar has quit IRC | 03:05 | |
| openstackgerrit | Kevin Carter (cloudnull) proposed openstack/openstack-ansible-ops master: Move most of the variables into the roles https://review.openstack.org/591218 | 03:15 |
| openstackgerrit | Kevin Carter (cloudnull) proposed openstack/openstack-ansible-ops master: move the bulk to templates into the new roles https://review.openstack.org/591217 | 03:19 |
| openstackgerrit | Kevin Carter (cloudnull) proposed openstack/openstack-ansible-ops master: Move most of the variables into the roles https://review.openstack.org/591218 | 03:20 |
| *** udesale has joined #openstack-ansible | 03:57 | |
| *** mmalchuk has quit IRC | 04:27 | |
| *** dave-mccowan has quit IRC | 04:28 | |
| openstackgerrit | Merged openstack/openstack-ansible-ops master: move the bulk to templates into the new roles https://review.openstack.org/591217 | 04:34 |
| openstackgerrit | Merged openstack/openstack-ansible-ops master: Move most of the variables into the roles https://review.openstack.org/591218 | 04:34 |
| *** viks_ has joined #openstack-ansible | 04:47 | |
| openstackgerrit | Kevin Carter (cloudnull) proposed openstack/openstack-ansible-ops master: Add SUSE support https://review.openstack.org/591269 | 04:49 |
| *** Bhujay has joined #openstack-ansible | 04:50 | |
| openstackgerrit | Kevin Carter (cloudnull) proposed openstack/openstack-ansible-ops master: Add SUSE support https://review.openstack.org/591269 | 04:53 |
| *** shyamb has joined #openstack-ansible | 05:01 | |
| *** dariko has joined #openstack-ansible | 05:11 | |
| *** shyamb has quit IRC | 05:12 | |
| *** shyamb has joined #openstack-ansible | 05:12 | |
| *** cshen has joined #openstack-ansible | 05:16 | |
| *** openstackgerrit has quit IRC | 05:18 | |
| *** cshen has quit IRC | 05:21 | |
| *** mmalchuk has joined #openstack-ansible | 05:21 | |
| *** chandankumar has joined #openstack-ansible | 05:41 | |
| *** hwoarang has joined #openstack-ansible | 05:53 | |
| *** pcaruana has joined #openstack-ansible | 05:59 | |
| *** shyamb has quit IRC | 06:03 | |
| *** pcaruana has quit IRC | 06:05 | |
| *** openstackgerrit has joined #openstack-ansible | 06:06 | |
| openstackgerrit | Merged openstack/openstack-ansible-ops master: Add SUSE support https://review.openstack.org/591269 | 06:06 |
| *** pcaruana has joined #openstack-ansible | 06:19 | |
| *** udesale has quit IRC | 06:25 | |
| *** dariko has quit IRC | 06:36 | |
| *** shyamb has joined #openstack-ansible | 06:40 | |
| *** shyamb has quit IRC | 06:47 | |
| *** shyamb has joined #openstack-ansible | 06:47 | |
| *** cshen has joined #openstack-ansible | 06:48 | |
| *** vnogin has joined #openstack-ansible | 06:53 | |
| *** ssbarnea has joined #openstack-ansible | 06:55 | |
| *** vnogin has quit IRC | 06:57 | |
| *** udesale has joined #openstack-ansible | 07:02 | |
| openstackgerrit | Merged openstack/openstack-ansible-nspawn_hosts stable/rocky: Document primary interfaces and add mv addresses https://review.openstack.org/591208 | 07:05 |
| *** evrardjp has joined #openstack-ansible | 07:08 | |
| *** shyamb has quit IRC | 07:24 | |
| jrosser | morning | 07:34 |
| openstackgerrit | Jonathan Rosser proposed openstack/openstack-ansible master: AIO - Improve robustness of network configuration. https://review.openstack.org/590424 | 07:34 |
| *** gillesMo has joined #openstack-ansible | 07:36 | |
| *** shyamb has joined #openstack-ansible | 07:37 | |
| *** rpittau has joined #openstack-ansible | 07:43 | |
| *** pcaruana has quit IRC | 07:43 | |
| *** olivierbourdon38 has joined #openstack-ansible | 07:44 | |
| openstackgerrit | zhulingjie proposed openstack/openstack-ansible-os_cloudkitty master: Make the owner vars configurable from default/main.yml https://review.openstack.org/591285 | 07:49 |
| evrardjp | morning | 07:51 |
| evrardjp | sorry for the late answer : ) | 07:51 |
| *** shyamb has quit IRC | 07:52 | |
| *** pcaruana has joined #openstack-ansible | 07:57 | |
| openstackgerrit | Jonathan Rosser proposed openstack/openstack-ansible master: Add test for Ubuntu Bionic https://review.openstack.org/591287 | 07:57 |
| *** dante-as has joined #openstack-ansible | 08:02 | |
| dante-as | Hi guys, is any way to have active Ansible Tower and Openstack Dashboard on the same localhost? | 08:03 |
| *** ppetit has joined #openstack-ansible | 08:04 | |
| openstackgerrit | zhulingjie proposed openstack/openstack-ansible master: Add the "x" to avoid "unary operator expected" https://review.openstack.org/591290 | 08:11 |
| evrardjp | dante-as: we generally have containers preventing this kind of mess :) | 08:11 |
| evrardjp | by default the openstack dashboard is in a container behind a LB, so you could have your tower on your localhost | 08:12 |
| *** DanyC has joined #openstack-ansible | 08:12 | |
| dante-as | evrardjp: I am not so familiar with Docker or any other containers and wanted to check if is any workaround to configure them in this way since I am running out of time | 08:16 |
| openstackgerrit | Maxime Guyot proposed openstack/openstack-ansible-ops master: Add Prometheus node_exporter https://review.openstack.org/587090 | 08:17 |
| dante-as | any link or suggestion would be helpful | 08:21 |
| evrardjp | dante-as: we are not using docker | 08:29 |
| evrardjp | dante-as: if you're willing to do something different, please spend the time for it -- there is no magic, no silver bullet | 08:30 |
| *** hamzaachi has joined #openstack-ansible | 08:33 | |
| openstackgerrit | Merged openstack/openstack-ansible-os_ironic master: use include_tasks instead of include https://review.openstack.org/585273 | 08:33 |
| dante-as | evrardjp: until now, I have tried some stuff: adding the new port for Openstack in the Puppets module ports file config, I have added a new host port, changed the Horizon port, I have added the port in the iptables, but since then, I have broke also the httpd | 08:35 |
| dante-as | Basically, wanted to have a direction if this way doesn't work | 08:36 |
| Tahvok | cloudnull: in container.. OSA doesn't control the hosts interface as far as I know? | 08:39 |
| *** rodolof has joined #openstack-ansible | 08:40 | |
| openstackgerrit | Jonathan Rosser proposed openstack/openstack-ansible master: Change AIO ZFS pool to be at /pool https://review.openstack.org/591304 | 08:40 |
| *** mmalchuk_ has joined #openstack-ansible | 08:41 | |
| Tahvok | dante-as: are you using openstack-ansible? Because it doesn't sound like it... | 08:41 |
| *** mmalchuk has quit IRC | 08:42 | |
| dante-as | Tahvok: I am using Ansible Tower and Packstack Openstack | 08:43 |
| dante-as | Both deployed on my laptop on a VM | 08:44 |
| *** shyamb has joined #openstack-ansible | 08:44 | |
| evrardjp | dante-as: then it's not for our channel :) | 08:46 |
| evrardjp | here it's to deploy with openstack-ansible | 08:46 |
| evrardjp | for those heading to the PTG: https://etherpad.openstack.org/p/osa-stein-ptg | 08:48 |
| evrardjp | mnaser: we probably need someone to head to ops meetup on Monday-Tuesday of the PTG - Do you know who will go there? I am busy with other things already. | 08:51 |
| dante-as | evrardjp: Thanks. Can you please suggest me which would be the channel that I would have to address to it? | 08:58 |
| *** shyamb has quit IRC | 09:00 | |
| *** gregoryo has quit IRC | 09:06 | |
| *** shyamb has joined #openstack-ansible | 09:12 | |
| *** shyamb has quit IRC | 09:14 | |
| *** shyamb has joined #openstack-ansible | 09:14 | |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible-os_masakari master: Update reno for stable/rocky https://review.openstack.org/590939 | 09:22 |
| *** noonedeadpunk has joined #openstack-ansible | 09:25 | |
| *** noonedeadpunk has quit IRC | 09:29 | |
| *** electrofelix has joined #openstack-ansible | 09:33 | |
| *** noonedeadpunk has joined #openstack-ansible | 09:34 | |
| *** rpittau has quit IRC | 09:36 | |
| *** rpittau has joined #openstack-ansible | 09:36 | |
| openstackgerrit | Merged openstack/openstack-ansible-os_masakari master: Update reno for stable/rocky https://review.openstack.org/590939 | 09:46 |
| *** tosky has joined #openstack-ansible | 09:59 | |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible-openstack_hosts stable/rocky: Remove pip_install role execution https://review.openstack.org/591327 | 10:04 |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible-openstack_hosts stable/rocky: Remove pip_install role execution https://review.openstack.org/591327 | 10:05 |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible-repo_build stable/rocky: Simplify pip/setuptools/wheel prep for venvs https://review.openstack.org/591328 | 10:06 |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible-repo_build stable/rocky: Simplify pip/setuptools/wheel prep for venvs https://review.openstack.org/591328 | 10:06 |
| *** shyamb has quit IRC | 10:07 | |
| *** shyamb has joined #openstack-ansible | 10:09 | |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible-openstack_hosts stable/rocky: Remove pip from packages installed https://review.openstack.org/591329 | 10:09 |
| openstackgerrit | Dmitriy Rabotjagov (noonedeadpunk) proposed openstack/openstack-ansible-os_gnocchi stable/rocky: Added drivers for coordination. https://review.openstack.org/591330 | 10:09 |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible stable/rocky: Eliminate installing pip on host/containers https://review.openstack.org/591331 | 10:10 |
| openstackgerrit | Dmitriy Rabotjagov (noonedeadpunk) proposed openstack/openstack-ansible-os_gnocchi stable/queens: Added drivers for coordination. https://review.openstack.org/591333 | 10:12 |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible stable/rocky: Eliminate installing pip on host/containers https://review.openstack.org/591331 | 10:14 |
| *** tosky has quit IRC | 10:19 | |
| openstackgerrit | Dmitriy Rabotjagov (noonedeadpunk) proposed openstack/openstack-ansible-os_ceilometer stable/rocky: Added drivers for coordination. https://review.openstack.org/591335 | 10:19 |
| openstackgerrit | Dmitriy Rabotjagov (noonedeadpunk) proposed openstack/openstack-ansible-os_horizon stable/rocky: Implements custom theme distribution https://review.openstack.org/591336 | 10:20 |
| odyssey4me | noonedeadpunk: you're doing backports before the master patch has merged - please do not do that | 10:20 |
| noonedeadpunk | odyssey4me: oh, sorry. Should I abbandon patches? | 10:22 |
| noonedeadpunk | or just workflow -1? | 10:22 |
| odyssey4me | noonedeadpunk: for now, best yes otherwise they create confusion - you can restore them once the master patch has merged, then refresh the cherry-pick | 10:22 |
| noonedeadpunk | no problem, sry | 10:23 |
| noonedeadpunk | odyssey4me: btw, can you review the master patch https://review.openstack.org/#/c/590815/ ? | 10:27 |
| openstackgerrit | Dmitriy Rabotjagov (noonedeadpunk) proposed openstack/openstack-ansible-lxc_container_create stable/rocky: Implements possibility to use physical networks in containers. https://review.openstack.org/591338 | 10:30 |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible master: Disable verbose output of log collection https://review.openstack.org/591339 | 10:33 |
| noonedeadpunk | Or whom I may ask to review except you and cloudnull (as he has already set + on it)? | 10:34 |
| *** shyamb has quit IRC | 10:36 | |
| odyssey4me | noonedeadpunk: all cores scan the review list when possible - be patient, as there's a lot going on right now | 10:36 |
| openstackgerrit | wu.chunyang proposed openstack/openstack-ansible-os_swift master: Add the project source code repository in README https://review.openstack.org/591340 | 10:36 |
| odyssey4me | if you've not got attention on a patch, then the tuesday meeting may be a good tie to raise it - you can add it to the agenda | 10:36 |
| odyssey4me | very often people don't review because they don't understand the purpose of the patch - and sometimes it's just because they didn't see it | 10:37 |
| jrosser | noonedeadpunk: i looked and didn't understand - it seems to install pip packages (drivers?) for things we don't deploy | 10:39 |
| odyssey4me | jrosser: ah yes, I suggested that either noonedeadpunk add the drivers to the existing list, or he add an optional set of pip packages which is installed after the venv build | 10:42 |
| noonedeadpunk | jrosser: yeh, as for now zookeeper role is not really functional (and it's not even working now). However, it's stated in documentation, that user is supposed to install coordination provider on his own. | 10:42 |
| odyssey4me | the fact that we don't have a role/playbook to implement redis/zookeeper is kinda immaterial | 10:42 |
| openstackgerrit | wu.chunyang proposed openstack/openstack-ansible-os_tempest master: Add the project source code repository in README https://review.openstack.org/591343 | 10:42 |
| noonedeadpunk | odyssey4me actually it's present, and I even tried to make it working https://review.openstack.org/#/q/project:openstack/ansible-role-zookeeper | 10:43 |
| odyssey4me | it sort of comes down to whether we'd prefer the venv to be leaner (only implement what's required to run the basic service) or not (add whatever is usable and won't conflict) | 10:43 |
| odyssey4me | noonedeadpunk: that's not a role we maintain, and not a role I'd personally advise being used either | 10:43 |
| odyssey4me | there are far better roles available outside the openstack namespace which do zookeeper | 10:44 |
| jrosser | it would probably be good to do a spec for a coordination service, beacuse designate really could use one too | 10:45 |
| odyssey4me | oh really? | 10:45 |
| noonedeadpunk | odyssey4me: yep, I already got it. But after release of rocky are you going to implement some coordinator inside of OSA project? As it seems pretty reasonable to get ceilometer distribute load among workers. | 10:45 |
| odyssey4me | I didn't even know this was a thing. | 10:46 |
| jrosser | odyssey4me: i think so, its optional, let me try to dig that out | 10:46 |
| odyssey4me | But if this is an emerging thing that would be useful, then yes - adding a spec for it would be great. | 10:46 |
| odyssey4me | If not something built-in, perhaps a user story for it. | 10:46 |
| openstackgerrit | wu.chunyang proposed openstack/openstack-ansible-os_watcher master: Add the project source code repository in README https://review.openstack.org/591344 | 10:47 |
| odyssey4me | But if this is useful across services, then perhaps built-in is best. | 10:47 |
| odyssey4me | jrosser: if this is isolated to ceilometer/gnocchi then I don't really see the harm in including the drivers in the venv for now to enable people to use the feature | 10:48 |
| odyssey4me | I would prefer to see them as optional packages, but that's probably too much overhead for these services which aren't heavily used in this community. | 10:49 |
| jrosser | odyssey4me: heres the designate bit https://docs.openstack.org/designate/latest/admin/ha.html | 10:52 |
| jrosser | i think without a coordinator/dlm designate-producer might do more work than is necessary | 10:52 |
| jrosser | but it'll work | 10:52 |
| odyssey4me | ah, interesting | 10:53 |
| *** cshen has quit IRC | 10:53 | |
| odyssey4me | I think the simplest thing for us to do right now is to just enable this to be used more easily | 10:53 |
| odyssey4me | there are several different approaches to take | 10:53 |
| odyssey4me | one is to enable the use of an optional set of packages, and another is to add them into the venv | 10:54 |
| openstackgerrit | wu.chunyang proposed openstack/openstack-ansible-rsyslog_server master: Add the project source code repository in README https://review.openstack.org/591348 | 10:55 |
| odyssey4me | meh, this has limited impact - and if anyone feels strongly about making them optional, they can patch it | 10:57 |
| *** shyamb has joined #openstack-ansible | 11:00 | |
| openstackgerrit | wu.chunyang proposed openstack/openstack-ansible-rsyslog_client master: Add the project source code repository in README https://review.openstack.org/591349 | 11:00 |
| openstackgerrit | wu.chunyang proposed openstack/openstack-ansible-repo_server master: Add the project source code repository in README https://review.openstack.org/591350 | 11:03 |
| *** cshen has joined #openstack-ansible | 11:05 | |
| openstackgerrit | wu.chunyang proposed openstack/openstack-ansible-repo_build master: Add the project source code repository in README https://review.openstack.org/591351 | 11:05 |
| *** vnogin has joined #openstack-ansible | 11:10 | |
| *** shyamb has quit IRC | 11:12 | |
| *** shyamb has joined #openstack-ansible | 11:12 | |
| *** jiteka has left #openstack-ansible | 11:15 | |
| noonedeadpunk | odyssey4me: I'm not sure what is the best way to illustrate path, where some archive is placed. https://review.openstack.org/#/c/586318/4/defaults/main.yml. So the idea is, that as user may have several themes, he provides path to directory, where he puts described in horizon_custom_themes themes. So, if theme_name is "custom", and it's placed in /opt/horizon/themes/custom.zip, than horizon_custom_themes_path: /opt/horizon/ | 11:25 |
| noonedeadpunk | themes/ | 11:25 |
| noonedeadpunk | probably I should write description in more obvious way... | 11:25 |
| evrardjp | are we okay to backport https://review.openstack.org/#/c/590778/ if it isn't merged already? | 11:26 |
| evrardjp | So that it unlocks stable/rocky | 11:26 |
| odyssey4me | noonedeadpunk: the defaults are documentation, so in the comments I've asked to provide a commented example so that people know how to structure the data | 11:26 |
| odyssey4me | evrardjp: yes, I think so | 11:26 |
| openstackgerrit | Jean-Philippe Evrard proposed openstack/openstack-ansible stable/rocky: Temporarily use tempest master for all builds https://review.openstack.org/591357 | 11:26 |
| evrardjp | else there is a risk of drift | 11:26 |
| odyssey4me | yep | 11:26 |
| openstackgerrit | Jean-Philippe Evrard proposed openstack/openstack-ansible stable/rocky: Freeze all SHAs for RC1 https://review.openstack.org/590503 | 11:27 |
| noonedeadpunk | odyssey4me: yeah, I know about docs. I'll try to write it in another words then. | 11:27 |
| *** udesale has quit IRC | 11:27 | |
| odyssey4me | evrardjp: oh the extra indexes were removed? | 11:28 |
| odyssey4me | the fragmanet I mean | 11:28 |
| openstackgerrit | wu.chunyang proposed openstack/openstack-ansible-os_almanach master: Add the project source code repository in README https://review.openstack.org/591358 | 11:28 |
| odyssey4me | bah *fragments | 11:28 |
| evrardjp | yeah it didn'twork | 11:29 |
| evrardjp | with the 2 fragments, it was not using the right version, with one fragment.. I can't remember but it wasn't working either wrong version or something | 11:30 |
| openstackgerrit | wu.chunyang proposed openstack/openstack-ansible-os_blazar master: fix wrong link https://review.openstack.org/591360 | 11:30 |
| evrardjp | odyssey4me: if that's not okay, I can define developer_mode: True by default in tempest role. | 11:30 |
| openstackgerrit | Merged openstack/openstack-ansible-os_gnocchi master: Added drivers for coordination. https://review.openstack.org/590815 | 11:30 |
| evrardjp | so we don't do this kind of things | 11:31 |
| evrardjp | oh yeah | 11:33 |
| evrardjp | "No matching distribution found for tempest==0.0.0" | 11:33 |
| evrardjp | too many cpu cycles for nothing | 11:33 |
| openstackgerrit | wu.chunyang proposed openstack/openstack-ansible-os_rally master: Add the project source code repository in README https://review.openstack.org/591361 | 11:35 |
| odyssey4me | evrardjp: the trouble with that is that the py_pkgs plugin will pick it up - so we need to revert it once the right version of tempest is published | 11:35 |
| evrardjp | odyssey4me: yeah I am tracking this | 11:36 |
| odyssey4me | oh wait - developer mode true... hmm, yeah then the tempest_git_repo var also has to be set by default | 11:36 |
| odyssey4me | but yes, that means that once the right tempest is published we must revert both settings | 11:36 |
| evrardjp | for now I'd say let's leave it like this | 11:36 |
| odyssey4me | that will mean that we can do without the integrated repo thing | 11:36 |
| evrardjp | I am tracking tempest releasing anyway | 11:36 |
| openstackgerrit | wu.chunyang proposed openstack/openstack-ansible-os_sahara master: Add the project source code repository in README https://review.openstack.org/591364 | 11:37 |
| evrardjp | I'd rather stick with what's there as this was closer to our "traditional way" | 11:37 |
| evrardjp | and it's just because requirements + tempest isn't released yet | 11:37 |
| evrardjp | requirements isn't branched and tempest isn't released | 11:37 |
| evrardjp | so I can keep that up in memory | 11:37 |
| evrardjp | odyssey4me: I think I had a patch to modify tempest 's default behavior that I abandonned. I can double check | 11:38 |
| *** tosky has joined #openstack-ansible | 11:41 | |
| *** ppetit has quit IRC | 11:43 | |
| openstackgerrit | Dmitriy Rabotjagov (noonedeadpunk) proposed openstack/openstack-ansible-os_horizon master: Implements custom theme distribution https://review.openstack.org/586318 | 11:44 |
| noonedeadpunk | odyssey4me: https://review.openstack.org/#/c/586318/5/defaults/main.yml is description okey for you? | 11:47 |
| odyssey4me | noonedeadpunk: It still wasn't clear to me, then I reliased there was a nother var involved - looking | 11:48 |
| mmalchuk_ | hello everyone | 11:49 |
| noonedeadpunk | hi | 11:49 |
| mmalchuk_ | odyssey4me could you please review https://review.openstack.org/#/q/status:open+project:openstack/openstack-ansible+branch:master+topic:air-gapped-env-deploy | 11:49 |
| openstackgerrit | Merged openstack/openstack-ansible-os_gnocchi master: Bring gather vars in line with other OSA roles https://review.openstack.org/586685 | 11:50 |
| *** dave-mccowan has joined #openstack-ansible | 11:52 | |
| openstackgerrit | James Denton proposed openstack/openstack-ansible master: Adds haproxy endpoints for networking-ovn support https://review.openstack.org/584069 | 11:54 |
| jamesdenton | mbuil thanks for the reviews! | 11:55 |
| *** shyamb has quit IRC | 12:03 | |
| mbuil | jamesdenton: you are welcome! Your patch was almost perfect from the beginning. Thanks for the contribution :) | 12:04 |
| jamesdenton | Don't let evrardjp see that. He will expect more from me | 12:06 |
| evrardjp | haha | 12:06 |
| evrardjp | alway! | 12:06 |
| evrardjp | always* | 12:07 |
| *** rpittau is now known as rpittau|afk | 12:10 | |
| *** rpittau|afk is now known as rpittau | 12:10 | |
| *** shyamb has joined #openstack-ansible | 12:11 | |
| openstackgerrit | Jonathan Rosser proposed openstack/openstack-ansible-repo_build master: Generate a filtered list of git repos to clone https://review.openstack.org/591368 | 12:14 |
| noonedeadpunk | evrardjp: it's pretty tricky to make just list of file extensions, as ve cannot normally iterate over them (as we've already have with_dict for task) | 12:15 |
| *** shyamb has quit IRC | 12:15 | |
| evrardjp | noonedeadpunk: not sure what you mean there? | 12:16 |
| noonedeadpunk | evrardjp your comment here https://review.openstack.org/#/c/586318/5/tasks/horizon_post_install.yml | 12:17 |
| evrardjp | with_ is merely calling a lookup | 12:17 |
| evrardjp | noonedeadpunk: which one? | 12:17 |
| evrardjp | noonedeadpunk: set_fact of the file to deploy | 12:18 |
| evrardjp | files* | 12:18 |
| evrardjp | or find: the files to deploy | 12:18 |
| evrardjp | then copy and wire them | 12:18 |
| evrardjp | (or unarchive them) | 12:18 |
| noonedeadpunk | evrardjp: ha, so you mean to add another task, where we will just register file, which was found? | 12:19 |
| mnaser | good morning everyone | 12:19 |
| odyssey4me | noonedeadpunk: why not just use with_fileglob ? | 12:19 |
| evrardjp | with fileglob is okay for that case. | 12:21 |
| noonedeadpunk | odyssey4me evrardjp and list of extensoins in regexp? | 12:22 |
| evrardjp | as I said: You want readability, you should probably have a set_fact. If not, you can do it directly in the task, but needs to be done cleverly :) | 12:22 |
| *** ppetit has joined #openstack-ansible | 12:22 | |
| mnaser | friendly reminder: https://etherpad.openstack.org/p/osa-stein-ptg :) | 12:22 |
| evrardjp | noonedeadpunk: that's not regexp, fileglob is just regular python glob | 12:23 |
| *** dante-as has left #openstack-ansible | 12:23 | |
| evrardjp | first found is not a gread idea, because there might be more than a file or zero | 12:23 |
| evrardjp | find on the other side, is giving you all the features you need | 12:24 |
| evrardjp | just run on localhost | 12:24 |
| kaiokmo | morning | 12:27 |
| noonedeadpunk | evrardjp: yep, it's just not so specific. isn't it better to do nothing, that to unarchive some file, which is not supposed to be unarchived? Of course we may rely on the user, but isn't it worth to check, that he didn't put anything strange into the folder? | 12:29 |
| evrardjp | noonedeadpunk: deployers have to read the documentation/release notes. | 12:33 |
| evrardjp | or at least should | 12:33 |
| evrardjp | I guess it's a minimum -- else they can audit the code | 12:34 |
| evrardjp | it's fairly readable | 12:34 |
| *** tosky has quit IRC | 12:35 | |
| noonedeadpunk | evrardjp: ok, probably you're right (but not every deployer would read documentation, actually). And how do you mean to improve notify? Yes, it's kind of horizon_custom_uploads, but into different path, and with, probably, much more files to upload, so I'm not sure about merging it into horizon_custom_uploads | 12:38 |
| evrardjp | I think it's fine to have a generic way | 12:39 |
| evrardjp | and tell users they should do "x" now | 12:39 |
| evrardjp | talking about Uploading horizon custom files right? | 12:39 |
| evrardjp | same for Enable project panels | 12:39 |
| evrardjp | that's quite silly now | 12:39 |
| evrardjp | I did that for clarity, vs what we had before | 12:40 |
| evrardjp | but I just checked the code in the plugins | 12:40 |
| *** fghaas has joined #openstack-ansible | 12:40 | |
| evrardjp | it's kinda always the same behavior -- that can be replaced by a find: and a file: task | 12:40 |
| evrardjp | we should do this | 12:40 |
| evrardjp | for your case,you can simplify the readability of the tasks | 12:41 |
| evrardjp | we can later improve things | 12:41 |
| evrardjp | (later means: in a separate commit, but it would be nice to be done earlier rather than later) | 12:41 |
| noonedeadpunk | evrardjp: ok, got it | 12:41 |
| evrardjp | I can definitely see https://github.com/openstack/openstack-ansible-os_horizon/blob/master/vars/main.yml#L16-L109 go away soon, and I would be happy about it. | 12:43 |
| evrardjp | :D | 12:43 |
| noonedeadpunk | evrardjp: yep, it would be nice) | 12:45 |
| kaiokmo | evrardjp: about that thing we discussed on friday: https://review.openstack.org/#/c/591101/ :D | 12:47 |
| *** nicolasbock has joined #openstack-ansible | 12:48 | |
| *** hamzaachi has quit IRC | 12:52 | |
| evrardjp | kaiokmo: I agree with your patch, but I'd like a discussion with cloudnull about systemd_service if we do it that way. | 12:54 |
| evrardjp | see my comment | 12:55 |
| evrardjp | else this patch is actively increasing technical debt and makes things harder to guess | 12:55 |
| evrardjp | but I am globally positive | 12:55 |
| evrardjp | thanks for the reminder kaiokmo | 12:55 |
| *** hwoarang has quit IRC | 12:56 | |
| odyssey4me | evrardjp kaiokmo yeah, agreed - rather use the built-in facility than ose the config override | 12:58 |
| odyssey4me | *use | 12:58 |
| *** hwoarang has joined #openstack-ansible | 12:59 | |
| *** jroll has quit IRC | 12:59 | |
| openstackgerrit | Andy Smith proposed openstack/openstack-ansible-os_nova master: Setup oslo.messaging extra packages for optional drivers https://review.openstack.org/584450 | 12:59 |
| *** jroll has joined #openstack-ansible | 13:00 | |
| openstackgerrit | Andy Smith proposed openstack/openstack-ansible-os_cinder master: Setup oslo.messaging extra packages for optional drivers https://review.openstack.org/582632 | 13:02 |
| evrardjp | odyssey4me: I am fine with the config override IF we remove the default var | 13:04 |
| kaiokmo | evrardjp: no problem | 13:04 |
| evrardjp | I wonder how we can make this easy though | 13:04 |
| evrardjp | because we won't be able to use |default(omit) | 13:04 |
| hwoarang | andymccr: hello. do you have a second to talk about ceph? :) | 13:04 |
| evrardjp | hwoarang: I am not sure we'll see andymccr talk about ceph soon. | 13:04 |
| andymccr | hey sorry just got on :) what's up with Ceph? | 13:05 |
| odyssey4me | hwoarang: andymccr has moved on to another company, doing other work outside of openstack - I don't know if he'll be around to discuss anything OSA any more | 13:05 |
| hwoarang | oh | 13:05 |
| hwoarang | :/ | 13:05 |
| odyssey4me | lol, and there he is | 13:05 |
| evrardjp | andymccr: maybe you should announce your departure, else ppl will ping you with many things ceph : ) | 13:06 |
| evrardjp | but you're always welcome | 13:06 |
| hwoarang | well andymccr congratulations then and I hope all the best in your new role ): | 13:06 |
| hwoarang | *:) | 13:06 |
| andymccr | thanks hwoarang ! im gonna try keep contributing esp around those bits - so lets see how it goes | 13:06 |
| andymccr | anyway whats up with Ceph? | 13:06 |
| hwoarang | quick question then. fatal: [aio1_ceph-rgw_container-8f1c8d43]: FAILED! => {"changed": false, "cmd": "ceph --cluster ceph -s -f json", "msg": "[Errno 2] No such file or directory", "rc": 2} | 13:06 |
| odyssey4me | thanks andymccr - appreciate it :) | 13:07 |
| hwoarang | i think prob is that we call ceph-defaults here https://github.com/openstack/openstack-ansible/blob/master/playbooks/ceph-rgw-install.yml#L45 | 13:07 |
| hwoarang | but the container seem to be missing package installation steps? | 13:07 |
| *** hamzaachi has joined #openstack-ansible | 13:07 | |
| hwoarang | so 'ceph' command is not available | 13:07 |
| hwoarang | at least i am hitting that on suse... | 13:07 |
| odyssey4me | hwoarang andymccr also FYI jrosser makes use of ceph, so he can possibly also help | 13:07 |
| hwoarang | ok good to know! | 13:08 |
| andymccr | let me take a look at our playbook quick | 13:08 |
| hwoarang | or maybe somethign went wrong in the ceph-install.yml playbook in the first place | 13:10 |
| andymccr | hwoarang, is this on master? | 13:10 |
| hwoarang | yes | 13:10 |
| hwoarang | hmm actually package installation happens on ceph-common but we call it after ceph-defaults | 13:11 |
| evrardjp | is that a ptyhon package? | 13:11 |
| andymccr | my personal opinion is that we should have some playbooks to configure the containers and do OSA specific bits - but then just call the site.yml from ceph-ansible (we could copy it over at each version bump if needed) | 13:17 |
| odyssey4me | andymccr: I wholeheartedly agree | 13:17 |
| *** shyamb has joined #openstack-ansible | 13:18 | |
| odyssey4me | We should also implement a different ansible runtime for when we run ceph-ansible's site.yml so that we're not tightly coupled to them. | 13:18 |
| andymccr | although in this case - defaults is called first, then common, then config then ceph-rgw role | 13:18 |
| hwoarang | i see that this task is being skipped on xenial | 13:19 |
| hwoarang | http://logs.openstack.org/78/590778/4/check/openstack-ansible-deploy-ceph-ubuntu-xenial/91db0b4/job-output.txt.gz#_2018-08-13_10_20_39_499760 | 13:19 |
| hwoarang | so question for me is to find out why suse executes it | 13:19 |
| andymccr | hwoarang, ahh so does this work on ubuntu just not on SuSE? | 13:23 |
| hwoarang | looks like it | 13:23 |
| andymccr | ok so the repo setup might not be correct | 13:23 |
| andymccr | that should be an easy enough fix | 13:23 |
| hwoarang | i originally thought that it was a package installation issue but it seems that this task should be skipped instead | 13:24 |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible stable/rocky: AIO: Tune the nova scheduler workers down https://review.openstack.org/591063 | 13:24 |
| odyssey4me | cores - could we get this backport in please? https://review.openstack.org/591041 | 13:25 |
| odyssey4me | and https://review.openstack.org/591043 | 13:25 |
| *** DanyC has quit IRC | 13:26 | |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible stable/rocky: Eliminate installing pip on host/containers https://review.openstack.org/591331 | 13:26 |
| openstackgerrit | Dmitriy Rabotjagov (noonedeadpunk) proposed openstack/openstack-ansible-os_gnocchi stable/queens: Added drivers for coordination. https://review.openstack.org/591385 | 13:26 |
| openstackgerrit | Dmitriy Rabotjagov (noonedeadpunk) proposed openstack/openstack-ansible-os_gnocchi stable/rocky: Added drivers for coordination. https://review.openstack.org/591386 | 13:26 |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible stable/rocky: AIO: Tune down the glance uwsgi processes https://review.openstack.org/591062 | 13:26 |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible stable/rocky: AIO: Tune the nova scheduler workers down https://review.openstack.org/591063 | 13:26 |
| andymccr | hwoarang, that is weird - that task shouldnt get included on SuSE because it should go to the "install_on_suse.yml" from ceph-common - which can't/won't include that | 13:27 |
| hwoarang | ok i think i got the bottom of it | 13:27 |
| hwoarang | the task that's failing has ansible_hostname != ansible_fqdn | 13:28 |
| hwoarang | and i get "ansible_hostname": "aio1-ceph-rgw-container-8f1c8d43" and "ansible_fqdn": "aio1-ceph-rgw-container-8f1c8d43.openstack.local" | 13:28 |
| hwoarang | meh | 13:28 |
| hwoarang | so many questions | 13:29 |
| mnaser | hwoarang: using unbound or not | 13:31 |
| mnaser | maybe we should look into the whole unbound vs hosts file | 13:31 |
| jrosser | ceph not found fail in one of my xenial check jobs too | 13:31 |
| hwoarang | mnaser: i am not using anything myself | 13:32 |
| hwoarang | whatever AIO configures | 13:32 |
| hwoarang | still trying to dig deeper | 13:32 |
| odyssey4me | mnaser: yes the plan was always to have unbound there, then switch it to the default for a cycle, then remove the hosts file changes the cycle after | 13:33 |
| odyssey4me | perhaps we should switch the default in rocky | 13:33 |
| hwoarang | so does that mean that ansible_hostname should return the fqdn from now on? | 13:33 |
| hwoarang | that's what i understand of them release note about openstack_domain variable | 13:34 |
| *** hamzaachi_ has joined #openstack-ansible | 13:35 | |
| *** spatel_ has joined #openstack-ansible | 13:36 | |
| *** hamzaachi has quit IRC | 13:37 | |
| hwoarang | to be it seems that ansible_hostname should not return fqdn given that the inventory does not container fqdn | 13:37 |
| *** hamzaachi_ has quit IRC | 13:38 | |
| hwoarang | but then again, looks like ceph is expecting ansible_hostname to be fqdn | 13:38 |
| hwoarang | s/to be/to me | 13:38 |
| *** hamzaachi_ has joined #openstack-ansible | 13:38 | |
| mnaser | odyssey4me: any idea if CI right now does unbound coverage or not? | 13:39 |
| odyssey4me | mnaser: it does not | 13:39 |
| *** vollman has quit IRC | 13:39 | |
| odyssey4me | mnaser: logan-'s external CI did - but I've not seen it posting results any more, so I think it's disabled while he figures out what's going on there | 13:40 |
| mnaser | i'd defer to evrardjp on his thoughts for rocky cycle decisions but i think it would help clean out al ot of things | 13:41 |
| mnaser | i've had weird issues with rabbitmq clusters failing to bootstrap because the ordering of stuff in /etc/hosts was weird | 13:41 |
| jrosser | if the rabbit@<hostname> you pass it doesnt match what it thinks the hostname is when setting up the policies it blows up, i had that today | 13:42 |
| openstackgerrit | Andy Smith proposed openstack/openstack-ansible master: [WIP] Add qdrouterd role for messaging backend deployment https://review.openstack.org/570261 | 13:42 |
| evrardjp | hwoarang: that's weird that ansible_hostname has to be an fqdn that's not what ansible is expecting | 13:43 |
| evrardjp | there is ansible_fqdn IIRC for that | 13:43 |
| jrosser | i'm seeing a ceph on xenial fail here http://logs.openstack.org/46/589146/2/check/openstack-ansible-deploy-ceph-ubuntu-xenial/3444486/job-output.txt.gz#_2018-08-13_12_01_32_881202 | 13:43 |
| evrardjp | changing default things to DNS is opening up a large series of problems, because most ppl can't debug dns properly | 13:44 |
| evrardjp | s/most/many/ | 13:44 |
| *** hamzaachi has joined #openstack-ansible | 13:44 | |
| evrardjp | yeah, that's experience talking here, not numbers | 13:44 |
| evrardjp | "woops I enabled dnssec and now everything is broken" | 13:44 |
| evrardjp | that's why I don't think it's good that we bring that in NOW -- if we bring it now. Master is fine | 13:45 |
| hwoarang | well i only state what i see here | 13:46 |
| hwoarang | https://github.com/ceph/ceph-ansible/blob/master/roles/ceph-defaults/tasks/facts.yml#L238 | 13:46 |
| hwoarang | so for us these two variables differ, ceph is not installed at this point, and task blows up | 13:46 |
| *** hamzaachi_ has quit IRC | 13:46 | |
| hwoarang | yeah jrosser exact same failure on suse too | 13:47 |
| jrosser | hwoarang: that only got changed very recently https://github.com/ceph/ceph-ansible/commit/f422efb1d6b56ce56a7d39a21736a471e4ed357c | 13:47 |
| hwoarang | yep | 13:48 |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible master: Remove the last remnants of get-pip.py https://review.openstack.org/591392 | 13:48 |
| hwoarang | the conditional is rather weird. maybe it should have been ansible_hostname != inventory_hostname | 13:50 |
| *** savvas has joined #openstack-ansible | 13:51 | |
| savvas | Anyone running into this issue when running os-octavia install? http://paste.openstack.org/show/727925/ | 13:55 |
| *** shyamb has quit IRC | 14:03 | |
| *** dxiri has joined #openstack-ansible | 14:04 | |
| *** fghaas has left #openstack-ansible | 14:04 | |
| openstackgerrit | Jonathan Rosser proposed openstack/openstack-ansible master: [WIP] Use parallel git clone https://review.openstack.org/588372 | 14:05 |
| *** SamYaple has joined #openstack-ansible | 14:13 | |
| *** mmalchuk_ has quit IRC | 14:14 | |
| *** cshen has quit IRC | 14:15 | |
| andymccr | hwoarang, might want a new issue for that - since the thing has merged already | 14:17 |
| *** hamzaachi has quit IRC | 14:17 | |
| andymccr | i'll ping guillaume | 14:17 |
| *** hamzaachi has joined #openstack-ansible | 14:17 | |
| hwoarang | andymccr: first i wanted to confirm that it is a bug on the ceph side and not on ours | 14:18 |
| hwoarang | thank you | 14:18 |
| andymccr | im not sure whether its a ceph side bug or not, but - it seems like an odd requirement | 14:21 |
| *** exodusftw has quit IRC | 14:24 | |
| *** Bhujay has quit IRC | 14:27 | |
| cloudnull | mornings | 14:33 |
| jrosser | o/ | 14:34 |
| openstackgerrit | Merged openstack/openstack-ansible-ops master: Add Prometheus node_exporter https://review.openstack.org/587090 | 14:48 |
| Tahvok | Guys, what happens if I added accidentaly a compute without running by the add compute docs? | 14:51 |
| Tahvok | I mean if I just ran setup-everything without the --skip-tags nova-key-distribute and everything that's needed when adding a compute? | 14:51 |
| *** tailDASHeF has joined #openstack-ansible | 14:51 | |
| odyssey4me | Tahvok: if you just ran setup-everything and it worked, problem solved? | 14:54 |
| Tahvok | odyssey4me: isn't there a reason for the skip-tags part? | 14:54 |
| odyssey4me | only to make it happen faster | 14:54 |
| *** hamzaachi_ has joined #openstack-ansible | 14:54 | |
| odyssey4me | those instructions are for adding a single host without having to run everything | 14:55 |
| *** hamzaachi has quit IRC | 14:57 | |
| *** hamzaachi has joined #openstack-ansible | 14:59 | |
| Tahvok | odyssey4me: then why you should first run "openstack-ansible setup-openstack.yml --skip-tags nova-key-distribute --limit NEW_HOST_NAME" and then still run "openstack-ansible setup-openstack.yml --tags ceph-config,nova-key --limit compute_hosts" | 15:00 |
| *** tailDASHeF has quit IRC | 15:00 | |
| *** vnogin has quit IRC | 15:01 | |
| *** hamzaachi_ has quit IRC | 15:01 | |
| odyssey4me | Tahvok: I think the idea is first to generate the new key on that host and gather its facts, then populate its ssh key to the other hosts after | 15:01 |
| odyssey4me | The keys are used for live migrations | 15:02 |
| Tahvok | odyssey4me: that's what I thought.. So the question is, what happens if I didn't do it this way? | 15:03 |
| odyssey4me | Tahvok: If you run without tags, then the playbook does all that for you. | 15:04 |
| odyssey4me | It takes a lot longer, because it's verifying all sorts of things, but it does the same things. | 15:05 |
| *** hamzaachi_ has joined #openstack-ansible | 15:06 | |
| openstackgerrit | Jean-Philippe Evrard proposed openstack/openstack-ansible master: Replace tempest with healthchecks https://review.openstack.org/587544 | 15:07 |
| *** hamzaachi has quit IRC | 15:07 | |
| openstackgerrit | Jonathan Rosser proposed openstack/openstack-ansible-lxc_hosts master: Install eatmydata inside containers https://review.openstack.org/591412 | 15:08 |
| *** dxiri has quit IRC | 15:08 | |
| openstackgerrit | Jonathan Rosser proposed openstack/openstack-ansible-nspawn_hosts master: Install eatmydata inside containers https://review.openstack.org/591414 | 15:08 |
| *** hamzaachi has joined #openstack-ansible | 15:09 | |
| Tahvok | Ok, that's a relief.. I was afraid we broke something.. Maybe we should specify in the docs that the only reason for the steps it's for faster run, but you could still do withe the full osa run? | 15:10 |
| *** hamzaachi_ has quit IRC | 15:12 | |
| *** cshen has joined #openstack-ansible | 15:15 | |
| *** rodolof has quit IRC | 15:18 | |
| evrardjp | odyssey4me: https://review.openstack.org/#/c/590503/ is based on https://review.openstack.org/#/c/591357/1 so you can probably upgrade your vote | 15:21 |
| *** mmalchuk has joined #openstack-ansible | 15:23 | |
| *** savvas has quit IRC | 15:23 | |
| *** mmalchuk_ has joined #openstack-ansible | 15:25 | |
| openstackgerrit | Chandan Kumar proposed openstack/openstack-ansible-os_tempest master: [DNM] testing https://review.openstack.org/591424 | 15:26 |
| *** mmalchuk has quit IRC | 15:28 | |
| *** vnogin has joined #openstack-ansible | 15:28 | |
| openstackgerrit | Jonathan Rosser proposed openstack/openstack-ansible master: Enable eatmydata for apt package manager on AIO builds https://review.openstack.org/591426 | 15:31 |
| cjloader | hello | 15:31 |
| *** DanyC has joined #openstack-ansible | 15:32 | |
| *** shyamb has joined #openstack-ansible | 15:34 | |
| *** dw has joined #openstack-ansible | 15:35 | |
| *** DanyC has quit IRC | 15:37 | |
| *** pcaruana has quit IRC | 15:37 | |
| *** hamzaachi has quit IRC | 15:42 | |
| *** DanyC has joined #openstack-ansible | 15:43 | |
| *** olivierbourdon38 has quit IRC | 15:44 | |
| *** mmalchuk has joined #openstack-ansible | 15:57 | |
| *** mmalchuk_ has quit IRC | 15:59 | |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible master: Remove the last remnants of get-pip.py https://review.openstack.org/591392 | 16:00 |
| *** cshen has quit IRC | 16:09 | |
| *** vakuznet has joined #openstack-ansible | 16:14 | |
| *** vnogin has quit IRC | 16:15 | |
| *** dxiri has joined #openstack-ansible | 16:16 | |
| *** gyee has joined #openstack-ansible | 16:23 | |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible-ops master: MNAIO: Use single vars file for Ubuntu trusty/xenial/bionic https://review.openstack.org/591454 | 16:25 |
| *** olivierbourdon38 has joined #openstack-ansible | 16:26 | |
| *** dxiri has quit IRC | 16:28 | |
| mnaser | jrosser: is there any thread or discussion regarding the eatmydata stuff | 16:30 |
| jrosser | mnaser: no there isnt - shall we discuss? | 16:32 |
| *** cshen has joined #openstack-ansible | 16:32 | |
| mnaser | jrosser: well i was just curious behind why it's being added (if there is an eventual ci thing) | 16:34 |
| jrosser | my local tests suggest at least 2x speedup for apt install | 16:34 |
| jrosser | however i reaaly should verify that also holds in CI | 16:35 |
| jrosser | hence the patches thensleves will become proof if is worth javing | 16:35 |
| jrosser | mnaser: evrardjp points out that i can roll them all into one patch by oerriding the right vars which is nicer | 16:38 |
| *** pcaruana has joined #openstack-ansible | 16:41 | |
| mnaser | jrosser: 2x speed up sounds amazing | 16:47 |
| mnaser | and we're in CI, don't think disk consistency is all that important | 16:47 |
| mnaser | yeah i think squashing it all into a change that changes it would be super super cool | 16:47 |
| jrosser | 2x speed up for apt install tasks only! | 16:49 |
| jrosser | not for everything but it stands a chance of being significant for ubuntu distro installs | 16:49 |
| jrosser | mnaser: i have been putting together an etherpad on the general topic of "go faster" which we could discuss tomorrow | 16:50 |
| mnaser | jrosser: will you be at the ptg? | 16:51 |
| *** hamzaachi has joined #openstack-ansible | 16:51 | |
| mnaser | jrosser: https://etherpad.openstack.org/p/osa-stein-ptg i'd love to see it discussed there | 16:51 |
| jrosser | sadly not this time | 16:52 |
| *** aludwar has quit IRC | 16:53 | |
| evrardjp | jrosser: don't forget to think about where this will be undefined if you pass that as extra vars :) | 16:54 |
| jrosser | yep indeed - better suggestion? | 16:55 |
| evrardjp | groups vars or we bring smarter overrides in openstack-hosts | 16:58 |
| evrardjp | I have not tried, it might work with blabla | default([]) | 16:58 |
| jrosser | i got bit by that with the elk stuff | 16:58 |
| jrosser | wanting user_variables which then failed due to a pile of undefined stuff | 16:59 |
| evrardjp | makes sense due to their evaluation order | 17:00 |
| evrardjp | group vars is easier :) | 17:01 |
| *** gillesMo has quit IRC | 17:01 | |
| jrosser | you are right there, i'll poke at this later | 17:01 |
| evrardjp | but I wouldn't mind if we had another way to define this in testing -- a different package list (two lists to maintain are not THAT hard) | 17:04 |
| evrardjp | I just think we should use the mechanism in there instead of writing code in the roles | 17:05 |
| odyssey4me | evrardjp: some roles already have a set of 'test' or 'developer_mode' packages - perhaps we should just extend that idea | 17:07 |
| odyssey4me | perhaps change it to an 'extra_distro_packages' list which is empty by default, and the tests add to it where needed | 17:07 |
| odyssey4me | we could do the same with the pip packages - make an 'extra_optional_pip_packages' var in each role, making it easy to extend the package lists | 17:08 |
| openstackgerrit | Kevin Carter (cloudnull) proposed openstack/openstack-ansible-os_neutron master: Update permissions on execution path folders https://review.openstack.org/591463 | 17:10 |
| openstackgerrit | Kevin Carter (cloudnull) proposed openstack/openstack-ansible-os_nova master: Cleanup files and templates using smart sources https://review.openstack.org/588951 | 17:11 |
| openstackgerrit | Kevin Carter (cloudnull) proposed openstack/openstack-ansible-os_cinder master: Cleanup files and templates using smart sources https://review.openstack.org/588953 | 17:13 |
| openstackgerrit | Kevin Carter (cloudnull) proposed openstack/openstack-ansible-os_heat master: Cleanup files and templates using smart sources https://review.openstack.org/588954 | 17:14 |
| *** dxiri has joined #openstack-ansible | 17:14 | |
| openstackgerrit | Kevin Carter (cloudnull) proposed openstack/openstack-ansible-os_keystone master: Cleanup files and templates using smart sources https://review.openstack.org/588960 | 17:16 |
| openstackgerrit | Kevin Carter (cloudnull) proposed openstack/openstack-ansible-os_glance master: Cleanup files and templates using smart sources https://review.openstack.org/588959 | 17:17 |
| *** openstackgerrit has quit IRC | 17:19 | |
| jrosser | odyssey4me: i would like that general "extra packages" style | 17:20 |
| odyssey4me | jrosser: I expect that a lot of people would, given that it's a fairly common ask to add more packages | 17:21 |
| *** openstackgerrit has joined #openstack-ansible | 17:21 | |
| openstackgerrit | Kevin Carter (cloudnull) proposed openstack/openstack-ansible-ops master: Add Redhat to the ELK deployment capabilities https://review.openstack.org/591467 | 17:21 |
| jrosser | it also avoids th horrid undefined vars when overriding stuff from roles | 17:21 |
| odyssey4me | it would also help reduce the need to add packages for drivers with a seperated task as it'd be a universal tool | 17:21 |
| odyssey4me | universal tools cut tech debt ;) | 17:22 |
| cloudnull | jrosser ^ idk if you'd seen but the ELK tools now support Ubuntu: trusty, Xenial, Bionic, and SUSE: 42.3. Soon to support Cent: 7. | 17:22 |
| cloudnull | id be great if you could give those a spin in your bionic lab to validate. | 17:22 |
| jrosser | cloudnull: i saw the patches fly past... nice work! | 17:22 |
| cloudnull | but it should have fairly comprehensive build tests, both AIO and clustered. | 17:23 |
| *** shyamb has quit IRC | 17:23 | |
| jrosser | bionic is just AIO atm, but i could certainly try to wedge a 1 node elk into there | 17:23 |
| odyssey4me | cloudnull jrosser does it make any sense to have the MNAIO tooling use zfs as an underlying storage subsystem for the VM's? | 17:23 |
| jrosser | yes | 17:23 |
| jrosser | always | 17:23 |
| odyssey4me | not the containers - the VM's | 17:23 |
| cloudnull | +1 | 17:23 |
| jrosser | but im biassed :) | 17:23 |
| odyssey4me | the containers are a different setting | 17:23 |
| odyssey4me | well, what I'd like to do is divorce the two from each other - allow one var to set how the host is laid out, and another for how the containers are laid out | 17:24 |
| odyssey4me | if it isn't already the case - busy digging | 17:25 |
| jrosser | are there helpers for kvm to do the required snapshots etc? | 17:25 |
| cloudnull | jrosser correct me if im wrong, but ZFS, like BTRFS, has a double write penalty when running workloads like a VM. | 17:25 |
| odyssey4me | I'm thinking of employing a task routing system too - because the playbooks are getting a little more hairy with each feature added. | 17:25 |
| cloudnull | In BTRFS we can set no-data-cow to fix that. | 17:25 |
| cloudnull | id assume there's a similar setting in / tunable in ZFS | 17:25 |
| cloudnull | (if the issue even exists) | 17:26 |
| jrosser | cloudnull: do you have a reference i can read? | 17:26 |
| cloudnull | yea, let me go grab that. | 17:26 |
| jrosser | whilst youre both thinking about FS things i think there is memory efficiency to be had with ZFS | 17:26 |
| cloudnull | not an official source, but I remember reading https://www.reddit.com/r/zfs/comments/7rcezf/vm_wzfs_on_kvm_host_wzfs/ | 17:27 |
| odyssey4me | jrosser cloudnull well, perhaps I should share my ultimate goal here - I need to deploy an MNAIO to whatever stage, then I want to save the disk images off to a remote location, then in another deployment I want to pull those images down and use them to kick off where I last left off | 17:27 |
| jrosser | i need to run it all again but aio+zfs uses significantly less ram than aio+dir | 17:27 |
| odyssey4me | So basically I was thinking make the VM's use a qemu disk, which can run on whatever we like underneath. | 17:27 |
| jrosser | zfs send/receive are good toys for that | 17:29 |
| jrosser | btrfs has similar i believe | 17:29 |
| cloudnull | ^ +1 | 17:29 |
| cloudnull | zfs is far better for send/recieve | 17:29 |
| jrosser | you just meed to be careful about combining that wih snapshots | 17:29 |
| cloudnull | BTRFS has it but its just not as well baked. | 17:29 |
| odyssey4me | oh this won't use snapshots | 17:29 |
| jrosser | because you have a parent/child relationship | 17:29 |
| odyssey4me | it'll shut the machine down, then save the disk | 17:30 |
| jrosser | but if you build a base inage, download, run, then upload the delta you are good | 17:30 |
| noonedeadpunk | cloudnull: hi) May you please look at https://review.openstack.org/#/c/590819/ | 17:30 |
| odyssey4me | using snapshots on a machine will be something to look at later for the convenience of using the system once you're on one | 17:30 |
| noonedeadpunk | and https://review.openstack.org/#/c/591338/ if you can | 17:31 |
| cloudnull | noonedeadpunk looking now | 17:31 |
| jrosser | anyway, i have a feeling that zfs is allowing my containers to share pages | 17:31 |
| jrosser | because things that are the same in the base inage are the same blocks on disk | 17:32 |
| jrosser | and therefore a memory efficiency | 17:32 |
| cloudnull | noonedeadpunk +2 from me. | 17:32 |
| jrosser | it is sad the btrfs seems to have a unique page cache per subvolime | 17:32 |
| noonedeadpunk | cloudnull thanks) | 17:32 |
| cloudnull | mind cherry-picking those comitts to stable/rocky ? | 17:33 |
| odyssey4me | noonedeadpunk: yep, once they merge - please cherry-pick | 17:33 |
| jrosser | odyssey4me: i am going to revisit overlayfs to see if the same memory efficiency shows there | 17:33 |
| noonedeadpunk | cloudnull odyssey4me yes, of course I will. Thanks) | 17:34 |
| cloudnull | jrosser yea the unique page system is a bummer. | 17:34 |
| odyssey4me | jrosser: I seem to recall it being worse, but I might be wrong. | 17:34 |
| odyssey4me | unique page system is more secure, surely? | 17:34 |
| cloudnull | more isolation, less efficient. | 17:34 |
| noonedeadpunk | actually chrry-pich for gnocchi is already here https://review.openstack.org/#/c/591386/ https://review.openstack.org/#/c/591385/ | 17:35 |
| jrosser | it is very sad that zfs isnt more universal as my zfs gate checks seem to swap much less than dir based ones | 17:36 |
| *** bgmccollum has quit IRC | 17:37 | |
| *** ianychoi_ has joined #openstack-ansible | 17:38 | |
| *** hamzaachi_ has joined #openstack-ansible | 17:38 | |
| *** hamzaachi has quit IRC | 17:39 | |
| *** jamesdenton has quit IRC | 17:39 | |
| *** ianychoi has quit IRC | 17:41 | |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible-ops master: Optomise vm_disk_mode conditionals https://review.openstack.org/591473 | 17:41 |
| *** ppetit has quit IRC | 17:43 | |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible-ops master: Optomise vm_disk_mode conditionals https://review.openstack.org/591473 | 17:44 |
| jrosser | odyssey4me: perhaps a daft question but has an LXD flavoured mnaio been tried? | 17:45 |
| odyssey4me | jrosser: the tooling doesn't cater to it | 17:45 |
| jrosser | because all these host/guest fs problems just go away | 17:45 |
| odyssey4me | sure, perhaps, but in our efforts to use things more universally available, we're moving away from lxc/lxd anyway | 17:46 |
| mnaser | andymccr: congrats! :D | 17:46 |
| odyssey4me | no idea whether lxd can do nested containers | 17:46 |
| *** bgmccollum has joined #openstack-ansible | 17:48 | |
| *** olivierbourdon38 has quit IRC | 17:52 | |
| *** pcaruana has quit IRC | 17:59 | |
| *** poopcat has joined #openstack-ansible | 18:01 | |
| *** bgmccollum has quit IRC | 18:02 | |
| *** jamesdenton has joined #openstack-ansible | 18:04 | |
| noonedeadpunk | could you point me to some rule of describing release files? evrardjp said, that I should place a release not, but I'm not sure what is it | 18:16 |
| *** electrofelix has quit IRC | 18:19 | |
| odyssey4me | noonedeadpunk: release note | 18:19 |
| odyssey4me | noonedeadpunk: https://docs.openstack.org/openstack-ansible/latest/contributor/code-rules.html#documentation-and-release-note-guidelines | 18:20 |
| odyssey4me | more specifically https://docs.openstack.org/openstack-ansible/latest/contributor/code-rules.html#reno | 18:20 |
| openstackgerrit | Merged openstack/openstack-ansible-os_ceilometer master: Added drivers for coordination. https://review.openstack.org/590819 | 18:21 |
| noonedeadpunk | odyssey4me got it. I was reading about reno, but was not sure, that it's the tool I need now. | 18:23 |
| noonedeadpunk | learn smth new today | 18:30 |
| jrosser | odyssey4me: do we have a pattern for these extra package lists, like an empty list in defaults vs. extra_packages|default([]) | 18:35 |
| odyssey4me | jrosser: better to have a var in defaults, as those are part of our documentation | 18:35 |
| jrosser | sure, makes sense | 18:35 |
| noonedeadpunk | odyssey4me: is it acceptably to edit existing release note? | 18:37 |
| noonedeadpunk | probably not, but just to double check | 18:38 |
| odyssey4me | noonedeadpunk: only if it's not yet in a stable branch because it's not yet released | 18:39 |
| odyssey4me | once it's release, it should not be changed | 18:39 |
| noonedeadpunk | odyssey4me: another new thing to me - does zuul gate also makes recheck on "recheck" comment? | 18:46 |
| odyssey4me | yes | 18:46 |
| openstackgerrit | Dmitriy Rabotjagov (noonedeadpunk) proposed openstack/openstack-ansible-os_horizon master: Implements custom theme distribution https://review.openstack.org/586318 | 18:57 |
| *** DanyC has quit IRC | 19:00 | |
| *** DanyC has joined #openstack-ansible | 19:00 | |
| noonedeadpunk | odyssey4me: I tried to do my best here https://review.openstack.org/586318 but it's by far the only option which was the working and reasonable one. | 19:01 |
| noonedeadpunk | or at least I don't see obvious things | 19:01 |
| *** hwoarang has quit IRC | 19:02 | |
| openstackgerrit | Jonathan Rosser proposed openstack/openstack-ansible-lxc_hosts master: Allow user-defined extra distro packages https://review.openstack.org/591412 | 19:04 |
| *** DanyC has quit IRC | 19:04 | |
| jrosser | cloudnull: https://www.reddit.com/r/zfs/comments/7rcezf/vm_wzfs_on_kvm_host_wzfs/ <- that seems to be simultaneously talking about lots of unrelated things | 19:08 |
| *** fdegir_ has joined #openstack-ansible | 19:17 | |
| *** wlmbasson_ has joined #openstack-ansible | 19:17 | |
| *** geofft_ has joined #openstack-ansible | 19:19 | |
| *** mgagne_ has joined #openstack-ansible | 19:24 | |
| *** fghaas has joined #openstack-ansible | 19:25 | |
| *** idlemind has quit IRC | 19:25 | |
| *** taildashef-work[ has quit IRC | 19:25 | |
| *** fdegir has quit IRC | 19:25 | |
| *** serverascode_ has quit IRC | 19:25 | |
| *** wlmbasson has quit IRC | 19:25 | |
| *** geofft has quit IRC | 19:25 | |
| *** mgagne has quit IRC | 19:25 | |
| *** eoseq has quit IRC | 19:25 | |
| *** Miouge has quit IRC | 19:25 | |
| *** charz has quit IRC | 19:25 | |
| *** wlmbasson_ is now known as wlmbasson | 19:25 | |
| *** fdegir_ is now known as fdegir | 19:25 | |
| *** Miouge- has joined #openstack-ansible | 19:25 | |
| openstackgerrit | Andy Smith proposed openstack/openstack-ansible master: [WIP] Add qdrouterd role for messaging backend deployment https://review.openstack.org/570261 | 19:29 |
| *** fghaas has quit IRC | 19:30 | |
| openstackgerrit | Merged openstack/openstack-ansible-os_gnocchi stable/queens: Added drivers for coordination. https://review.openstack.org/591385 | 19:30 |
| openstackgerrit | Merged openstack/openstack-ansible-os_gnocchi stable/rocky: Added drivers for coordination. https://review.openstack.org/591386 | 19:30 |
| *** vollman has joined #openstack-ansible | 19:35 | |
| openstackgerrit | Merged openstack/openstack-ansible-lxc_container_create stable/rocky: Implements possibility to use physical networks in containers. https://review.openstack.org/591338 | 19:35 |
| *** DanyC has joined #openstack-ansible | 19:41 | |
| *** jproulx has quit IRC | 19:42 | |
| *** DanyC has quit IRC | 19:44 | |
| *** DanyC has joined #openstack-ansible | 19:44 | |
| *** cshen has quit IRC | 19:48 | |
| *** geofft_ is now known as geofft | 19:49 | |
| openstackgerrit | Merged openstack/openstack-ansible-os_monasca-agent master: Add support for CentOS 7 https://review.openstack.org/588456 | 20:02 |
| *** dxiri has quit IRC | 20:03 | |
| *** bgmccollum has joined #openstack-ansible | 20:04 | |
| *** dxiri has joined #openstack-ansible | 20:07 | |
| *** dxiri has quit IRC | 20:07 | |
| *** dxiri has joined #openstack-ansible | 20:08 | |
| *** cshen has joined #openstack-ansible | 20:12 | |
| openstackgerrit | Dmitriy Rabotjagov (noonedeadpunk) proposed openstack/openstack-ansible-os_ceilometer stable/rocky: Added drivers for coordination. https://review.openstack.org/591335 | 20:18 |
| *** hamzaachi_ has quit IRC | 20:24 | |
| *** vakuznet has quit IRC | 20:39 | |
| *** hamzaachi has joined #openstack-ansible | 20:51 | |
| *** spatel_ has quit IRC | 20:55 | |
| *** throwsb1 has joined #openstack-ansible | 21:05 | |
| *** hwoarang has joined #openstack-ansible | 21:18 | |
| *** hamzaachi has quit IRC | 21:27 | |
| *** throwsb1 has quit IRC | 21:32 | |
| openstackgerrit | Merged openstack/openstack-ansible-os_ironic stable/rocky: update pip packages for ironic drivers https://review.openstack.org/591043 | 21:41 |
| *** spatel has joined #openstack-ansible | 21:58 | |
| openstackgerrit | Kevin Carter (cloudnull) proposed openstack/openstack-ansible-ops master: Add Redhat to the ELK deployment capabilities https://review.openstack.org/591467 | 22:01 |
| *** spatel has quit IRC | 22:03 | |
| *** cshen has quit IRC | 22:09 | |
| *** DanyC has quit IRC | 22:23 | |
| *** DanyC has joined #openstack-ansible | 22:23 | |
| *** DanyC has quit IRC | 22:28 | |
| openstackgerrit | Merged openstack/openstack-ansible-ops master: MNAIO: Use single vars file for Ubuntu trusty/xenial/bionic https://review.openstack.org/591454 | 22:29 |
| *** poopcat has quit IRC | 22:41 | |
| *** poopcat has joined #openstack-ansible | 22:42 | |
| openstackgerrit | Kevin Carter (cloudnull) proposed openstack/openstack-ansible-ops master: Add Redhat to the ELK deployment capabilities https://review.openstack.org/591467 | 22:52 |
| *** spatel has joined #openstack-ansible | 23:02 | |
| *** spatel has quit IRC | 23:05 | |
Generated by irclog2html.py 2.15.3 by Marius Gedminas - find it at mg.pov.lt!