| *** aedc has quit IRC | 00:00 | |
| *** DanyC_ has quit IRC | 00:07 | |
| *** markvoelker has joined #openstack-ansible | 00:09 | |
| *** markvoelker has quit IRC | 00:14 | |
| *** sdake has joined #openstack-ansible | 00:26 | |
| *** chandankumar has quit IRC | 00:28 | |
| *** chandankumar has joined #openstack-ansible | 00:29 | |
| *** cmart has quit IRC | 00:31 | |
| *** vnogin has joined #openstack-ansible | 00:34 | |
| cloudnull | cmart still around? -- I've not given OSA a go in such an environment. the largest I ever managed in a single region was: ~500 compute nodes, ~200 storage nodes, ~5 infra hosts, 3 L3 network nodes. | 00:37 |
|---|---|---|
| cloudnull | the ops side of things was not really tough, but it did take a while to roll out updates. | 00:37 |
| cloudnull | that was in the mitaka timeframe , so I suspect that same deployment would be a lot simpler to manage today than it was then | 00:38 |
| *** vnogin has quit IRC | 00:39 | |
| cloudnull | we had to create a lot of monitoring and metric collection on services, we cared less about specific host downs and really drove toward scheduler available. so a lot of monitoring was focused on resource available and capacity planning | 00:39 |
| jamesdenton | good times had by all | 00:40 |
| cloudnull | indeed! | 00:40 |
| cloudnull | host down was all about: quick fix if possible (reboot), evacuate (if possible), say sorry (if not), call hot hands to replace gear. | 00:41 |
| cloudnull | but it wasn't a death march and wasn't something that required a lot of heads. | 00:41 |
| cloudnull | jamesdenton knows what I'm talking about :) | 00:41 |
| jamesdenton | :D | 00:41 |
| cloudnull | 10/8 for life! | 00:41 |
| jamesdenton | i miss having those machines | 00:42 |
| cloudnull | me too. | 00:42 |
| cloudnull | hows it jamesdenton? | 00:43 |
| jamesdenton | it's goin. trying to get thru the nighttime routine and maybe sit down with some scotch and poke at an AIO | 00:44 |
| jamesdenton | brb | 00:47 |
| cloudnull | ooh, scotch and cloud. some great times to be had by all | 00:54 |
| cloudnull | bgmccollum did you make it go ? | 00:54 |
| *** tosky has quit IRC | 00:57 | |
| *** sdake has quit IRC | 00:58 | |
| *** markvoelker has joined #openstack-ansible | 01:10 | |
| *** sdake has joined #openstack-ansible | 01:14 | |
| *** sdake has quit IRC | 01:16 | |
| *** sdake has joined #openstack-ansible | 01:21 | |
| *** sdake has quit IRC | 01:42 | |
| *** sdake has joined #openstack-ansible | 01:43 | |
| *** markvoelker has quit IRC | 01:44 | |
| *** sdake has quit IRC | 01:54 | |
| *** sdake has joined #openstack-ansible | 02:00 | |
| *** sdake has quit IRC | 02:22 | |
| openstackgerrit | Andriy Shevchenko proposed openstack/openstack-ansible-galera_server master: fix tox python3 overrides https://review.openstack.org/615490 | 02:31 |
| *** markvoelker has joined #openstack-ansible | 02:41 | |
| *** hwoarang has quit IRC | 02:59 | |
| *** hwoarang has joined #openstack-ansible | 03:02 | |
| *** markvoelker has quit IRC | 03:13 | |
| openstackgerrit | Kevin Carter (cloudnull) proposed openstack/openstack-ansible-ops master: Update grafana, use vendored role, and add es lb https://review.openstack.org/637672 | 03:20 |
| *** mgariepy has quit IRC | 03:33 | |
| chandankumar | cloudnull: Hello | 03:33 |
| cloudnull | o/ | 03:33 |
| chandankumar | cloudnull: I have successfully deployed openstack-ansible aio_metal_heat scenario | 03:34 |
| cloudnull | sweet! | 03:35 |
| chandankumar | using gate-check-commit.sh script | 03:35 |
| chandankumar | from source | 03:35 |
| cloudnull | nice! | 03:35 |
| chandankumar | cloudnull: now the problem is there I am currently ssh into the box it is dumping here [root@aio1 ~] | 03:36 |
| chandankumar | cloudnull: there is no lxc related commands there | 03:36 |
| chandankumar | cloudnull: or I missing something? | 03:36 |
| chandankumar | there only openstack-ansible and tempest workspace exists | 03:36 |
| chandankumar | and openrc file | 03:36 |
| cloudnull | aio_metal_heat would deploy everything on the single host without using any containers | 03:37 |
| cloudnull | so all of the services should be deployed and running on that one AIO | 03:37 |
| chandankumar | cloudnull: so there is no lxc containers | 03:37 |
| cloudnull | if you check `ps auxf` or systemd-cgls you'll see all of the cgroups and processes but no containers. | 03:37 |
| cloudnull | chandankumar yes. there are no containers when using that scenario | 03:38 |
| chandankumar | cloudnull: awesome | 03:38 |
| chandankumar | cloudnull: now I can see the venv and other stuff | 03:38 |
| cloudnull | yes. a | 03:38 |
| chandankumar | cloudnull: why the version tempest-18.0.0.0rc2.dev370 ends with 18.0.0? | 03:38 |
| cloudnull | all of the venvs will be in the /openstack directory | 03:38 |
| cloudnull | the venv or the version of the service in the venv directory | 03:39 |
| cloudnull | I would suspect "18.0.0.0rc2.dev370" is the version of tempest deployed | 03:40 |
| chandankumar | cloudnull: venv-<version> | 03:40 |
| chandankumar | cloudnull: tempest 19.0.0 is installed there on sourceing the venv | 03:40 |
| cloudnull | so that's the venv version ? | 03:41 |
| chandankumar | cloudnull: http://paste.openstack.org/show/745300/ | 03:41 |
| cloudnull | something like /openstack/venvs/tempest-18.0.0.0rc2.dev370 | 03:41 |
| chandankumar | (tempest-18.0.0.0rc2.dev370) [root@aio1 venvs]# virtualenv --version | 03:41 |
| chandankumar | 15.1.0 | 03:41 |
| cloudnull | (tempest-18.0.0.0rc2.dev370) is the activated venv | 03:42 |
| cloudnull | which in that case would be the OSA version | 03:42 |
| chandankumar | cloudnull: other venv http://paste.openstack.org/show/745301/ | 03:42 |
| chandankumar | cloudnull: ok, got it | 03:42 |
| cloudnull | you can see that in the /etc/openstack-release file | 03:42 |
| cloudnull | we mark all of the venvs we deploy with the the OSA code version | 03:43 |
| cloudnull | which makes it easier to roll back if needed | 03:43 |
| chandankumar | cloudnull: DISTRIB_RELEASE="18.0.0.0rc2.dev370" | 03:43 |
| cloudnull | ++ | 03:43 |
| chandankumar | cool, thanks, Now I can debug heat stuff now :-) | 03:43 |
| cloudnull | sweet! | 03:43 |
| cloudnull | you should also be able to redeploy from new sources as needed. | 03:45 |
| *** shyamb has joined #openstack-ansible | 03:45 | |
| *** dave-mccowan has joined #openstack-ansible | 03:45 | |
| cloudnull | a friend of mine, Miguel did a talk about this a while back https://developer.rackspace.com/blog/life-without-devstack-openstack-development-with-osa/ | 03:46 |
| cloudnull | http://confreaks.tv/videos/openstacksummittokyo2015-life-without-devstack-upstream-development-with-osad | 03:46 |
| cloudnull | which was cool! | 03:46 |
| chandankumar | cloudnull: will check that thanks :-) | 03:51 |
| *** udesale has joined #openstack-ansible | 03:54 | |
| *** mgariepy has joined #openstack-ansible | 04:06 | |
| *** shyamb has quit IRC | 04:06 | |
| *** markvoelker has joined #openstack-ansible | 04:10 | |
| *** vnogin has joined #openstack-ansible | 04:10 | |
| *** hamzy has quit IRC | 04:13 | |
| *** vnogin has quit IRC | 04:15 | |
| *** jamesdenton has quit IRC | 04:16 | |
| *** jamesdenton has joined #openstack-ansible | 04:16 | |
| *** chandankumar is now known as chkumar|ruck | 04:32 | |
| *** markvoelker has quit IRC | 04:44 | |
| openstackgerrit | Chandan Kumar proposed openstack/openstack-ansible-os_tempest master: Added dependency of os_tempest role https://review.openstack.org/632726 | 04:46 |
| *** irclogbot_0 has quit IRC | 04:55 | |
| *** medberry has quit IRC | 05:07 | |
| *** shyamb has joined #openstack-ansible | 05:08 | |
| *** dave-mccowan has quit IRC | 05:13 | |
| *** mgariepy has quit IRC | 05:29 | |
| *** mgariepy has joined #openstack-ansible | 05:30 | |
| *** udesale has quit IRC | 05:33 | |
| *** markvoelker has joined #openstack-ansible | 05:41 | |
| *** udesale has joined #openstack-ansible | 05:43 | |
| *** shyamb has quit IRC | 06:03 | |
| *** shyamb has joined #openstack-ansible | 06:09 | |
| *** markvoelker has quit IRC | 06:14 | |
| *** shyamb has quit IRC | 06:37 | |
| *** shyamb has joined #openstack-ansible | 06:39 | |
| openstackgerrit | Chandan Kumar proposed openstack/openstack-ansible-os_tempest master: Added tempest.conf for heat_plugin https://review.openstack.org/632021 | 06:47 |
| openstackgerrit | Chandan Kumar proposed openstack/openstack-ansible-os_tempest master: Added tempest.conf for heat_plugin https://review.openstack.org/632021 | 06:50 |
| openstackgerrit | Chandan Kumar proposed openstack/openstack-ansible-os_tempest master: Use the correct heat tests https://review.openstack.org/630695 | 06:50 |
| openstackgerrit | Chandan Kumar proposed openstack/openstack-ansible-os_heat master: [DNM] heat tempest tests finding https://review.openstack.org/630694 | 06:50 |
| *** bgmccollum has quit IRC | 06:53 | |
| *** rgogunskiy has joined #openstack-ansible | 06:56 | |
| *** DanyC has joined #openstack-ansible | 06:58 | |
| *** miloa has joined #openstack-ansible | 07:02 | |
| *** kopecmartin|off is now known as kopecmartin | 07:02 | |
| *** markvoelker has joined #openstack-ansible | 07:11 | |
| *** bgmccollum has joined #openstack-ansible | 07:14 | |
| *** DanyC has quit IRC | 07:28 | |
| *** shyamb has quit IRC | 07:33 | |
| *** markvoelker has quit IRC | 07:43 | |
| *** shyamb has joined #openstack-ansible | 07:45 | |
| *** udesale has quit IRC | 07:58 | |
| *** shyam89 has joined #openstack-ansible | 08:01 | |
| *** udesale has joined #openstack-ansible | 08:03 | |
| openstackgerrit | Chandan Kumar proposed openstack/openstack-ansible-os_tempest master: Added tempest.conf for heat_plugin https://review.openstack.org/632021 | 08:03 |
| openstackgerrit | Chandan Kumar proposed openstack/openstack-ansible-os_tempest master: Use the correct heat tests https://review.openstack.org/630695 | 08:04 |
| *** shyamb has quit IRC | 08:05 | |
| *** udesale has quit IRC | 08:07 | |
| *** udesale has joined #openstack-ansible | 08:08 | |
| *** pcaruana has joined #openstack-ansible | 08:11 | |
| *** shardy has joined #openstack-ansible | 08:15 | |
| *** shardy has quit IRC | 08:15 | |
| *** shyam89 has quit IRC | 08:16 | |
| *** dswebb has joined #openstack-ansible | 08:22 | |
| fnpanic | morning | 08:24 |
| *** sdake has joined #openstack-ansible | 08:28 | |
| *** goldenfri has quit IRC | 08:31 | |
| *** shardy has joined #openstack-ansible | 08:31 | |
| *** shyamb has joined #openstack-ansible | 08:31 | |
| *** evrardjp_ is now known as evrardjp | 08:31 | |
| *** sdake has quit IRC | 08:37 | |
| *** electrofelix has joined #openstack-ansible | 08:38 | |
| *** markvoelker has joined #openstack-ansible | 08:40 | |
| *** tosky has joined #openstack-ansible | 08:42 | |
| *** goldenfri has joined #openstack-ansible | 08:44 | |
| *** sdake has joined #openstack-ansible | 08:45 | |
| *** hamzaachi has joined #openstack-ansible | 08:49 | |
| *** aedc has joined #openstack-ansible | 08:51 | |
| *** shyamb has quit IRC | 08:51 | |
| *** shyamb has joined #openstack-ansible | 08:52 | |
| *** DanyC has joined #openstack-ansible | 08:57 | |
| *** priteau has joined #openstack-ansible | 08:59 | |
| *** DanyC has quit IRC | 09:08 | |
| *** DanyC has joined #openstack-ansible | 09:09 | |
| CeeMac | morning | 09:14 |
| *** markvoelker has quit IRC | 09:14 | |
| *** ostackz has quit IRC | 09:16 | |
| *** shyamb has quit IRC | 09:18 | |
| *** shyamb has joined #openstack-ansible | 09:19 | |
| *** sdake has quit IRC | 09:25 | |
| openstackgerrit | Chandan Kumar proposed openstack/openstack-ansible-os_tempest master: Revert "Only init a workspace if doesn't exists" https://review.openstack.org/637801 | 09:27 |
| *** sdake has joined #openstack-ansible | 09:28 | |
| dswebb | can anyone tell me what exactly needs to be set to deploy a cinder-volume container with osa? this is a snippet of the config I thought would do it, but all I have is a cinder-api container that only runns the scheduler / api. https://pastebin.com/4CGUrzhn | 09:45 |
| *** sdake has quit IRC | 09:49 | |
| *** shyamb has quit IRC | 09:55 | |
| *** shyamb has joined #openstack-ansible | 09:56 | |
| *** markvoelker has joined #openstack-ansible | 10:11 | |
| *** DanyC has quit IRC | 10:18 | |
| *** DanyC has joined #openstack-ansible | 10:19 | |
| odyssey4me | dswebb cinder-volume goes on to a host by default, not into a container | 10:32 |
| odyssey4me | dswebb this is a sample config: https://github.com/openstack/openstack-ansible/blob/master/etc/openstack_deploy/conf.d/cinder.yml.aio#L29-L44 | 10:32 |
| odyssey4me | that will do cinder-volume with a lvm back-end | 10:33 |
| dswebb | onto one of the controller nodes by default? | 10:36 |
| *** cshen has joined #openstack-ansible | 10:36 | |
| odyssey4me | guilhermesp I expect you're going to want to port https://review.openstack.org/637275 back to the appropriate stable branches? | 10:38 |
| dswebb | I've just ended up with an install with no cinder-volume running at all, so trying to backtrack and figure out where it went wrong | 10:39 |
| *** priteau has quit IRC | 10:41 | |
| odyssey4me | dswebb sorry - lost connection there, did you see any of my replies? | 10:42 |
| dswebb | I did, the sample config | 10:42 |
| chkumar|ruck | odyssey4me: hey | 10:43 |
| chkumar|ruck | odyssey4me: in morning, I finally have the osa installed system | 10:43 |
| chkumar|ruck | odyssey4me: using gate commit check script | 10:43 |
| chkumar|ruck | odyssey4me: heat tempest conf tests getting generated | 10:43 |
| *** markvoelker has quit IRC | 10:43 | |
| chkumar|ruck | odyssey4me: and the other issue guilhermesp was telling | 10:44 |
| odyssey4me | chkumar|ruck great! | 10:44 |
| CeeMac | jrosser, you around? | 10:44 |
| openstackgerrit | Chandan Kumar proposed openstack/openstack-ansible-os_tempest master: Revert "Only init a workspace if doesn't exists" https://review.openstack.org/637801 | 10:45 |
| odyssey4me | dswebb ok, I now see your response in http://eavesdrop.openstack.org/irclogs/%23openstack-ansible/%23openstack-ansible.2019-02-19.log.html - you need to choose where your volume service will run... if you're using ceph, then a controller will do. if you're using an LVM back-end, then you may wish to choose a dedicated box. | 10:51 |
| odyssey4me | It'll depend how much volume for cinder-backup you'll expect, because it can be heavy, depending on the back-end. In most cases, just using a controller is fine. | 10:52 |
| *** shyamb has quit IRC | 10:56 | |
| *** shyamb has joined #openstack-ansible | 11:01 | |
| jrosser | CeeMac: hello! i'm around | 11:02 |
| CeeMac | hi! | 11:03 |
| CeeMac | did a bit more digging through the logs on the hosts that are having network issues and it always seems to be around the lxc_hosts role with the apt install | 11:03 |
| CeeMac | systemd upgrade is restarting the networking | 11:04 |
| CeeMac | digging in more closely to the underlay, I don't think the Cisco blade switches I'm using implement lacp correctly / at all in the chassis | 11:04 |
| CeeMac | which explains why I lose network connectivity when the active port is restarted | 11:04 |
| CeeMac | my controller host randomly lost networking at 6am today too | 11:05 |
| CeeMac | not sure how this wasn't an issue when I first deployed OSA, but its not loving it now | 11:06 |
| CeeMac | so I'm probably going to have to revisit my network setup :( | 11:07 |
| CeeMac | might still look at networkd as an alternative to netplan (if I feel brave later) | 11:07 |
| CeeMac | jrosser ^ sorry should have tagged you at the start of that :) | 11:09 |
| *** udesale has quit IRC | 11:10 | |
| *** jawad_axd has joined #openstack-ansible | 11:28 | |
| *** cshen has quit IRC | 11:36 | |
| jrosser | CeeMac: i needed extra config on my cisco switches for lacp, but that really very much depends on how the servers/switches are wired | 11:38 |
| guilhermesp | mornings | 11:38 |
| guilhermesp | ah yeah odyssey4me will do that, thanks | 11:38 |
| CeeMac | jrosser, I have the bonds set to active/backup, so I didn't think lacp would be necessary | 11:39 |
| CeeMac | unless there is something blocking GARP somewhere | 11:39 |
| guilhermesp | chkumar|ruck: nice too! Let me know if you need the hold I have for the upgrade jobs too, that could be helpful for debugging | 11:39 |
| openstackgerrit | Guilherme Steinmuller Pimentel proposed openstack/openstack-ansible-galera_server stable/rocky: Iterate over list of values of PPC packages dict https://review.openstack.org/637825 | 11:39 |
| jrosser | CeeMac: ah ok i had trouble with active/active, but have fixed that now | 11:40 |
| *** markvoelker has joined #openstack-ansible | 11:40 | |
| CeeMac | i'm also a little perturbed by the issues with native vlan tag, through to vlan interface | 11:40 |
| openstackgerrit | Guilherme Steinmuller Pimentel proposed openstack/openstack-ansible-galera_server stable/queens: Iterate over list of values of PPC packages dict https://review.openstack.org/637826 | 11:41 |
| CeeMac | doesn't seem to like having a port on the native vlan where the vlan is also trunked on the same interface | 11:42 |
| CeeMac | really need to rethink my networking layout for the environment | 11:42 |
| jrosser | i'm not sure that is possible | 11:42 |
| CeeMac | it worked on the 'base' ubuntu install | 11:42 |
| jamesdenton | Are you doing VLAN pruning? Where the vlan is allowed on the interface but set as native? | 11:43 |
| CeeMac | fell over when i created tagged interfaces on the bond | 11:43 |
| CeeMac | yes | 11:43 |
| CeeMac | as its passed through for the container too | 11:43 |
| jamesdenton | As long as you don't tag that vlan, you should be OK. | 11:43 |
| CeeMac | oh, hey jamesdenton :) | 11:44 |
| jamesdenton | So a neutron 'flat' network would be appropriate. good morning | 11:44 |
| *** ansmith has joined #openstack-ansible | 11:44 | |
| CeeMac | yeah, I'm thinking I'll be better off using a seperate OOB management vlan as the native for SSH access to the host, then keeping the br-mgmt vlan as tagged end to end | 11:44 |
| jrosser | ^ good move | 11:44 |
| jamesdenton | that's what we do | 11:45 |
| CeeMac | jamesdenton, any thoughts on active/backup bond to independant switches having issues when the active interface goes down? | 11:45 |
| CeeMac | or is bond across independant switches effectively not supported/possible? | 11:46 |
| CeeMac | I do it this way for ESXi, so figured normal linux bond would work too (massive assumption maybe) | 11:46 |
| jamesdenton | active/passive to multiple switches should be OK. We have done that numerous times. | 11:47 |
| CeeMac | sorry, a bit off-topic for OSA | 11:47 |
| jamesdenton | when the active interface goes down... I've seen that behavior vary based on which side is acting up | 11:48 |
| CeeMac | i can't help but think this is related somehow to the issue i was seeing in VMware with virtual switches | 11:48 |
| CeeMac | its like GARP isn't working, so when the active interface goes down and the ARP timeout expires, the MAC disapears | 11:48 |
| CeeMac | the GARP from the passive interface isn't being registered | 11:49 |
| *** jawad_axd has quit IRC | 11:50 | |
| CeeMac | except for 6am today which I haven't worked out yet, its always been associate with systemd update while doing minor upgrade | 11:50 |
| *** jorti has quit IRC | 11:50 | |
| CeeMac | setup-hosts and setup-infrastructure both seem to reference lxc-hosts which upgrades systemd | 11:50 |
| CeeMac | active interface goes down, bond goes down. comes back up on the host, but not on the network | 11:51 |
| *** shyamb has quit IRC | 11:51 | |
| CeeMac | I then have to physically shut/no shut both switch ports and networking comes back on the host | 11:51 |
| *** shyamb has joined #openstack-ansible | 11:51 | |
| jrosser | CeeMac: out of interest how have you configured your host networking? networkd? | 11:51 |
| CeeMac | jrosser, netplan | 11:52 |
| jrosser | hmmmm | 11:52 |
| CeeMac | renderer: networkd | 11:52 |
| jamesdenton | brb in 20 | 11:52 |
| CeeMac | k | 11:52 |
| CeeMac | i thought I was OK with netplan, but it still feels a little flaky for want of a better term | 11:53 |
| CeeMac | really can't decide between going pure networkd or switching back to ifupdown | 11:53 |
| CeeMac | i feel ifupdown would be more stable, but networkd more futureproof? | 11:53 |
| jrosser | well. i don't know, everyone has their own opinion on that | 11:54 |
| CeeMac | yeah | 11:55 |
| *** jorti has joined #openstack-ansible | 11:55 | |
| CeeMac | that was a throw-away comment really :) | 11:55 |
| jrosser | my 2p is that i've learned the very hard way that anything that restarts the networking regardless of ifupdown/systemd unleashes a lot of brokenness | 11:55 |
| *** jawad_axd has joined #openstack-ansible | 11:55 | |
| CeeMac | oh | 11:55 |
| chkumar|ruck | odyssey4me: cloudnull any idea how to fix it http://logs.openstack.org/26/632726/12/check/tripleo-ci-centos-7-standalone-os-tempest/d22a454/job-output.txt.gz#_2019-02-19_09_46_00_505465 ? | 11:55 |
| jrosser | so anything you can do to ensure that never happens unless you really really want it to is a good thing | 11:55 |
| CeeMac | that sounds ominous | 11:55 |
| chkumar|ruck | or do I need to change the path | 11:55 |
| CeeMac | i'm not a big fan of a lot of brokenness in general | 11:56 |
| jrosser | CeeMac: as an example, i have ansible playbooks that set up the host networking, and they are designed as one-shot, so only apply changes on a fresh box | 11:57 |
| CeeMac | jrosser, i still need to build out my build process, hand building right now | 11:57 |
| jrosser | you have to explicily pass "-e yes_really_restart_networking=true" sort of variable to the ansible-playbook run to make it ever do it again | 11:57 |
| CeeMac | that sounds emininently sensible | 11:58 |
| CeeMac | are you using a pxe server plugin with ansible to manage the inventory/ o/s build prior to running your config playbooks? | 11:59 |
| CeeMac | or are you netboot/preseeding? | 11:59 |
| jrosser | pretty much, pxe server built with ansible | 11:59 |
| jrosser | using the ansible inventory to write out the tftp/dhcp config files and the preseeds per host | 12:00 |
| CeeMac | that sounds fun/complicated, I'll get around to taking a look at that sort of thing eventually | 12:00 |
| *** cshen has joined #openstack-ansible | 12:01 | |
| *** jawad_axd has quit IRC | 12:03 | |
| CeeMac | need to work on network stability first though | 12:06 |
| *** ansmith has quit IRC | 12:08 | |
| jrosser | guilhermesp: i am very close to getting magnum working, but not there yet - have you looked at magnum at all? | 12:11 |
| *** prometheanfire has quit IRC | 12:12 | |
| CeeMac | hmmm | 12:12 |
| CeeMac | could actually be an issue on the juniper swithches | 12:12 |
| *** prometheanfire has joined #openstack-ansible | 12:14 | |
| *** markvoelker has quit IRC | 12:14 | |
| jamesdenton | CeeMac what did you find on the juniper side? | 12:19 |
| CeeMac | at least for the SVI, update on garp is disabled by default | 12:20 |
| CeeMac | which would affect the gateway | 12:20 |
| CeeMac | not sure that would affect the other hosts on the same vlan though | 12:20 |
| CeeMac | will try set gratuitous-arp-reply and test later this afternoon | 12:21 |
| CeeMac | if i get time | 12:22 |
| jamesdenton | yeah, you may need to have an eye on the host and switches simultaneously, and put on your troubleshooting cap and go to town | 12:24 |
| CeeMac | yeah | 12:24 |
| CeeMac | might leave that for later in the week, got other tasks to do like build out the RC environment to actually have multiple nodes under the HA roles | 12:25 |
| CeeMac | so at least I can move things around during upgrades which would part way mitigate the problem | 12:25 |
| *** mkuf_ has joined #openstack-ansible | 12:29 | |
| *** mkuf has quit IRC | 12:32 | |
| guilhermesp | jrosser: for the moment I have a functional magnum cluster with rocky in one of our customers | 12:48 |
| guilhermesp | I think using 18.1.1. Are you facing any others issues? | 12:48 |
| jrosser | guilhermesp: sort of basic question, is that with fedora atomic or coreos? | 12:49 |
| jrosser | this in particular seems to need an update https://github.com/openstack/openstack-ansible-os_magnum/blob/master/defaults/main.yml#L146 | 12:50 |
| jrosser | as it points you to fedora 25 which seems to not go so well | 12:50 |
| *** udesale has joined #openstack-ansible | 12:55 | |
| guilhermesp | i've always using fedora-atomic. We need to use the version that matches the openstack release. I think we are using fedora-atomic-28, not sure. I can check it out | 12:56 |
| *** shyamb has quit IRC | 12:57 | |
| guilhermesp | and yes, I agree with an update of this link you sent as we have particular versions of fedora-atomic for each version | 12:57 |
| jrosser | it says 27 here https://docs.openstack.org/magnum/rocky/install/launch-instance.html#upload-the-images-required-for-your-clusters-to-the-image-service | 12:57 |
| kaiokmo | magnum should use fedora atomic 27 since rocky, IIRC | 12:57 |
| guilhermesp | yes, both of you are right jrosser and kaiokmo https://docs.openstack.org/magnum/rocky/user/#kubernetes-on-fedora-atomic | 12:58 |
| jrosser | ok so my specific issue is that on the master node i can't find where this is laid down https://github.com/openstack/magnum/blob/stable/rocky/magnum/drivers/k8s_fedora_ironic_v1/templates/kubemaster.yaml#L424 | 12:58 |
| jrosser | so it never completes (this is a different issue to internal/external endpoints) | 12:59 |
| jrosser | the magnum code looks like it should be creating a systemd service on the master node which signals completion | 12:59 |
| *** mkuf has joined #openstack-ansible | 13:00 | |
| jrosser | oh sorry, thats the wrong link, thats ironic | 13:01 |
| *** mkuf_ has quit IRC | 13:04 | |
| guilhermesp | that's odd for me... the only issue I had with magnum after fixing the heat bugs was specific network issues from the company side | 13:04 |
| jrosser | so actually thats the thing, i really understand how the overall completion of the deployment is signalled in the fedora-atomic case | 13:04 |
| jrosser | *dont | 13:04 |
| *** vnogin has joined #openstack-ansible | 13:11 | |
| *** markvoelker has joined #openstack-ansible | 13:11 | |
| openstackgerrit | Luis Filipe Domingues proposed openstack/openstack-ansible master: Fix keepalive pingable_check_script https://review.openstack.org/637839 | 13:15 |
| *** phasespace has joined #openstack-ansible | 13:20 | |
| dswebb | odyssey4me cheers for the help, finally managed to get osa to deploy the volumes containers. worth mentioning that the mnaio ceph install doesn't work out of the box. | 13:21 |
| dswebb | unless I happened to miss a flag I should have set, which is totally possible | 13:21 |
| odyssey4me | dswebb it'd be useful if you could register a bug with what didn't work, and the config you needed to add to make it work. | 13:23 |
| *** sdake has joined #openstack-ansible | 13:23 | |
| dswebb | will do | 13:23 |
| odyssey4me | Or even better, if you're able to push up a review to fix it. :) | 13:24 |
| *** vnogin has quit IRC | 13:24 | |
| dswebb | I might try at the weekend, need to hand this env over to my devs for the week | 13:24 |
| *** ArchiFleKs has quit IRC | 13:31 | |
| chkumar|ruck | jrosser: odyssey4me one query, when we create public and private network do we need to use the same network_type while creating or different one? | 13:31 |
| chkumar|ruck | at one place it is flat for public network and vxlan for private network | 13:32 |
| *** gokhani has joined #openstack-ansible | 13:32 | |
| chkumar|ruck | https://github.com/openstack/openstack-ansible-os_tempest/blob/master/defaults/main.yml#L147 and https://github.com/openstack/openstack-ansible-os_tempest/blob/master/defaults/main.yml#L135 | 13:32 |
| odyssey4me | chkumar|ruck that would depend on what you want, or have setup on the host as I understand it - whatever you set there needs to work with the host networking environment - jrosser knows that stuff better, and jamesdenton would also be of help | 13:33 |
| *** med_ has joined #openstack-ansible | 13:34 | |
| chkumar|ruck | odyssey4me: it is related to creating networks for os_tempest | 13:34 |
| jamesdenton | I'm not sure the private network is really tested. But the public network is likely flat so you can ssh via floating up | 13:36 |
| jamesdenton | *ip | 13:36 |
| jrosser | i guess the private network is implictly tested otherwise you'd not be able to ssh to the vm | 13:36 |
| jamesdenton | touche | 13:36 |
| chkumar|ruck | odyssey4me: jamesdenton jrosser https://review.openstack.org/#/c/637838/ on tripleo side we are using vlan at each place so asked | 13:38 |
| chkumar|ruck | *using flat at both place | 13:38 |
| jamesdenton | You'd need two interfaces | 13:41 |
| *** markvoelker has quit IRC | 13:43 | |
| *** ansmith has joined #openstack-ansible | 13:44 | |
| *** hamzy has joined #openstack-ansible | 13:44 | |
| *** chkumar|ruck is now known as chandankumar | 13:45 | |
| *** ArchiFleKs has joined #openstack-ansible | 13:48 | |
| *** dave-mccowan has joined #openstack-ansible | 13:49 | |
| *** sdake has quit IRC | 14:05 | |
| jamesdenton | chandankumar Are you trying to solve for issues with multi-node testing? Curious why the change from vxlan to flat | 14:09 |
| *** sdake has joined #openstack-ansible | 14:13 | |
| *** masterpe has joined #openstack-ansible | 14:13 | |
| kaiokmo | jrosser: I think maybe https://github.com/openstack/magnum/blob/stable/rocky/magnum/drivers/common/templates/kubernetes/fragments/wc-notify-master.sh#L14-L22 is what you are looking for | 14:14 |
| *** sdake has quit IRC | 14:19 | |
| *** phasespace has quit IRC | 14:31 | |
| *** sdake has joined #openstack-ansible | 14:37 | |
| *** markvoelker has joined #openstack-ansible | 14:41 | |
| *** sdake has quit IRC | 14:53 | |
| *** udesale has quit IRC | 14:53 | |
| chandankumar | jamesdenton: there we are doing single node testing | 14:58 |
| chandankumar | jamesdenton: here is the ara logs http://logs.openstack.org/60/634060/12/check/tripleo-ci-centos-7-standalone-os-tempest/2981018/logs/ara_oooq/result/11f3eab6-8fdf-4cda-afb6-8de4ff78bafd/ | 14:59 |
| *** sdake has joined #openstack-ansible | 15:00 | |
| cloudnull | mornings all | 15:01 |
| cloudnull | anyone want to give https://review.openstack.org/#/c/588959/ a shove | 15:01 |
| *** vnogin has joined #openstack-ansible | 15:02 | |
| jamesdenton | chandankumar Unless I'm wrong, in os_tempest, the 'flat' label maps to eth12, and you'd only get one flat network out of that. You'd have to build a separate mapping (flat2:eth13) and then have to specify the label on each network (public and private). | 15:04 |
| chandankumar | jamesdenton: not sure 2019-02-19 08:13:14.027 ERROR /var/log/extra/errors.txt: ERROR /var/log/containers/neutron/server.log: 30 ERROR ovsdbapp.event [-] Unexpected exception in notify_loop: KeyError: 'neutron:ovn-metadata-sb-cfg' | 15:04 |
| jamesdenton | as for that error, do you happen to have access to the configuration of ml2_conf.ini from that build? | 15:04 |
| chandankumar | 2019-02-19 08:13:14.027 ERROR /var/log/extra/errors.txt: ERROR /var/log/containers/neutron/server.log: 30 ERROR ovsdbapp.event Traceback (most recent call last): | 15:04 |
| chandankumar | 2019-02-19 08:13:14.027 ERROR /var/log/extra/errors.txt: ERROR /var/log/containers/neutron/server.log: 30 ERROR ovsdbapp.event File "/usr/lib/python2.7/site-packages/ovsdbapp/event.py", line 137, in notify_loop | 15:04 |
| chandankumar | 2019-02-19 08:13:14.027 ERROR /var/log/extra/errors.txt: ERROR /var/log/containers/neutron/server.log: 30 ERROR ovsdbapp.event match.run(event, row, updates) | 15:04 |
| chandankumar | 2019-02-19 08:13:14.027 ERROR /var/log/extra/errors.txt: ERROR /var/log/containers/neutron/server.log: 30 ERROR ovsdbapp.event File "/usr/lib/python2.7/site-packages/networking_ovn/ovsdb/ovsdb_monitor.py", line 98, in run | 15:04 |
| chandankumar | 2019-02-19 08:13:14.027 ERROR /var/log/extra/errors.txt: ERROR /var/log/containers/neutron/server.log: 30 ERROR ovsdbapp.event self._metadata_nb_cfg(row)) | 15:04 |
| chandankumar | 2019-02-19 08:13:14.027 ERROR /var/log/extra/errors.txt: ERROR /var/log/containers/neutron/server.log: 30 ERROR ovsdbapp.event File "/usr/lib/python2.7/site-packages/networking_ovn/ovsdb/ovsdb_monitor.py", line 86, in _metadata_nb_cfg | 15:04 |
| chandankumar | 2019-02-19 08:13:14.027 ERROR /var/log/extra/errors.txt: ERROR /var/log/containers/neutron/server.log: 30 ERROR ovsdbapp.event return int(row.external_ids[ovn_const.OVN_AGENT_METADATA_SB_CFG_KEY]) | 15:04 |
| jamesdenton | which plugin are you testing? | 15:04 |
| chandankumar | 2019-02-19 08:13:14.027 ERROR /var/log/extra/errors.txt: ERROR /var/log/containers/neutron/server.log: 30 ERROR ovsdbapp.event KeyError: 'neutron:ovn-metadata-sb-cfg' | 15:04 |
| chandankumar | sorry | 15:04 |
| chandankumar | jamesdenton: http://logs.openstack.org/60/634060/12/check/tripleo-ci-centos-7-standalone-os-tempest/2981018/logs/undercloud/var/log/extra/errors.txt.gz#_2019-02-19_08_13_14_027 | 15:04 |
| chandankumar | jamesdenton: this might be the issue | 15:04 |
| jamesdenton | If you're testing OVN, then vxlan definitely isn't supported. It would need to be GENEVE | 15:04 |
| chandankumar | jamesdenton: they have recently migrating to ovn | 15:05 |
| *** vnogin has quit IRC | 15:06 | |
| jamesdenton | in that case, consider setting tempest_private_net_provider_type to geneve and see how that goes | 15:06 |
| *** shananigans has joined #openstack-ansible | 15:10 | |
| chandankumar | jamesdenton: thanks, updated that | 15:12 |
| *** markvoelker has quit IRC | 15:14 | |
| openstackgerrit | Merged openstack/openstack-ansible-ops master: Update grafana, use vendored role, and add es lb https://review.openstack.org/637672 | 15:22 |
| openstackgerrit | Chandan Kumar proposed openstack/openstack-ansible-os_tempest master: Set tempest_private_net_provider_type to geneve as default https://review.openstack.org/637875 | 15:22 |
| chandankumar | jamesdenton: as per dalverz, we can set geneve as a default in os_tempest | 15:23 |
| jamesdenton | Are you using a forked branch? | 15:24 |
| chandankumar | jamesdenton: nope | 15:24 |
| jamesdenton | the default today is vxlan: https://github.com/openstack/openstack-ansible-os_tempest/blob/master/defaults/main.yml#L135. So I'm guessing you just mean override that on your side? | 15:25 |
| chandankumar | jamesdenton: http://codesearch.openstack.org/?q=tempest_private_net_provider_type&i=nope&files=.yml&repos= | 15:26 |
| jamesdenton | yes | 15:27 |
| chandankumar | in OSA and os_tempest is vxlan by defalt | 15:27 |
| jamesdenton | yes | 15:28 |
| chandankumar | jamesdenton: let me comment on review itself on tripleo side | 15:28 |
| jamesdenton | you mentioned changing the default to geneve, which would break the OSA gate today | 15:28 |
| *** lucasagomes has joined #openstack-ansible | 15:28 | |
| jamesdenton | thus the suggestion to override | 15:28 |
| *** sdake has quit IRC | 15:28 | |
| chandankumar | lucasagomes: Hello | 15:29 |
| chandankumar | lucasagomes: on osa side http://codesearch.openstack.org/?q=tempest_private_net_provider_type&i=nope&files=.yml&repos= vxlan is default, so changing in os_tempest will break it | 15:30 |
| chandankumar | jamesdenton: https://review.openstack.org/#/c/637838/2 | 15:30 |
| lucasagomes | chandankumar, hi there, lemme take a look | 15:31 |
| lucasagomes | chandankumar, ok, hmm I have to check how we can proceed then because vxlan is not a valid net type in OVN | 15:33 |
| *** sdake has joined #openstack-ansible | 15:35 | |
| *** sreejithp has joined #openstack-ansible | 15:37 | |
| *** vnogin has joined #openstack-ansible | 15:37 | |
| openstackgerrit | Jonathan Rosser proposed openstack/openstack-ansible-os_magnum master: Update fedora-atomic image to version 27 following upstream docs https://review.openstack.org/637912 | 15:42 |
| *** cmart has joined #openstack-ansible | 15:42 | |
| *** weshay has joined #openstack-ansible | 15:51 | |
| weshay | 0/ | 15:52 |
| *** sdake has quit IRC | 15:52 | |
| *** dalvarez has joined #openstack-ansible | 15:53 | |
| dalvarez | hey | 15:53 |
| dalvarez | chandankumar: | 15:53 |
| *** vnogin has quit IRC | 15:53 | |
| chandankumar | dalvarez: lucasagomes is looking into it | 15:54 |
| mnaser | bonjour everyone | 15:54 |
| *** cshen has quit IRC | 15:55 | |
| *** sdake has joined #openstack-ansible | 15:58 | |
| openstackgerrit | Merged openstack/openstack-ansible-ops master: Add monitorstack data collection into ES https://review.openstack.org/637455 | 15:59 |
| mnaser | mass ping alert | 16:00 |
| mnaser | cloudnull, DimGR, d34dh0r53, hughsaunders, b3rnard0, palendae, odyssey4me, serverascode, rromans, erikmwilson, mancdaz, _shaps_, BjoernT, claco, echiu, dstanek, jwagner, ayoung, prometheanfire, evrardjp, arbrandes, scarlisle, luckyinva, ntt, javeriak, spotz, vdo, jmccrory, alextricity25, jasondotstar, admin0, michaelgugino, ametts, bgmccollum, darrenc, JRobinson__, colinmcnamara, thorst, adreznec, eil397, | 16:00 |
| mnaser | qwang,nishpatwa_, cathrichardson, drifterza, hwoarang, cshen, ullbeking, mnaser, nicolasbock, jrosser, cjloader, antonym, dcdamien, jamesdenton | 16:00 |
| mnaser | #startmeeting openstack_ansible_meeting | 16:00 |
| openstack | Meeting started Tue Feb 19 16:00:14 2019 UTC and is due to finish in 60 minutes. The chair is mnaser. Information about MeetBot at http://wiki.debian.org/MeetBot. | 16:00 |
| openstack | Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. | 16:00 |
| *** openstack changes topic to " (Meeting topic: openstack_ansible_meeting)" | 16:00 | |
| openstack | The meeting name has been set to 'openstack_ansible_meeting' | 16:00 |
| guilhermesp | o/ | 16:00 |
| mnaser | #topic rollcall | 16:00 |
| *** openstack changes topic to "rollcall (Meeting topic: openstack_ansible_meeting)" | 16:00 | |
| mnaser | o/ | 16:00 |
| evrardjp | o/ | 16:00 |
| admin0 | \o | 16:00 |
| chandankumar | \o/ | 16:00 |
| jrosser | hello! | 16:00 |
| fnpanic | hi | 16:00 |
| *** ianychoi has quit IRC | 16:00 | |
| mnaser | #topic highlights from last week | 16:02 |
| *** openstack changes topic to "highlights from last week (Meeting topic: openstack_ansible_meeting)" | 16:02 | |
| mnaser | anyone want to start sharing some of the work they did last week ? :) | 16:02 |
| prometheanfire | o/ | 16:02 |
| *** sdake has quit IRC | 16:03 | |
| guilhermesp | I think the bug squash days were a success. We should keep the mood | 16:03 |
| mnaser | i think so too, it was awesome, we burned down through so much stuff | 16:04 |
| mnaser | unfortunately i couldnt participate as much as i would have but hey that's life | 16:04 |
| evrardjp | sorry I couldn't attend to that | 16:04 |
| evrardjp | it's been epic here | 16:04 |
| jrosser | the bug day was great | 16:05 |
| mnaser | i think it was just fun as well because we got to hack on plenty of things and i twas pretty feel-good overall :p | 16:05 |
| evrardjp | other highlights is that releasing will be easier in the future | 16:05 |
| evrardjp | mnaser: damn you're hitting my FOMO nerve | 16:05 |
| guilhermesp | are we going to keep the idea? | 16:05 |
| mnaser | guilhermesp: i think maybe one every few weeks or so is better than like | 16:05 |
| mnaser | a bug triage every meeting | 16:06 |
| guilhermesp | sounds good | 16:06 |
| mnaser | i dunno, thoughts? | 16:06 |
| guilhermesp | I think once a month and keep the bug triage every meeting | 16:06 |
| mnaser | yeah maybe we could do that | 16:07 |
| evrardjp | guilhermesp: some bugs need fast triaging | 16:07 |
| mnaser | other than that, any other last week updates to talk about? | 16:07 |
| jrosser | i have been digging at heat/magnum after a few folks had trouble with all the endpoint stuff | 16:08 |
| guilhermesp | yes evrardjp that's why is important to keep in the meeting too every week | 16:08 |
| evrardjp | guilhermesp: sorry I don't know why I pinged you.. We were agreeing with each other :) | 16:08 |
| guilhermesp | evrardjp: np! | 16:08 |
| guilhermesp | jrosser: I saw you asked a question in openstack-containers | 16:09 |
| jrosser | yes, it's all a bit quiet in there | 16:09 |
| guilhermesp | but no reply till now | 16:09 |
| guilhermesp | maybe tag strigazi | 16:09 |
| mnaser | jrosser: can i help? | 16:09 |
| mnaser | i have contributed a bunch to magnum | 16:09 |
| mnaser | i can help make the appropriate fixes if they need to be done | 16:09 |
| jrosser | i'm sure, can we catch up after the meeting, i've found a few things i don't understand in the code | 16:10 |
| mnaser | sure | 16:10 |
| mnaser | happy to | 16:10 |
| jrosser | i think i've found a bug in heat about the endpoints too | 16:10 |
| odyssey4me | o/ | 16:10 |
| mnaser | ok, i'd be happy to ping the right people to get those fixed up | 16:10 |
| chandankumar | jrosser: http://logs.openstack.org/94/630694/13/check/openstack-ansible-deploy-aio_metal_heat-centos-7/bb028ec/job-output.txt.gz#_2019-02-19_09_22_50_865174 | 16:11 |
| *** markvoelker has joined #openstack-ansible | 16:11 | |
| chandankumar | jrosser: is it somehthing like that on heat | 16:12 |
| *** ianychoi has joined #openstack-ansible | 16:12 | |
| chandankumar | i can discuss after meeting | 16:12 |
| mnaser | cool cool | 16:12 |
| mnaser | #topic open discussion | 16:13 |
| *** openstack changes topic to "open discussion (Meeting topic: openstack_ansible_meeting)" | 16:13 | |
| mnaser | i've been slowly trying to make this more office hoursy and looks like we have more .. people showing up and sticking around :) | 16:13 |
| mnaser | PTG is approaching, anyone wanna volunteer setup the initial etherpad? :) | 16:14 |
| guilhermesp | If I have a previous layout to be based on I can start this new etherpad mnaser | 16:15 |
| mnaser | guilhermesp: https://etherpad.openstack.org/p/osa-stein-ptg => osa-train-ptg ? | 16:15 |
| cloudnull | o/ | 16:16 |
| * cloudnull late but here | 16:16 | |
| mnaser | #action guilhermesp create ptg etherpad | 16:17 |
| mnaser | any other subjects that we wanna bring up? maybe someone can help finish the last steps of aio nspawn in gate so we can maybe do the switch this cycle? | 16:17 |
| guilhermesp | ok mnaser | 16:17 |
| mnaser | we're running out of time | 16:17 |
| chandankumar | mnaser: are we going to built container of tempest on nspawn? | 16:20 |
| odyssey4me | chandankumar we have all the bits in place, but there are a few outstanding issues with it which need addressing | 16:22 |
| odyssey4me | The general plan is to move OSA to use nspawn instead of LXC for machine containers. | 16:22 |
| guilhermesp | https://etherpad.openstack.org/p/osa-train-ptg WIP | 16:22 |
| chandankumar | odyssey4me: is it possible to run docker temepst container with nspawn? as tripleo uses kolla container so | 16:22 |
| cloudnull | chandankumar: docker no. but app container yes. | 16:23 |
| odyssey4me | chandankumar it is already, yes - we already do that with lxc - however, we use it as a machine container so we add systemd into it | 16:23 |
| *** dswebb has quit IRC | 16:23 | |
| cloudnull | systemd-nspawn is the basis for rkt, which can run application containers. | 16:23 |
| odyssey4me | as cloudnull said, we could move to use an app container model | 16:24 |
| odyssey4me | but that's a fair way off at this stage | 16:24 |
| cloudnull | ^ | 16:24 |
| chandankumar | tempest is just an app, not a service, so I think it will work | 16:24 |
| chandankumar | ok | 16:24 |
| odyssey4me | we have a *lot* of assumptions around the services running on a full host, so the app container model does not work with OSA at this time | 16:24 |
| cloudnull | that said, chandankumar if you have cycles to poke at it I'd be happy to help where I can | 16:24 |
| evrardjp | I am not sure we should outright kill that idea, it's just a binary blob after all :p | 16:25 |
| chandankumar | cloudnull: we are thinking to work on container support may be next cycle | 16:25 |
| cloudnull | sadly my time to focus on a lot of those things has largely evaporated | 16:25 |
| evrardjp | but we are basically alone though | 16:25 |
| chandankumar | on os_tempest side | 16:25 |
| odyssey4me | yes, doing app containers is absolutely something we've talked about, but no-one has had the cycles to put into it | 16:25 |
| odyssey4me | chandankumar so if you folks have cycles to figure out a model, that'd be fantastic | 16:26 |
| cloudnull | +1 | 16:26 |
| chandankumar | sure! | 16:26 |
| jamesdenton | folks, i have to bounce out for a bit. bbiaf | 16:27 |
| cloudnull | chao | 16:27 |
| *** cmart has quit IRC | 16:31 | |
| *** cshen has joined #openstack-ansible | 16:33 | |
| odyssey4me | FYI I've made some more progress on the python build simplification - just ironing out a few things before releasing them for review. Unfortunately my changes break the integrated build, so a few tweaks are still needed. | 16:33 |
| odyssey4me | It's an issue with stackviz building left, which I just need a little more time on to get done. | 16:35 |
| odyssey4me | Heh, I see the issue and should have a patch up to solve it a bit later. | 16:37 |
| chandankumar | jrosser: still around? | 16:40 |
| *** cshen has quit IRC | 16:40 | |
| jrosser | yes i am | 16:40 |
| chandankumar | jrosser: http://logs.openstack.org/94/630694/13/check/openstack-ansible-deploy-aio_metal_heat-centos-7/bb028ec/job-output.txt.gz#_2019-02-19_09_22_50_865174 | 16:40 |
| mnaser | odyssey4me: cool, when we're almost done, it would be good to get some info on that so we can replicate things across | 16:41 |
| chandankumar | jrosser: can you take a look at this failure, it is related to heat endpoint | 16:41 |
| jrosser | chandankumar: it is the same as other folks were seeing | 16:41 |
| jrosser | after the meeting we can discuss some more | 16:41 |
| chandankumar | jrosser: https://github.com/openstack/heat-tempest-plugin/blob/9ca1fc2c55bceba078a6066d6d81447b1c8644d3/heat_tempest_plugin/tests/api/test_heat_api.py#L57 | 16:43 |
| chandankumar | jrosser: sure | 16:43 |
| *** markvoelker has quit IRC | 16:44 | |
| mnaser | I think finding some more time to do reviews would be good too | 16:45 |
| *** macza has joined #openstack-ansible | 16:47 | |
| mnaser | i need to do more as well | 16:49 |
| jrosser | it would be great to get more of us on the case - we've all got different experience to bring | 16:49 |
| *** hamzy has quit IRC | 16:50 | |
| odyssey4me | mnaser yep, I'll get the standard AIO done - then work up an etherpad for the rest for everyone to climb in while I shift focus to getting the AIO used for all tests | 16:54 |
| mnaser | odyssey4me: awesome | 16:54 |
| chandankumar | odyssey4me: is it possible to clone only those repo which are needed for a particular scenario? | 16:54 |
| chandankumar | in AIO | 16:54 |
| odyssey4me | chandankumar for now, no - but I have some ideas around how we could achieve that which we can discuss at the PTG, assuming I'm able to get there - unfortunately my travel is not yet approved | 16:56 |
| * chandankumar is not going due to us visa issue :-) | 16:56 | |
| odyssey4me | chandankumar :( At the next one then. | 16:56 |
| chandankumar | odyssey4me: yes for sure :-) | 16:57 |
| *** pcaruana has quit IRC | 16:57 | |
| *** chandankumar is now known as kmrchdn | 17:12 | |
| odyssey4me | mnaser did I miss the end of the meeting? | 17:13 |
| mnaser | #endmeeting | 17:14 |
| mnaser | no | 17:14 |
| *** openstack changes topic to "Launchpad: https://launchpad.net/openstack-ansible || Weekly Meetings: https://wiki.openstack.org/wiki/Meetings/openstack-ansible || Review Dashboard: http://bit.ly/2xA1eZC" | 17:14 | |
| openstack | Meeting ended Tue Feb 19 17:14:01 2019 UTC. Information about MeetBot at http://wiki.debian.org/MeetBot . (v 0.1.4) | 17:14 |
| mnaser | :) | 17:14 |
| openstack | Minutes: http://eavesdrop.openstack.org/meetings/openstack_ansible_meeting/2019/openstack_ansible_meeting.2019-02-19-16.00.html | 17:14 |
| openstack | Minutes (text): http://eavesdrop.openstack.org/meetings/openstack_ansible_meeting/2019/openstack_ansible_meeting.2019-02-19-16.00.txt | 17:14 |
| odyssey4me | hahaha | 17:14 |
| openstack | Log: http://eavesdrop.openstack.org/meetings/openstack_ansible_meeting/2019/openstack_ansible_meeting.2019-02-19-16.00.log.html | 17:14 |
| *** lucasagomes is now known as lucas-afk | 17:20 | |
| *** lucas-afk is now known as lucasagomes | 17:20 | |
| *** miloa has quit IRC | 17:21 | |
| jrosser | mnaser: got a minute for heat/magnum things? | 17:38 |
| mnaser | jrosser: sure, gimme 2 mins, do you wanna hop on zoom or something higher bandwidth to make it quicker/easier to explain? | 17:39 |
| *** markvoelker has joined #openstack-ansible | 17:41 | |
| *** cmart has joined #openstack-ansible | 17:41 | |
| *** mma has quit IRC | 17:42 | |
| jrosser | yes - i have about 20mins till i need to travel | 17:42 |
| *** DanyC_ has joined #openstack-ansible | 17:43 | |
| mnaser | jrosser: im ready if you'd like | 17:46 |
| *** DanyC has quit IRC | 17:47 | |
| mnaser | jrosser: https://zoom.us/j/330771324 if thats easier? | 17:47 |
| mnaser | unless you prefer ascii :) | 17:47 |
| * jrosser tries the link | 17:47 | |
| *** kopecmartin is now known as kopecmartin|off | 17:53 | |
| *** markvoelker has quit IRC | 18:14 | |
| *** sreejithp has quit IRC | 18:14 | |
| *** sreejithp has joined #openstack-ansible | 18:14 | |
| *** hamzy has joined #openstack-ansible | 18:17 | |
| kmrchdn | mnaser: anything we can improve in os_tempest report? | 18:26 |
| kmrchdn | it appears kind of boring, when I put all the all the reviews there | 18:26 |
| *** kmrchdn is now known as chandankumar | 18:26 | |
| chandankumar | as none wants to read long emails, | 18:26 |
| chandankumar | what about adding stuff which I learned each week? | 18:27 |
| *** hamzaachi has quit IRC | 18:27 | |
| mnaser | chandankumar: i really enjoy them tbh, i think they do a good update and i wish i can get myself to do the same for OSA | 18:33 |
| mnaser | i read through every single one and i find them informative, but maybe we can have some goals in there and then update each week about these goals and targets ? | 18:33 |
| chandankumar | mnaser: may we can reproduce a newletter as keystone or zuul does? | 18:33 |
| chandankumar | mnaser: open an etherpad people will highlight the stuff and we can can send it together | 18:34 |
| mnaser | chandankumar: yes, i think the neat thing that zuul does is they maintain an etherpad where they always add things into | 18:34 |
| mnaser | yeayh | 18:34 |
| chandankumar | mnaser: glad you liked the status update :-) | 18:34 |
| odyssey4me | chandankumar I like it too, now that I have fixed my access to the ML. :) | 18:36 |
| chandankumar | odyssey4me: :-) | 18:36 |
| openstackgerrit | Cam J. Loader (cjloader) proposed openstack/openstack-ansible master: Add iputils to utility container https://review.openstack.org/637970 | 18:46 |
| cjloader | odyssey4me: mind reviewing ^? | 18:50 |
| odyssey4me | chandankumar cjloader given that https://github.com/openstack/openstack-ansible-os_tempest/commit/f30f002d4b5db1cb30a0b7d669ba19e774e99e8e added a dependency on ping (because the ping test is on by default), then surely the os_tempest role should install the requisite package for it? | 18:53 |
| odyssey4me | chandankumar alternatively, perhaps the default should be to disable that test? | 18:54 |
| chandankumar | odyssey4me: yes it should install required packages | 18:55 |
| chandankumar | odyssey4me: we have added ping stuff for debugging in lsx containers specifically | 18:55 |
| chandankumar | *lxc | 18:55 |
| odyssey4me | chandankumar I personally prefer to see dependencies implemented in the roles that depend on them. It helps with maintenance later, should that dependency ever go away. | 18:56 |
| prometheanfire | odyssey4me: is there some kind of gathering that takes place so that all the deps can be installed at once (union of nova/foo/bar/tempest)? | 18:58 |
| odyssey4me | prometheanfire no, ansible doesn't work that way unfortunatey | 18:58 |
| prometheanfire | odyssey4me: that's what I thought, not too big of a deal though | 18:59 |
| odyssey4me | execution is task by task - it is very simple that way, even if it is very slow | 18:59 |
| prometheanfire | yep | 19:00 |
| prometheanfire | registering vars would only work for following roles | 19:01 |
| cjloader | so chandankumar odyssey4me prometheanfire are we agreeing it should go in the role? | 19:03 |
| prometheanfire | the tempest role? yes | 19:03 |
| *** electrofelix has quit IRC | 19:03 | |
| odyssey4me | cjloader yep | 19:04 |
| cjloader | okay | 19:04 |
| odyssey4me | prometheanfire can you propose the backports for https://review.openstack.org/636388 on top of https://review.openstack.org/#/q/Id4872caed2cba6b2ef31e850cde63a2a88d55250 please? | 19:07 |
| prometheanfire | odyssey4me: without looking, pike? | 19:07 |
| prometheanfire | nope | 19:07 |
| odyssey4me | prometheanfire well, with the other patches, the suplicate section is now present in all the other branches where those backports were proposed | 19:07 |
| odyssey4me | *duplicate | 19:08 |
| prometheanfire | ah, thought I proposed it, will do then | 19:08 |
| odyssey4me | and I'm guessing you're going to want to port https://review.openstack.org/637210 back too once it merges, and that will need to be on top of the backports of https://review.openstack.org/636388 | 19:08 |
| odyssey4me | awesome, thanks | 19:08 |
| odyssey4me | I'd really prefer not to have the duplicate section make it into a release by mistake. | 19:09 |
| *** DanyC_ has quit IRC | 19:09 | |
| prometheanfire | odyssey4me: I might just wait for stuff to merge since they are all workflowed back to pike | 19:09 |
| prometheanfire | and ya, that'd be nice to backport but as you noted it still technically works | 19:09 |
| prometheanfire | it's formatting only | 19:10 |
| *** markvoelker has joined #openstack-ansible | 19:11 | |
| odyssey4me | I think it'd be good to do - it reduces confusion when trying to figure out why something's not working. | 19:11 |
| prometheanfire | yep | 19:11 |
| openstackgerrit | Kevin Carter (cloudnull) proposed openstack/openstack-ansible-ops master: Ensure upper and lower memory limits are whole numbers https://review.openstack.org/637975 | 19:12 |
| odyssey4me | OK folks, I'm out for the night - and possibly the week too. I've got to sand and paint my kitchen this week, but I may check in from time to time to give my tired arms a break. ;) | 19:15 |
| cjloader | wait odyssey4me | 19:15 |
| cjloader | why shouldn't the utility container have the ability to ping | 19:16 |
| prometheanfire | odyssey4me: nn, sanding sucks, good luck | 19:16 |
| cjloader | (from d34dh0r53) | 19:16 |
| prometheanfire | use machines as you can | 19:16 |
| odyssey4me | cjloader d34dh0r53 that's not the point - the point is that the dependence on ping is in the tempest role... if someone would like to ping from the utility container in general, then they can override the utility container package list or we can add another var for extra things... we've done a ton of work this cycle to reduce unnecessary package installation which slow down the deployment and increase the bloat | 19:18 |
| odyssey4me | OSA only needs to deploy what it requires - anything additional desired by a deployer can be implemented afterwards | 19:18 |
| odyssey4me | of course we can also implement something which is opt-in which adds extra nice-to-have packages to the utility container... but then the gate doesn't need it, so it won't opt-in | 19:20 |
| openstackgerrit | Kevin Carter (cloudnull) proposed openstack/openstack-ansible-ops master: Add dynamic group detection to openstack metrics https://review.openstack.org/637980 | 19:26 |
| *** hamzaachi has joined #openstack-ansible | 19:33 | |
| *** sdake has joined #openstack-ansible | 19:34 | |
| *** mgariepy has quit IRC | 19:40 | |
| *** markvoelker has quit IRC | 19:44 | |
| cjloader | well if the gates deploy os_tempest.....its a reqd pkg | 19:46 |
| *** DanyC has joined #openstack-ansible | 19:49 | |
| openstackgerrit | Michael Vollman proposed openstack/openstack-ansible-os_nova master: Avoid distro installing unused services https://review.openstack.org/633275 | 19:51 |
| *** sdake has quit IRC | 19:52 | |
| *** sdake has joined #openstack-ansible | 19:55 | |
| *** DanyC has quit IRC | 19:56 | |
| *** DanyC has joined #openstack-ansible | 19:57 | |
| *** sdake has quit IRC | 19:58 | |
| *** openstackgerrit has quit IRC | 20:09 | |
| *** alvinstarr has quit IRC | 20:25 | |
| *** DanyC has quit IRC | 20:28 | |
| *** DanyC has joined #openstack-ansible | 20:28 | |
| *** openstackgerrit has joined #openstack-ansible | 20:34 | |
| openstackgerrit | Kevin Carter (cloudnull) proposed openstack/openstack-ansible-ops master: Correct service name when running with upstart https://review.openstack.org/637992 | 20:34 |
| openstackgerrit | Kevin Carter (cloudnull) proposed openstack/openstack-ansible-ops master: Add dynamic group detection to openstack metrics https://review.openstack.org/637980 | 20:35 |
| openstackgerrit | Kevin Carter (cloudnull) proposed openstack/openstack-ansible-ops master: Add dynamic group detection to openstack metrics https://review.openstack.org/637980 | 20:35 |
| openstackgerrit | Merged openstack/openstack-ansible-os_nova stable/queens: Update the pci config for nova. https://review.openstack.org/636397 | 20:35 |
| openstackgerrit | Merged openstack/openstack-ansible-os_nova stable/rocky: Update the pci config for nova. https://review.openstack.org/636391 | 20:35 |
| openstackgerrit | Merged openstack/openstack-ansible-os_nova master: update pci_passthrough template variables https://review.openstack.org/637210 | 20:35 |
| *** DanyC has quit IRC | 20:38 | |
| *** markvoelker has joined #openstack-ansible | 20:41 | |
| *** DanyC has joined #openstack-ansible | 20:43 | |
| *** coolj has joined #openstack-ansible | 20:44 | |
| openstackgerrit | Matthew Thode proposed openstack/openstack-ansible-os_nova stable/rocky: update pci_passthrough template variables https://review.openstack.org/638003 | 20:52 |
| openstackgerrit | Matthew Thode proposed openstack/openstack-ansible-os_nova stable/queens: update pci_passthrough template variables https://review.openstack.org/638004 | 20:54 |
| openstackgerrit | Matthew Thode proposed openstack/openstack-ansible-os_nova stable/pike: update pci_passthrough template variables https://review.openstack.org/638006 | 20:56 |
| *** DanyC has quit IRC | 20:57 | |
| prometheanfire | odyssey4me: all backported ^ https://review.openstack.org/#/q/is:open+project:openstack/openstack-ansible-os_nova+topic:pci-passthrough | 20:58 |
| openstackgerrit | Merged openstack/openstack-ansible-ops master: Ensure upper and lower memory limits are whole numbers https://review.openstack.org/637975 | 21:06 |
| *** markvoelker has quit IRC | 21:13 | |
| *** hamzaachi has quit IRC | 21:20 | |
| openstackgerrit | Shannon Mitchell proposed openstack/openstack-ansible-os_tempest master: Update workspace tempest.conf on changes https://review.openstack.org/638014 | 21:25 |
| openstackgerrit | Jacob Wagner proposed openstack/openstack-ansible-ops master: Add ability to deploy designate with BIND9 servers https://review.openstack.org/635611 | 21:28 |
| *** ArchiFleKs has quit IRC | 21:32 | |
| *** ansmith has quit IRC | 21:34 | |
| openstackgerrit | Merged openstack/openstack-ansible-ops master: Add dynamic group detection to openstack metrics https://review.openstack.org/637980 | 21:35 |
| openstackgerrit | Kevin Carter (cloudnull) proposed openstack/openstack-ansible-ops master: Correct service name when running with upstart https://review.openstack.org/637992 | 21:39 |
| *** hamzy has quit IRC | 21:41 | |
| *** devx has quit IRC | 21:47 | |
| *** ArchiFleKs has joined #openstack-ansible | 21:51 | |
| *** yetiszaf has quit IRC | 21:52 | |
| openstackgerrit | Cam J. Loader (cjloader) proposed openstack/openstack-ansible master: Add iputils to utility container https://review.openstack.org/637970 | 21:54 |
| cjloader | d34dh0r53: noonedeadpunk ^ | 21:54 |
| jrosser | i thought that the conclusion was that dependencies of tempest should go in the tempest role | 21:58 |
| *** sdake has joined #openstack-ansible | 21:59 | |
| jrosser | and the package names are not the same on all distros either afaik | 21:59 |
| *** DanyC has joined #openstack-ansible | 22:03 | |
| openstackgerrit | Merged openstack/openstack-ansible-os_nova stable/pike: Update deprecated option for pci passthrough https://review.openstack.org/636398 | 22:04 |
| *** dave-mccowan has quit IRC | 22:07 | |
| openstackgerrit | Shannon Mitchell proposed openstack/openstack-ansible-os_tempest master: Update workspace tempest.conf on changes https://review.openstack.org/638014 | 22:07 |
| *** DanyC has quit IRC | 22:07 | |
| *** sdake has quit IRC | 22:09 | |
| openstackgerrit | Shannon Mitchell proposed openstack/openstack-ansible-os_tempest master: Update workspace tempest.conf on changes https://review.openstack.org/638014 | 22:10 |
| *** markvoelker has joined #openstack-ansible | 22:10 | |
| openstackgerrit | Merged openstack/openstack-ansible-os_nova stable/pike: Update the pci config for nova. https://review.openstack.org/636399 | 22:12 |
| cjloader | well, d34dh0r53 had a question | 22:12 |
| cjloader | so i'm updating the PR until resolved | 22:13 |
| cjloader | jrosser | 22:13 |
| openstackgerrit | Shannon Mitchell proposed openstack/openstack-ansible-os_tempest master: Update workspace tempest.conf on changes https://review.openstack.org/638014 | 22:16 |
| *** phasespace has joined #openstack-ansible | 22:17 | |
| *** DanyC has joined #openstack-ansible | 22:23 | |
| openstackgerrit | Merged openstack/openstack-ansible-ops master: Correct service name when running with upstart https://review.openstack.org/637992 | 22:27 |
| openstackgerrit | Shannon Mitchell proposed openstack/openstack-ansible-os_tempest stable/rocky: Update workspace tempest.conf on changes https://review.openstack.org/638030 | 22:38 |
| *** markvoelker has quit IRC | 22:44 | |
| openstackgerrit | Shannon Mitchell proposed openstack/openstack-ansible-os_tempest stable/queens: Update workspace tempest.conf on changes https://review.openstack.org/638032 | 22:45 |
| openstackgerrit | Shannon Mitchell proposed openstack/openstack-ansible-os_tempest stable/pike: Update workspace tempest.conf on changes https://review.openstack.org/638036 | 22:48 |
| *** vnogin has joined #openstack-ansible | 23:05 | |
| *** shananigans has quit IRC | 23:06 | |
| *** vnogin has quit IRC | 23:09 | |
| *** ansmith has joined #openstack-ansible | 23:14 | |
| *** sreejithp has quit IRC | 23:16 | |
| prometheanfire | just want to make sure, but is it ever planned for OSA to support running one distro's container on another distro's host? | 23:34 |
| *** markvoelker has joined #openstack-ansible | 23:41 | |
| *** tosky has quit IRC | 23:58 | |
Generated by irclog2html.py 2.15.3 by Marius Gedminas - find it at mg.pov.lt!