| *** aaronsheffield has quit IRC | 00:20 | |
| *** sthussey has quit IRC | 01:03 | |
| *** ianychoi has quit IRC | 01:20 | |
| openstackgerrit | Matthew Heler proposed openstack/openstack-helm-infra master: Increase the cpu and memory resource limits for Ceph OSDs https://review.openstack.org/623617 | 01:27 |
|---|---|---|
| *** cfriesen has quit IRC | 02:38 | |
| *** jamesgu_ has quit IRC | 02:41 | |
| *** jistr has quit IRC | 02:42 | |
| *** jistr has joined #openstack-helm | 02:50 | |
| *** jamesgu_ has joined #openstack-helm | 03:14 | |
| *** cfriesen has joined #openstack-helm | 03:42 | |
| openstackgerrit | Tin Lam proposed openstack/openstack-helm-infra master: Add securityContext helm-toolkit function https://review.openstack.org/613611 | 03:59 |
| openstackgerrit | Tin Lam proposed openstack/openstack-helm master: Add startingDeadlineSeconds field to cronJobs https://review.openstack.org/609859 | 04:03 |
| openstackgerrit | Tin Lam proposed openstack/openstack-helm-infra master: Add Egress Helm-toolkit function https://review.openstack.org/612506 | 04:04 |
| openstackgerrit | Cliff Parsons proposed openstack/openstack-helm master: Add apparmor security annotation https://review.openstack.org/602365 | 05:33 |
| *** jamesgu_ has quit IRC | 05:42 | |
| *** sdake has quit IRC | 06:35 | |
| *** sdake has joined #openstack-helm | 06:40 | |
| *** cfriesen has quit IRC | 06:48 | |
| *** unicell has joined #openstack-helm | 06:51 | |
| openstackgerrit | Merged openstack/openstack-helm-infra master: helm-toolkit: Support standard kubernetes/helm labels https://review.openstack.org/621698 | 07:11 |
| evrardjp | o/ | 07:59 |
| evrardjp | portdirect: are you still there? | 07:59 |
| *** unicell has quit IRC | 08:02 | |
| *** unicell has joined #openstack-helm | 08:05 | |
| *** unicell has quit IRC | 08:11 | |
| *** lemko has joined #openstack-helm | 08:33 | |
| *** rgogunskiy has joined #openstack-helm | 09:17 | |
| *** aojea has joined #openstack-helm | 09:40 | |
| openstackgerrit | Jean-Philippe Evrard proposed openstack/openstack-helm-images master: [WIP] Add LOCI to OSH images https://review.openstack.org/621206 | 11:52 |
| *** sgrasley has joined #openstack-helm | 13:59 | |
| *** sthussey has joined #openstack-helm | 14:00 | |
| portdirect | evrardjp: itw as a bit late for me last night :( | 14:15 |
| portdirect | srwilkers, lrensing, v1k0d3n, lamt, portdirect, renmak, jayahn, korzen, alanmeadows, mattmceuen, raymaika, evrardjp, MegHeisler, howell: meeting time in 45 mins | 14:15 |
| evrardjp | portdirect: no worries | 14:15 |
| evrardjp | I have a few questions | 14:15 |
| portdirect | over in #openstack-meeting-5, the Agenda is here https://etherpad.openstack.org/p/openstack-helm-meeting-2018-12-11 | 14:16 |
| portdirect | sure - sup evrardjp ? | 14:16 |
| evrardjp | portdirect: for the OVS image -- what's the reason to build from source, vs using a package? | 14:17 |
| portdirect | as at the time, the packages were broken - also it provides an easy way of adding things like dpdk | 14:18 |
| portdirect | samyaple and i talked about a sister project to loci for things like this - that would let us build required things like ovs/libvirt/qemu rather than rely on distro packages that often lagg behind where we want, or expose more features than we need | 14:19 |
| openstackgerrit | Steve Wilkerson proposed openstack/openstack-helm master: Armada: Manifest updates for Ocata release https://review.openstack.org/591808 | 14:22 |
| evrardjp | well I beg to disagree there, except if you're a dpdk dev :) | 14:22 |
| evrardjp | ok let me ask another question | 14:23 |
| evrardjp | would you be strongly against an image that has different implementations per distro? For example: ubuntu using sources, opensuse using packages | 14:24 |
| evrardjp | I trust my colleagues to build good packages. Else someone probably has a nearby whip I can use. | 14:24 |
| portdirect | works for me | 14:24 |
| portdirect | honestly - I'm quite up for moving the default base to debian | 14:25 |
| portdirect | as it removes many of these distro issues | 14:26 |
| portdirect | and then we could move ubuntu to packages as well | 14:26 |
| portdirect | but i think that should come later | 14:26 |
| evrardjp | I am fine with that | 14:27 |
| evrardjp | if you saw the pattern I have set in that repo, all the images will have different versions, per distro, per openstack branch (if relevant), per "VERSION" (latest, 1.0, ...) | 14:28 |
| evrardjp | so we should be fine | 14:28 |
| portdirect | nice - yeah it looks good evrardjp | 14:29 |
| portdirect | i'll try to get an account on quay.io this week so we can store these images somewhere | 14:30 |
| evrardjp | perfect that's what I wanted to ask you | 14:33 |
| evrardjp | create an account somewhere | 14:33 |
| evrardjp | so that I can use it for publishing images | 14:34 |
| lemko | Hi. With one physical network in openstack, everything works But the problem comes when I want to add other network. The comments in neutron/values.yaml are really helpful and I believe I did everything right. I'm adding more bridges in auto_bridge_add, then bridge_mappings: "net1:br-ex,net2:br-ex2", them in flat_network: "net1,net2". With more network physical network in openstack, it doesn't work anymore, not possible | 14:48 |
| lemko | to access the network. What could be the cause of this? I looked at the bridge configuration in ovs and everything seems good. | 14:48 |
| *** cfriesen has joined #openstack-helm | 14:58 | |
| jayahn | portdirect: evrardjp: I am trying to summarize our thought before irc meeting, however I probably need a bit more time. pls read this week etherpad if the current writing makes any sense for you guys. otherwise, I will take some more time to properly write up the exact thought I want to pass. | 15:00 |
| evrardjp | great | 15:00 |
| jayahn | long story short, skt can contribute anything relared to kolla image based deployment, but we really need on loci side. | 15:01 |
| portdirect | jayahn: would also be great to get an update on the docs repo? | 15:01 |
| portdirect | understand the kolla issues - would be great to get you invovled there, esp as over time i expect we will move away from supporting default kolla images as discussed at the ptgs | 15:02 |
| jayahn | jaesang has been looking into docs repo for the last several days, I need to get an update from him. I will do that the first thing in tomorrow. | 15:02 |
| *** jamesgu_ has joined #openstack-helm | 15:10 | |
| *** aaronsheffield has joined #openstack-helm | 15:12 | |
| openstackgerrit | Rahul Khiyani proposed openstack/openstack-helm master: Keystone podSecurityContext https://review.openstack.org/614011 | 15:14 |
| openstackgerrit | Mohan V proposed openstack/openstack-helm-infra master: Do-not run nginx as root for Ingress helm chart https://review.openstack.org/622359 | 15:58 |
| *** nzoueidi has quit IRC | 16:02 | |
| jayahn | portdirect: party is gone. (sad). we recently had big organizational changes, and it is now time to do all the planning.... | 16:03 |
| *** nzoueidi has joined #openstack-helm | 16:04 | |
| jayahn | my team is still intact though. (fortunately!) got more tasks... (i am not sure if it is good thing.. :) ) | 16:04 |
| jayahn | I really want to invite all of you to Korea next year, need to come up with a very good excuse for that. ;) | 16:05 |
| openstackgerrit | Pete Birley proposed openstack/openstack-helm-infra master: Ingress: Remove server headers from response https://review.openstack.org/624429 | 16:06 |
| openstackgerrit | Steve Wilkerson proposed openstack/openstack-helm master: Armada: Manifest updates for Ocata release https://review.openstack.org/591808 | 16:07 |
| openstackgerrit | Steve Wilkerson proposed openstack/openstack-helm-infra master: Kibana: Add basic Helm tests https://review.openstack.org/609703 | 16:11 |
| *** jamesgu_ has quit IRC | 16:15 | |
| *** jamesgu_ has joined #openstack-helm | 16:19 | |
| openstackgerrit | Steve Wilkerson proposed openstack/openstack-helm-infra master: DNM: Test elastic query in post run job https://review.openstack.org/624435 | 16:20 |
| *** munimeha1 has joined #openstack-helm | 16:33 | |
| *** aaronsheffield has quit IRC | 16:36 | |
| *** aaronsheffield has joined #openstack-helm | 16:37 | |
| openstackgerrit | Arun Kant proposed openstack/openstack-helm-images master: [WIP] Add LOCI to OSH images https://review.openstack.org/621206 | 16:55 |
| cfriesen | question about the various "controller" services. Is there anything in the charts to ensure that multiple pods (for cinder-api, or glance-api, or ingress, etc.) don't end up on the same k8s node? | 16:57 |
| evrardjp | portdirect: are you there? | 17:08 |
| evrardjp | portdirect: remember when we discussed consumption of images, you said we could build locally, tag the image with docker.io/... and then it would be re-used correctly in the helm charts. Did I understand that correctly? | 17:09 |
| *** jamesgu__ has joined #openstack-helm | 17:35 | |
| openstackgerrit | Steve Wilkerson proposed openstack/openstack-helm-infra master: DNM: Test elastic query in post run job https://review.openstack.org/624435 | 17:36 |
| srwilkers | cfriesen: the charts should have antiaffinity defined to help prevent that where possible | 17:39 |
| *** jamesgu_ has quit IRC | 17:39 | |
| cfriesen | srwilkers: I see it, but it's "preferred" rather than required. I noticed it because when I rebooted one of the two controllers I ended up with most of the pods on the other controller and they never redistributed themselves | 17:42 |
| cfriesen | If I want one API pod per controller for host-based redundancy, I think "required" anti-affinity might make more sense. | 17:43 |
| cfriesen | unless the intent is just a pool of pods and we hope that statistically they get distributed reasonably | 17:45 |
| srwilkers | cfriesen: they can be overridden just like the other values in values.yaml. I feel preferred is a sane default | 17:50 |
| cfriesen | how many controller nodes does openstack-helm normally test with? | 17:51 |
| supamatt | the gates? | 17:52 |
| cfriesen | or typical use at AT&T | 17:52 |
| portdirect | cfriesen: thats a piece of string question, and thankfully pretty similar whether talking best practive or within ATT | 18:00 |
| portdirect | I'd recomend a minimum of three controllers | 18:01 |
| *** aojea has quit IRC | 18:01 | |
| portdirect | beyond there is driven by a balance of load, and physical layout | 18:01 |
| portdirect | i think most people running osh, tend to run a controller per rack | 18:02 |
| portdirect | and run at least three racks in a region | 18:02 |
| cfriesen | fair enough. we have a driver to go down to single node and 1+1 configurations, which means we run into issues sometimes. | 18:02 |
| portdirect | for this - id expect things to 'just work' - though as allways your millage may vary | 18:03 |
| portdirect | it was one of the reasons we default to soft anti affinity | 18:03 |
| portdirect | as though like you, i much prefer hard - it was a less friendly experience for people getting up and running, cicd, and smaller sites | 18:04 |
| cfriesen | makes sense...thanks for the explanation | 18:04 |
| portdirect | evrardjp: re the images, its a bit of a hack | 18:07 |
| portdirect | as we default to 'pull if not present', you can make k8s run any image thats pre-loaded onto the host, provided its tag matches whats expected | 18:07 |
| portdirect | its for this reason that i'd caution against that param in production, and also recommend pulling by sha over tag if possible | 18:08 |
| portdirect | but for quick and dirty testing ive used it a few times, though it would be leaner to just provide an over-ride to the chart | 18:08 |
| portdirect | *cleaner | 18:09 |
| evrardjp | I see | 18:09 |
| openstackgerrit | Steve Wilkerson proposed openstack/openstack-helm master: Update upgrade host playbook in osh to match osh-infra https://review.openstack.org/624454 | 18:11 |
| openstackgerrit | Steve Wilkerson proposed openstack/openstack-helm master: Armada: Manifest updates for Ocata release https://review.openstack.org/591808 | 18:20 |
| openstackgerrit | Steve Wilkerson proposed openstack/openstack-helm-infra master: DNM: Test elastic query in post run job https://review.openstack.org/624435 | 18:31 |
| openstackgerrit | Pete Birley proposed openstack/openstack-helm-infra master: Ingress: Update sleep function to not require dumb-init https://review.openstack.org/624465 | 18:54 |
| openstackgerrit | Jean-Philippe Evrard proposed openstack/openstack-helm-images master: [WIP] Add LOCI to OSH images https://review.openstack.org/621206 | 19:06 |
| openstackgerrit | Steve Wilkerson proposed openstack/openstack-helm-infra master: DNM: Test elastic query in post run job https://review.openstack.org/624435 | 19:21 |
| openstackgerrit | Jean-Philippe Evrard proposed openstack/openstack-helm-images master: [WIP] Add LOCI to OSH images https://review.openstack.org/621206 | 19:41 |
| cfriesen | has anyone seen issues like this with the rabbit pods? It seems like the one pod has multiple servers running but isn't actually working. http://paste.openstack.org/show/737050/ kubectl says the pods/containers are running, but I'm having issues with RPC in nova. | 19:43 |
| cfriesen | (this is with the stock docker image, after rebooting controller-1) | 19:44 |
| *** lemko has quit IRC | 19:50 | |
| openstackgerrit | Vladyslav Drok proposed openstack/openstack-helm-infra master: Do not expose ingress nginx version by default https://review.openstack.org/624483 | 20:17 |
| openstackgerrit | Drew Walters proposed openstack/openstack-helm-addons master: Add aptly chart https://review.openstack.org/623314 | 20:35 |
| openstackgerrit | Merged openstack/openstack-helm-infra master: Elasticsearch: Remove default Curator action configuration https://review.openstack.org/620092 | 20:58 |
| openstackgerrit | Mike Pham proposed openstack/openstack-helm master: WIP: Implement East-West EGRESS policies for the Cloud Core services https://review.openstack.org/623315 | 21:04 |
| openstackgerrit | Drew Walters proposed openstack/openstack-helm-addons master: [WIP] Add aptly chart https://review.openstack.org/623314 | 21:04 |
| openstackgerrit | Rahul Khiyani proposed openstack/openstack-helm master: Keystone podSecurityContext https://review.openstack.org/614011 | 21:19 |
| openstackgerrit | Merged openstack/openstack-helm master: Update upgrade host playbook in osh to match osh-infra https://review.openstack.org/624454 | 21:23 |
| openstackgerrit | Mike Pham proposed openstack/openstack-helm master: WIP: Implement East-West EGRESS policies for the Cloud Core services https://review.openstack.org/623315 | 21:47 |
| openstackgerrit | Merged openstack/openstack-helm-infra master: Ingress: Update sleep function to not require dumb-init https://review.openstack.org/624465 | 22:01 |
| openstackgerrit | Mike Pham proposed openstack/openstack-helm master: WIP: Implement East-West EGRESS policies for the Cloud Core services https://review.openstack.org/623315 | 22:24 |
| openstackgerrit | Stephanie Miller proposed openstack/openstack-helm-images master: [WIP] Add openSUSE Leap 15 openvswitch image https://review.openstack.org/621145 | 23:14 |
| *** jamesgu_ has joined #openstack-helm | 23:29 | |
| *** jamesgu__ has quit IRC | 23:32 | |
| *** cliff_ has joined #openstack-helm | 23:43 | |
| *** cliff_ has quit IRC | 23:50 | |
Generated by irclog2html.py 2.15.3 by Marius Gedminas - find it at mg.pov.lt!