| *** kranthikirang has joined #openstack-helm | 00:58 | |
| *** kranthikirang has joined #openstack-helm | 00:59 | |
| *** kranthikirang has quit IRC | 01:19 | |
| *** phuoc has quit IRC | 04:55 | |
| *** phuoc has joined #openstack-helm | 04:56 | |
| *** gagehugo has quit IRC | 07:14 | |
| *** gagehugo has joined #openstack-helm | 07:17 | |
| *** sgrasley has quit IRC | 07:40 | |
| *** todin has quit IRC | 09:20 | |
| *** georgk has joined #openstack-helm | 10:11 | |
| *** georgk has quit IRC | 10:12 | |
| *** todin has joined #openstack-helm | 16:24 | |
| v1k0d3n | hey portdirect have a question that you'd probably have the quickest answer for. is there a way to add options to the `/mnt/rootfs/etc/kubernetes/kubeadm-conf.yaml` manifest for the kube-aio deployment? i wanted to try something to see if i could squeeze a little snappier performance out of an AIO OSH deployment. if it works, i can submit it back. | 16:54 |
|---|---|---|
| v1k0d3n | i've been back trying to explore airship components, but have hit a couple of snags. figured i would dial back and just see if i can swap my old kolla environment for OSH if it's stable enough, and if i can have a little more control of the kubernetes side. | 16:57 |
| v1k0d3n | have you guys been keeping up with the slack side? seems like some questions about ceph/mariadb have been coming up, but i haven't played OSH in a long time, so i can't confirm or help answer any questions quite yet. maybe after i get reacquainted with where things are at now i can potentially field some q's (if that would help out at all). | 17:00 |
| portdirect | https://github.com/openstack/openstack-helm-infra/blob/master/tools/images/kubeadm-aio/assets/opt/playbooks/roles/deploy-kubeadm-master/templates/kubeadm-conf.yaml.j2 | 17:00 |
| portdirect | Not super templated ATM, but you should be able to add things in there | 17:01 |
| portdirect | I'm looking to get the slack sync back up, as without it I've been concious that there's not been much activity from core over there | 17:01 |
| v1k0d3n | ah! dude, thank you. | 17:04 |
| v1k0d3n | yeah, i figured it was just that sync wasn't in place, and it was just going a bit unnoticed. hard to keep up with everything. | 17:05 |
| v1k0d3n | plus, it's not like you guys have meetings or anything like that :P | 17:05 |
| srwilkers | Meetings? What’re those? | 17:25 |
| *** kranthikirang has joined #openstack-helm | 17:29 | |
| v1k0d3n | :) | 17:57 |
| *** kranthikirang has quit IRC | 18:08 | |
| *** kranthikirang has joined #openstack-helm | 18:24 | |
| *** kranthikirang has quit IRC | 18:40 | |
| *** kranthikirang has joined #openstack-helm | 18:45 | |
| v1k0d3n | so @portdirect as expected, the ipvs mods appear to be working out of the box. not sure if this is something you guys or not though? | 18:57 |
| v1k0d3n | ubuntu@osh-testing:~$ kubectl logs kube-proxy-rsq9r -n kube-system | 18:57 |
| v1k0d3n | I0127 18:38:52.093485 1 feature_gate.go:226] feature gates: &{{} map[SupportIPVSProxyMode:true]} | 18:57 |
| v1k0d3n | I0127 18:38:52.112070 1 server_others.go:183] Using ipvs Proxier. | 18:57 |
| v1k0d3n | W0127 18:38:52.786617 1 server.go:601] Failed to retrieve node info: nodes "osh-testing" not found | 18:57 |
| v1k0d3n | W0127 18:38:52.788907 1 proxier.go:289] invalid nodeIP, initializing kube-proxy with 127.0.0.1 as nodeIP | 18:57 |
| v1k0d3n | W0127 18:38:52.788936 1 proxier.go:304] IPVS scheduler not specified, use rr by default | 18:57 |
| v1k0d3n | I0127 18:38:52.789240 1 server_others.go:209] Tearing down inactive rules. | 18:57 |
| v1k0d3n | I0127 18:38:52.878288 1 server.go:444] Version: v1.10.9 | 18:57 |
| v1k0d3n | hmm...that didn't paste like i wanted. | 18:58 |
| v1k0d3n | https://www.irccloud.com/pastebin/qvrYsx31/ | 18:58 |
| v1k0d3n | https://www.irccloud.com/pastebin/GjrrEEeV/ | 18:58 |
| *** kranthikirang has quit IRC | 19:00 | |
| v1k0d3n | there's still quite a bit of iptables being spun up, which I don't see as much of in our Flagship deployment side. so that gives me some pause on if it's working all exactly as i would expect. but it is definitely enabled and working. | 19:00 |
| v1k0d3n | new services are definitely being activated via the ipvs kernel lb...so that's a really good sign. | 19:03 |
| v1k0d3n | https://www.irccloud.com/pastebin/TJFwCGnH/ | 19:04 |
| portdirect | Yeah it's easy to do | 19:05 |
| portdirect | https://github.com/project-azorian/halibut/blob/master/scripts/snippets/deploy-kubernetes.sh & https://github.com/project-azorian/halibut/blob/master/ansible/kubernetes.yaml | 19:06 |
| portdirect | Would be nice to add it to the gates I suppose, though I would not expect a night and day difference | 19:06 |
| v1k0d3n | yes, it's super easy to add. it is, however, something i've seen a lot of people overlook, including a lot of vendors we've come in contact with. if it's not that useful and you already have it figured out, it's not really something i need to spend time on. | 19:15 |
| *** kranthikirang has joined #openstack-helm | 19:37 | |
| *** kranthikirang has quit IRC | 19:41 | |
| *** jamesgu has joined #openstack-helm | 20:15 | |
| *** kranthikirang has joined #openstack-helm | 20:45 | |
| *** jamesgu has quit IRC | 21:13 | |
| *** jamesgu has joined #openstack-helm | 21:13 | |
| *** kranthikirang has quit IRC | 21:48 | |
| *** kranthikirang has joined #openstack-helm | 23:04 | |
| *** kranthikirang has quit IRC | 23:10 | |
| *** kranthikirang has joined #openstack-helm | 23:42 | |
| *** kranthikirang has quit IRC | 23:47 | |
Generated by irclog2html.py 2.15.3 by Marius Gedminas - find it at mg.pov.lt!