| NobodyCam | anyone ever see a minion node change it internal ip when adding a port | 00:21 |
|---|---|---|
| NobodyCam | I added a Ip tp the minion and the internal ip changed | 00:22 |
| flwang | so you mean adding a port to an existing node which caused the minion ip changed? | 00:40 |
| NobodyCam | yes | 00:42 |
| NobodyCam | can I give more then one external network on cluster template create? | 00:44 |
| NobodyCam | never mind I don't think you can have more then one floating ip | 00:52 |
| flwang | no, you can't | 00:53 |
| flwang | k8s will always poll your network from openstack to reflect the latest network status | 00:53 |
| flwang | so i'm not surprise that your internal ip changed TBH | 00:53 |
| NobodyCam | can I pin to the real internal ip? | 00:54 |
| NobodyCam | the networks I added are listed as external the real internal is not | 00:58 |
| *** ricolin has joined #openstack-containers | 01:00 | |
| flwang | NobodyCam: i'm not sure, i never try that | 01:39 |
| *** rcernin has quit IRC | 01:49 | |
| *** rcernin has joined #openstack-containers | 01:53 | |
| *** hongbin has joined #openstack-containers | 01:54 | |
| *** ianychoi has quit IRC | 02:03 | |
| *** ianychoi has joined #openstack-containers | 02:03 | |
| openstackgerrit | Feilong Wang proposed openstack/magnum master: [fedora_atomic] Support auto healing for k8s https://review.openstack.org/631378 | 02:23 |
| *** ykarel|away has joined #openstack-containers | 02:52 | |
| jakeyip | anybody using mesos? https://lists.apache.org/thread.html/b162dd624dc088cd634292f0402282a1d1d0ce853baeae8205bc033c@%3Cdev.mesos.apache.org%3E | 03:05 |
| *** hongbin has quit IRC | 03:06 | |
| *** ramishra has joined #openstack-containers | 03:07 | |
| *** ykarel|away has quit IRC | 03:32 | |
| *** ykarel|away has joined #openstack-containers | 03:46 | |
| *** udesale has joined #openstack-containers | 03:46 | |
| *** ykarel|away is now known as ykarel | 03:57 | |
| *** udesale has quit IRC | 04:44 | |
| *** udesale has joined #openstack-containers | 04:46 | |
| *** janki has joined #openstack-containers | 04:50 | |
| *** ramishra_ has joined #openstack-containers | 06:41 | |
| *** ramishra has quit IRC | 06:41 | |
| *** ramishra_ has quit IRC | 06:53 | |
| *** ramishra has joined #openstack-containers | 06:55 | |
| *** alisanhaji has quit IRC | 07:27 | |
| *** ykarel is now known as ykarel|lunch | 07:29 | |
| *** ivve has joined #openstack-containers | 07:47 | |
| *** ykarel|lunch is now known as ykarel | 08:25 | |
| openstackgerrit | OpenStack Proposal Bot proposed openstack/magnum-ui master: Imported Translations from Zanata https://review.openstack.org/648034 | 08:30 |
| openstackgerrit | OpenStack Proposal Bot proposed openstack/magnum-ui stable/rocky: Imported Translations from Zanata https://review.openstack.org/648035 | 08:31 |
| *** ricolin has quit IRC | 08:40 | |
| *** alisanhaji has joined #openstack-containers | 08:48 | |
| *** pcaruana has joined #openstack-containers | 08:57 | |
| *** ttsiouts has joined #openstack-containers | 09:00 | |
| *** ramishra_ has joined #openstack-containers | 09:04 | |
| *** ramishra has quit IRC | 09:04 | |
| *** ramishra_ has quit IRC | 09:10 | |
| *** ttsiouts has quit IRC | 09:10 | |
| *** ttsiouts has joined #openstack-containers | 09:13 | |
| *** ramishra has joined #openstack-containers | 09:14 | |
| *** ricolin has joined #openstack-containers | 10:33 | |
| *** rcernin has quit IRC | 10:51 | |
| *** ricolin has quit IRC | 10:52 | |
| *** ramishra has quit IRC | 10:56 | |
| *** ramishra has joined #openstack-containers | 10:59 | |
| *** ricolin has joined #openstack-containers | 11:40 | |
| openstackgerrit | Merged openstack/magnum-ui stable/rocky: Imported Translations from Zanata https://review.openstack.org/648035 | 11:47 |
| openstackgerrit | Merged openstack/magnum-ui master: Imported Translations from Zanata https://review.openstack.org/648034 | 11:47 |
| *** goldyfruit has quit IRC | 12:23 | |
| *** pcaruana has quit IRC | 12:39 | |
| *** pcaruana has joined #openstack-containers | 12:42 | |
| *** pcaruana has quit IRC | 12:42 | |
| *** pcaruana has joined #openstack-containers | 12:43 | |
| *** udesale has quit IRC | 12:48 | |
| *** udesale has joined #openstack-containers | 12:49 | |
| *** itlinux has quit IRC | 12:57 | |
| *** ricolin_phone has joined #openstack-containers | 13:02 | |
| *** ricolin_phone has quit IRC | 13:04 | |
| *** ricolin_phone has joined #openstack-containers | 13:06 | |
| *** ricolin_phone_ has joined #openstack-containers | 13:07 | |
| brtknr | ttsiouts: ng list and show working perfectly :) | 13:09 |
| ttsiouts | brtknr: Awesome news!!! | 13:09 |
| *** jhesketh has quit IRC | 13:10 | |
| *** ricolin_phone_ has quit IRC | 13:10 | |
| *** goldyfruit has joined #openstack-containers | 13:10 | |
| ttsiouts | brtknr: do you have any suggestions on the show output? | 13:11 |
| *** ricolin_phone has quit IRC | 13:12 | |
| ttsiouts | brtknr: i mean add/remove nodegroup fields to/from the output | 13:12 |
| brtknr | ttsiouts: what is the significance of the default nodegroup? | 13:13 |
| brtknr | ttsiouts: are those non deletable? | 13:13 |
| ttsiouts | brtknr: default NGs are the ones created with the cluster | 13:14 |
| brtknr | ttsiouts: Can those be deleted or not? | 13:15 |
| ttsiouts | brtknr: this is the intention, to not allow their deletion | 13:15 |
| brtknr | ttsiouts: ah okay... | 13:15 |
| brtknr | maybe they should be marked clearly? | 13:15 |
| ttsiouts | brtknr: thoughts on this? | 13:15 |
| brtknr | or they have a reserved role names | 13:16 |
| ttsiouts | brtknr: hmmm | 13:17 |
| ttsiouts | brtknr: for now their names come from <cluster_name>+<role> | 13:17 |
| ttsiouts | brtknr: I don't know if this is enough | 13:18 |
| ttsiouts | brtknr: on the other hand, API will throw errors if the user tries to delete them | 13:18 |
| brtknr | ttsiouts: perhaps the roles do not need to be visible then... thats reduntant information during list | 13:18 |
| brtknr | are roles unique? | 13:18 |
| ttsiouts | brtknr: nope | 13:19 |
| brtknr | or can multiple NG have the same role? | 13:19 |
| brtknr | oh okay | 13:19 |
| *** mrodriguez has joined #openstack-containers | 13:19 | |
| ttsiouts | brtknr: well for now master role will be reserved for master NGs | 13:19 |
| brtknr | it me, it makes sense to have unique role names | 13:19 |
| brtknr | it me, it makes sense to have unique role names *per cluster | 13:20 |
| ttsiouts | brtknr: hmmm | 13:20 |
| brtknr | can there be more role names than master and worker? | 13:20 |
| *** ricolin_phone has joined #openstack-containers | 13:20 | |
| ttsiouts | it's a string field | 13:20 |
| brtknr | okay... | 13:20 |
| ttsiouts | could be anything | 13:20 |
| ttsiouts | this is because we did not want to limit what the user can choose | 13:21 |
| brtknr | i am just trying to figure out what functional purpose it serves over than using the name | 13:21 |
| *** ricolin_phone has quit IRC | 13:21 | |
| ttsiouts | brtknr: for now there is a constraint in the DB to have unique NG names per cluster | 13:22 |
| brtknr | ttsiouts: so that will indirectly enforce unique role per NG | 13:23 |
| brtknr | ttsiouts: the way things are... | 13:23 |
| ttsiouts | brtknr: the <cluster_name>+<role> convention applies to the default NGs | 13:24 |
| ttsiouts | brtknr: when adding new NGs the user should be able to specify the name he wants | 13:25 |
| brtknr | hmm | 13:25 |
| brtknr | i am not sure if we need both role and name for nodegroup... | 13:25 |
| ttsiouts | brtknr: I think we do | 13:26 |
| ttsiouts | let's say that the user has two NGs with the same role but in different availability zones for HA | 13:26 |
| brtknr | ttsiouts: hmm | 13:27 |
| brtknr | ttsiouts: i see your point | 13:28 |
| ttsiouts | brtknr: this way we have the name so that the user can distinguish between the NGs with the same role | 13:28 |
| brtknr | ttsiouts: i think the default names can just be the role name in that case | 13:29 |
| brtknr | for master and worker | 13:29 |
| ttsiouts | brtknr: yeah.. I see your point | 13:30 |
| ttsiouts | Do you want to drop a comment in gerrit? So the other reviewers can share their thoughts on this? | 13:31 |
| brtknr | I just tried to create a NG :D sadly it failed :( but still exciting! created a DB entry :) | 13:31 |
| ttsiouts | brtknr: yeah the driver throws a NotImplemented still | 13:33 |
| brtknr | ttsiouts: would be cool to list nodegroup by role too... | 13:34 |
| brtknr | do you think thats easily doable? | 13:34 |
| ttsiouts | brtknr: yes | 13:34 |
| brtknr | as in `openstack coe nodegroup list k8s --role worker | 13:34 |
| brtknr | ` | 13:34 |
| ttsiouts | the DB api already supports role as filter | 13:34 |
| brtknr | ttsiouts: nice | 13:34 |
| ttsiouts | brtknr: I can add it | 13:35 |
| ttsiouts | brtknr: again thanks for your effort and your great suggestions!! | 13:35 |
| brtknr | ttsiouts: my pleasure! | 13:36 |
| brtknr | also, can we make CLI invocation `nodegroup` instead of `node-group`? its easier to write :) also easier to change it now than later | 13:37 |
| brtknr | after everyone gets used to it | 13:37 |
| ttsiouts | brtknr: sure | 13:44 |
| brtknr | ttsiouts: btw, if i create a NG without specifying a role, the default value is `minion` | 13:47 |
| brtknr | whereas for the default worker, the role is `worker` | 13:47 |
| ttsiouts | brtknr: it's hardcoded in the API controller but it will change | 13:49 |
| ttsiouts | brtknr: https://review.openstack.org/#/c/647792/1/magnum/api/controllers/v1/nodegroup.py L240 | 13:49 |
| *** ykarel is now known as ykarel|away | 13:52 | |
| *** hongbin has joined #openstack-containers | 13:56 | |
| *** ykarel|away has quit IRC | 13:59 | |
| goldyfruit | o/ | 14:20 |
| *** itlinux has joined #openstack-containers | 14:27 | |
| *** janki has quit IRC | 14:41 | |
| *** munimeha1 has joined #openstack-containers | 14:45 | |
| *** alisanhaji has quit IRC | 14:51 | |
| *** ykarel|away has joined #openstack-containers | 14:53 | |
| *** ykarel|away is now known as ykarel | 14:58 | |
| *** udesale has quit IRC | 15:35 | |
| *** alisanhaji has joined #openstack-containers | 15:39 | |
| dioguerra | should i also make a MR for stable/stein of https://review.openstack.org/#/c/645632/? | 15:50 |
| dioguerra | https://review.openstack.org/#/c/645632/ | 15:50 |
| *** ykarel is now known as ykarel|away | 16:16 | |
| openstackgerrit | Theodoros Tsioutsias proposed openstack/magnum master: APIs for listing and showing nodegroups https://review.openstack.org/604824 | 16:26 |
| openstackgerrit | Theodoros Tsioutsias proposed openstack/magnum master: [WIP] APIs for nodegroup CRUD operations https://review.openstack.org/647792 | 16:26 |
| openstackgerrit | Theodoros Tsioutsias proposed openstack/python-magnumclient master: Add nodegroup list/show commands https://review.openstack.org/642077 | 16:28 |
| openstackgerrit | Theodoros Tsioutsias proposed openstack/python-magnumclient master: Add nodegroup CRUD commands https://review.openstack.org/647793 | 16:28 |
| NobodyCam | Good Morning Folks | 16:28 |
| ttsiouts | brtknr: ^ changed node=group to nodegroup and added list by role | 16:28 |
| ttsiouts | brtknr: s/node=group/node-group | 16:29 |
| brtknr | ttsiouts: I just saw :) thank you for your hard work... also saw list by roles :) | 16:31 |
| ttsiouts | brtknr: :) | 16:32 |
| NobodyCam | anyone happen to know where k8's would get the external network ID for things like creating a load balancer? looks like its selecting a incorrect network for me. | 16:32 |
| NobodyCam | network from load balancer service: FloatingNetworkID:130c6451-5f71-44f5-86f9-68fb44f1f01e / network in template external_network_id | b2b181e9-4cab-434f-84f7-54fef0b70483 | 16:33 |
| brtknr | NobodyCam: thats strange... | 16:33 |
| brtknr | NobodyCam: is that the subnet ID? | 16:34 |
| brtknr | NobodyCam: for the floating network? | 16:34 |
| *** ttsiouts has quit IRC | 16:34 | |
| *** ttsiouts has joined #openstack-containers | 16:35 | |
| NobodyCam | which one? the template id is a external network ID the FloatingIPnetwork is another network | 16:35 |
| NobodyCam | it should be getting a ip from vlan 100 but its attempting to use an ip from vlan8 : http://paste.openstack.org/show/K8jF2BmXOZnLlfjVtX5H/ | 16:38 |
| *** ttsiouts has quit IRC | 16:40 | |
| goldyfruit | NobodyCam: When you create your Magnum template you have to specify the external network | 16:40 |
| goldyfruit | Option --external-network UUID | 16:40 |
| NobodyCam | yep I did! | 16:40 |
| NobodyCam | openstack coe cluster template create k8s-test-fedora-27_5 --coe kubernetes --network-driver calico --docker-volume-size 6 --volume-driver cinder --flavor 223e626b-bc0e-43e1-9a0b-9d167e06355d --external-network b2b181e9-4cab-434f-84f7-54fef0b70483 --fixed-network 1bfcdfcd-224a-4aee-898e-9f0a2f3ab4a3 --fixed-subnet 12a1beb0-553d-429d-b98d-8a527255fda3 --keypair AdminSshKey --image PsFedora27WithHelm --public | 16:41 |
| NobodyCam | --labels kube_tag=v1.11.8,tiller_enabled=True,availability_zone=nova,calico_tag=v2.6.11,calico_cni_tag=v1.11.7,calico_kube_controllers_tag=v1.0.4 | 16:41 |
| NobodyCam | oh sorry for the large paste | 16:41 |
| goldyfruit | I'm using using these labels "tiller_enabled=True,availability_zone=nova" | 16:42 |
| goldyfruit | But I got almost the same thing than your template | 16:42 |
| goldyfruit | https://paste.api-zulu.com/raw/cicusiwilu | 16:43 |
| NobodyCam | oh. I see you have --master-lb-enabled | 16:44 |
| goldyfruit | This options will create 2 LB for your Kubernetes cluster, 1 for etcd and 1 for the API | 16:45 |
| goldyfruit | Don't use this label with your version: heat_container_agent_tag=stein-dev | 16:46 |
| goldyfruit | Not sure of the impact | 16:46 |
| goldyfruit | Previously we had to use heat_container_agent_tag=rawhide to be able to spawn a cluster | 16:47 |
| goldyfruit | But we never got your issue | 16:47 |
| goldyfruit | b2b181e9-4cab-434f-84f7-54fef0b70483 is your floating IP network right ? | 16:47 |
| NobodyCam | yeppers | 16:48 |
| NobodyCam | I'n not at all understanding where the 130c6451-5f71-44f5-86f9-68fb44f1f01e network came from its not called for anywhere | 16:49 |
| *** ricolin has quit IRC | 16:50 | |
| goldyfruit | In your octavia confiuration you have to put a network | 16:51 |
| goldyfruit | Nothing in there related to this network ? | 16:51 |
| goldyfruit | What is this network 130c6451-5f71-44f5-86f9-68fb44f1f01e ? | 16:51 |
| goldyfruit | What is you remove this options "--fixed-network 1bfcdfcd-224a-4aee-898e-9f0a2f3ab4a3 --fixed-subnet 12a1beb0-553d-429d-b98d-8a527255fda3" and let Magnum create the private network ? | 16:53 |
| NobodyCam | I have not attempted that.. | 16:54 |
| NobodyCam | looking at the octavia conf I only see one network: it is the br-lbaas network: `amp_boot_network_list = bcda6bb9-f1bf-4fbe-b444-6f76c9346906` | 16:55 |
| NobodyCam | am I missing a octavia config? | 16:56 |
| goldyfruit | Don't think | 16:56 |
| NobodyCam | the 130 network is just a vlan network not at all in user for this. | 16:59 |
| brtknr | NobodyCam: if you do `openstack floating ip list` what UUIDs are next to the IP assigned to your node/loadbalancer | 17:03 |
| NobodyCam | all from the correct network: http://paste.openstack.org/show/a3fc9Q8G1KHUPybcm2kY/ | 17:05 |
| goldyfruit | This network is public ? | 17:05 |
| goldyfruit | or private ? | 17:05 |
| NobodyCam | public | 17:07 |
| * NobodyCam verifies that | 17:07 | |
| NobodyCam | oh | shared | False | 17:08 |
| goldyfruit | That is ok | 17:09 |
| goldyfruit | router:external is set ? | 17:09 |
| goldyfruit | DHCP is disabled on the subnet ? | 17:10 |
| NobodyCam | dhcp is true: http://paste.openstack.org/show/TIRKCvnenyX00RC1cxf9/ | 17:13 |
| goldyfruit | https://docs.openstack.org/neutron/latest/admin/config-service-subnets.html#example-2-dvr-configuration | 17:19 |
| goldyfruit | Should be disabled | 17:19 |
| goldyfruit | Forget about the --service-type 'network:floatingip' part | 17:19 |
| *** ramishra has quit IRC | 17:37 | |
| NobodyCam | I'm defiantly not a go lang person but if I read this correctly the cloud-provider-openstack picks any external network???? https://github.com/kubernetes/cloud-provider-openstack/blob/b100f24bb0e4b70c71587e24cac6ae84ddfa77d4/pkg/cloudprovider/providers/openstack/openstack_loadbalancer.go#L774 | 17:43 |
| goldyfruit | It seems when multiple external network are detected it picks the first one | 17:47 |
| goldyfruit | But only when no external network have been defined | 17:47 |
| goldyfruit | glog.V(4).Infof("find multiple external networks, pick the first one when there are no explicit configuration.") | 17:47 |
| NobodyCam | yea, I have several external networks for other VM workloads | 17:47 |
| goldyfruit | Magnum is using the version 0.2.0 of cloud provider (choose the correct tag when you look into the code) | 17:48 |
| NobodyCam | oh :p | 17:48 |
| goldyfruit | if externalNet.External { floatingNetworkIds = append(floatingNetworkIds, externalNet.ID) } | 17:50 |
| goldyfruit | It should pick the network you specified | 17:51 |
| goldyfruit | https://github.com/openstack/magnum/blob/stable/rocky/magnum/drivers/k8s_fedora_atomic_v1/templates/kubecluster.yaml#L329-L336 | 17:52 |
| *** ykarel|away has quit IRC | 18:16 | |
| *** zul has quit IRC | 18:23 | |
| *** hongbin has quit IRC | 19:15 | |
| *** hongbin has joined #openstack-containers | 19:28 | |
| *** pcaruana has quit IRC | 19:53 | |
| openstackgerrit | Feilong Wang proposed openstack/magnum master: Add cluster upgrade to the API https://review.openstack.org/514959 | 20:03 |
| *** hongbin has quit IRC | 20:07 | |
| NobodyCam | hey hey goldyfruit question is your lbaas-mgmt network marked as external? | 20:31 |
| *** itlinux has quit IRC | 20:46 | |
| openstackgerrit | Feilong Wang proposed openstack/magnum master: Kubernetes images release https://review.openstack.org/648243 | 21:03 |
| *** henriqueof has quit IRC | 21:07 | |
| *** jmlowe has quit IRC | 21:20 | |
| *** gsimondon has joined #openstack-containers | 21:33 | |
| goldyfruit | NobodyCam: My lbaas-mgnt is a provider network | 21:37 |
| goldyfruit | Marked as internal | 21:37 |
| NobodyCam | ack. :) | 21:38 |
| NobodyCam | I still can find where my system is getting that network id :p | 21:38 |
| *** gsimondon has quit IRC | 21:40 | |
| *** alisanhaji has quit IRC | 21:40 | |
| *** gsimondon has joined #openstack-containers | 21:46 | |
| goldyfruit | :( | 21:46 |
| *** ivve has quit IRC | 21:50 | |
| *** itlinux has joined #openstack-containers | 22:14 | |
| *** jmlowe has joined #openstack-containers | 22:19 | |
| *** goldyfruit has quit IRC | 22:26 | |
| *** rcernin has joined #openstack-containers | 22:31 | |
| *** mrodriguez has quit IRC | 22:32 | |
| *** munimeha1 has quit IRC | 22:51 | |
| openstackgerrit | Merged openstack/magnum master: Kubernetes images release https://review.openstack.org/648243 | 23:12 |
| *** gsimondon has quit IRC | 23:44 | |
Generated by irclog2html.py 2.15.3 by Marius Gedminas - find it at mg.pov.lt!