| openstackgerrit | Peter Pouliot proposed openstack/magnum master: Fixing typos and spelling errors in driver template files https://review.opendev.org/688017 | 01:18 |
|---|---|---|
| openstackgerrit | Peter Pouliot proposed openstack/magnum master: Fixing typos and spelling errors in driver template files https://review.opendev.org/688017 | 02:21 |
| *** hongbin has joined #openstack-containers | 02:37 | |
| *** ramishra has joined #openstack-containers | 03:24 | |
| *** hongbin has quit IRC | 03:24 | |
| *** ykarel|away has joined #openstack-containers | 03:33 | |
| *** dave-mccowan has quit IRC | 04:39 | |
| *** pcaruana has joined #openstack-containers | 04:55 | |
| *** osmanlicilegi has joined #openstack-containers | 05:02 | |
| *** osmanlicilegi has left #openstack-containers | 05:03 | |
| *** ykarel|away is now known as ykarel | 05:43 | |
| *** ykarel is now known as ykarel|afk | 05:48 | |
| *** ykarel|afk has quit IRC | 05:54 | |
| *** tbarron has quit IRC | 05:54 | |
| *** udesale has joined #openstack-containers | 05:56 | |
| *** irclogbot_1 has quit IRC | 05:56 | |
| *** irclogbot_2 has joined #openstack-containers | 05:57 | |
| *** lpetrut has joined #openstack-containers | 06:04 | |
| *** ykarel|afk has joined #openstack-containers | 06:10 | |
| *** ykarel|afk is now known as ykarel | 06:14 | |
| *** trident has quit IRC | 06:53 | |
| *** trident has joined #openstack-containers | 06:55 | |
| *** ttsiouts has joined #openstack-containers | 07:00 | |
| *** rcernin has quit IRC | 07:03 | |
| openstackgerrit | juhyun.park proposed openstack/magnum master: make ${INSTANCE_NAME} variable. https://review.opendev.org/688046 | 07:08 |
| openstackgerrit | Spyros Trigazis proposed openstack/python-magnumclient stable/train: Allow cluster config for any cluster state https://review.opendev.org/688047 | 07:10 |
| *** lpetrut has quit IRC | 07:15 | |
| *** tobberydberg has quit IRC | 07:20 | |
| *** ttsiouts has quit IRC | 07:24 | |
| *** ttsiouts has joined #openstack-containers | 07:25 | |
| *** tobberydberg has joined #openstack-containers | 07:26 | |
| *** ttsiouts has quit IRC | 07:30 | |
| *** lpetrut has joined #openstack-containers | 07:34 | |
| *** lpetrut has quit IRC | 07:53 | |
| *** ivve has joined #openstack-containers | 07:55 | |
| *** ykarel is now known as ykarel|lunch | 07:56 | |
| *** ttsiouts has joined #openstack-containers | 08:00 | |
| openstackgerrit | Merged openstack/python-magnumclient stable/train: Allow cluster config for any cluster state https://review.opendev.org/688047 | 08:08 |
| *** lsimngar has joined #openstack-containers | 08:12 | |
| openstackgerrit | Spyros Trigazis proposed openstack/magnum stable/train: ng-6: Add new fields to nodegroup objects https://review.opendev.org/685912 | 08:13 |
| openstackgerrit | Spyros Trigazis proposed openstack/magnum stable/train: ng-7: Adapt parameter and output mappings https://review.opendev.org/685918 | 08:13 |
| openstackgerrit | Spyros Trigazis proposed openstack/magnum stable/train: ng-8: APIs for nodegroup CRUD operations https://review.opendev.org/685919 | 08:13 |
| openstackgerrit | Spyros Trigazis proposed openstack/magnum stable/train: ng-9: Driver for nodegroup operations https://review.opendev.org/685920 | 08:13 |
| openstackgerrit | Spyros Trigazis proposed openstack/magnum stable/train: Drop deprecated APIs for kube v1.16 support https://review.opendev.org/688056 | 08:13 |
| openstackgerrit | Spyros Trigazis proposed openstack/magnum stable/train: Change the order of resource creation https://review.opendev.org/688057 | 08:13 |
| openstackgerrit | Spyros Trigazis proposed openstack/magnum stable/train: Improve log of k8s health status check https://review.opendev.org/688058 | 08:13 |
| openstackgerrit | Spyros Trigazis proposed openstack/magnum stable/train: Set cniVersion for flannel https://review.opendev.org/688059 | 08:13 |
| openstackgerrit | Spyros Trigazis proposed openstack/magnum stable/train: Add hostname-override to kube-proxy https://review.opendev.org/688060 | 08:13 |
| openstackgerrit | Spyros Trigazis proposed openstack/magnum stable/train: k8s_fedora: Label master nodes with kubectl https://review.opendev.org/688061 | 08:13 |
| openstackgerrit | Spyros Trigazis proposed openstack/magnum stable/train: k8s_fedora: Move rp_filter=1 for calico up https://review.opendev.org/688062 | 08:13 |
| openstackgerrit | Spyros Trigazis proposed openstack/magnum stable/train: Delete the ca-rotate api-ref 'ca-rotate' hasn't been supported, 'ERROR: 'rotate_ca_certificate' is not supported by this driver (HTTP 400)' will be returned. So, I think we should supply the api after it's realization. https://review.opendev.org/688063 | 08:13 |
| openstackgerrit | Spyros Trigazis proposed openstack/magnum stable/train: Pass ssh public key as string https://review.opendev.org/688064 | 08:13 |
| openstackgerrit | Spyros Trigazis proposed openstack/magnum stable/train: k8s_atomic: Run all syscontainer with podman https://review.opendev.org/688065 | 08:13 |
| openstackgerrit | Spyros Trigazis proposed openstack/magnum stable/train: Drop deprecated APIs for kube v1.16 support https://review.opendev.org/688056 | 08:17 |
| openstackgerrit | Spyros Trigazis proposed openstack/magnum stable/train: Change the order of resource creation https://review.opendev.org/688057 | 08:17 |
| openstackgerrit | Spyros Trigazis proposed openstack/magnum stable/train: Improve log of k8s health status check https://review.opendev.org/688058 | 08:17 |
| openstackgerrit | Spyros Trigazis proposed openstack/magnum stable/train: Set cniVersion for flannel https://review.opendev.org/688059 | 08:17 |
| openstackgerrit | Spyros Trigazis proposed openstack/magnum stable/train: Add hostname-override to kube-proxy https://review.opendev.org/688060 | 08:17 |
| openstackgerrit | Spyros Trigazis proposed openstack/magnum stable/train: k8s_fedora: Label master nodes with kubectl https://review.opendev.org/688061 | 08:17 |
| openstackgerrit | Spyros Trigazis proposed openstack/magnum stable/train: k8s_fedora: Move rp_filter=1 for calico up https://review.opendev.org/688062 | 08:17 |
| openstackgerrit | Spyros Trigazis proposed openstack/magnum stable/train: Delete the ca-rotate api-ref 'ca-rotate' hasn't been supported, 'ERROR: 'rotate_ca_certificate' is not supported by this driver (HTTP 400)' will be returned. So, I think we should supply the api after it's realization. https://review.opendev.org/688063 | 08:17 |
| openstackgerrit | Spyros Trigazis proposed openstack/magnum stable/train: Pass ssh public key as string https://review.opendev.org/688064 | 08:17 |
| openstackgerrit | Spyros Trigazis proposed openstack/magnum stable/train: k8s_atomic: Run all syscontainer with podman https://review.opendev.org/688065 | 08:17 |
| openstackgerrit | Spyros Trigazis proposed openstack/magnum stable/train: ng-6: Add new fields to nodegroup objects https://review.opendev.org/685912 | 08:21 |
| openstackgerrit | Spyros Trigazis proposed openstack/magnum stable/train: ng-7: Adapt parameter and output mappings https://review.opendev.org/685918 | 08:21 |
| openstackgerrit | Spyros Trigazis proposed openstack/magnum stable/train: ng-8: APIs for nodegroup CRUD operations https://review.opendev.org/685919 | 08:21 |
| openstackgerrit | Spyros Trigazis proposed openstack/magnum stable/train: ng-9: Driver for nodegroup operations https://review.opendev.org/685920 | 08:21 |
| *** dioguerra has quit IRC | 08:25 | |
| openstackgerrit | Spyros Trigazis proposed openstack/magnum stable/train: Drop deprecated APIs for kube v1.16 support https://review.opendev.org/688056 | 08:40 |
| openstackgerrit | Spyros Trigazis proposed openstack/magnum stable/train: Change the order of resource creation https://review.opendev.org/688057 | 08:40 |
| openstackgerrit | Spyros Trigazis proposed openstack/magnum stable/train: Improve log of k8s health status check https://review.opendev.org/688058 | 08:40 |
| openstackgerrit | Spyros Trigazis proposed openstack/magnum stable/train: Set cniVersion for flannel https://review.opendev.org/688059 | 08:40 |
| openstackgerrit | Spyros Trigazis proposed openstack/magnum stable/train: Add hostname-override to kube-proxy https://review.opendev.org/688060 | 08:40 |
| openstackgerrit | Spyros Trigazis proposed openstack/magnum stable/train: k8s_fedora: Label master nodes with kubectl https://review.opendev.org/688061 | 08:40 |
| openstackgerrit | Spyros Trigazis proposed openstack/magnum stable/train: k8s_fedora: Move rp_filter=1 for calico up https://review.opendev.org/688062 | 08:40 |
| openstackgerrit | Spyros Trigazis proposed openstack/magnum stable/train: Delete the ca-rotate api-ref https://review.opendev.org/688063 | 08:40 |
| openstackgerrit | Spyros Trigazis proposed openstack/magnum stable/train: Pass ssh public key as string https://review.opendev.org/688064 | 08:40 |
| openstackgerrit | Spyros Trigazis proposed openstack/magnum stable/train: k8s_atomic: Run all syscontainer with podman https://review.opendev.org/688065 | 08:40 |
| *** ykarel|lunch is now known as ykarel | 08:43 | |
| strigazi | brtknr: flwang1 here? | 08:45 |
| brtknr | strigazi: hey :) ive been +2ing the backports | 08:46 |
| strigazi | brtknr :) | 08:46 |
| brtknr | im surprised it will be 9.1.0 instead of 9.0.0 | 08:46 |
| strigazi | this is just a topic name | 08:46 |
| strigazi | brtknr: it will be determinted in the releases repo | 08:47 |
| brtknr | strigazi: fair enough | 08:48 |
| brtknr | strigazi: looks like train rc2 got merged yesterday | 08:49 |
| *** dioguerra has joined #openstack-containers | 08:49 | |
| strigazi | 2019-10-10 was a deadline by the release team. | 08:49 |
| brtknr | strigazi: is there a deadline for a final release? | 08:50 |
| brtknr | or is that when we want? | 08:50 |
| strigazi | there is, I don't know when exactly | 08:50 |
| ykarel | brtknr, strigazi they will just tag final rc with the ga tag:- https://review.opendev.org/#/c/687991/ | 08:55 |
| ykarel | and would possibly merge on 16th | 08:56 |
| ykarel | https://review.opendev.org/#/c/687991/1/deliverables/train/magnum.yaml | 08:56 |
| brtknr | thanks ykarel :) | 08:57 |
| brtknr | i suppose we can try to change that if we want | 08:57 |
| strigazi | thanks ykarel | 08:59 |
| ykarel | i think u would need rc3 tag with new patches, then it should be possible to adjust ^^ | 08:59 |
| ykarel | but good to confirm from release team | 08:59 |
| ykarel | as rc deadline is today as per release schedule | 09:00 |
| strigazi | where is flwang1 , we need to merge stuff | 09:00 |
| strigazi | brtknr: that is why I said 9.1.0. | 09:00 |
| strigazi | or implied, not said | 09:00 |
| *** ivve has quit IRC | 09:10 | |
| openstackgerrit | Spyros Trigazis proposed openstack/magnum master: [WIP] Support Fedora CoreOS 30 https://review.opendev.org/678458 | 09:13 |
| openstackgerrit | Merged openstack/magnum stable/train: ng-6: Add new fields to nodegroup objects https://review.opendev.org/685912 | 09:15 |
| *** ivve has joined #openstack-containers | 09:18 | |
| brtknr | ttsiouts: my cluster is still in UPGRADE_IN_PROGRESS since yesterday | 09:21 |
| brtknr | when I look at the nodegroups, only the worker is in UPDATE_IN_PROGRESS | 09:22 |
| ttsiouts | brtknr: how about the stacks? | 09:22 |
| brtknr | ttsiouts: stack is COMPLETE | 09:22 |
| ttsiouts | brtknr: checking | 09:23 |
| brtknr | flwang1: thanks for +2+w | 09:23 |
| openstackgerrit | Merged openstack/magnum stable/train: ng-7: Adapt parameter and output mappings https://review.opendev.org/685918 | 09:26 |
| openstackgerrit | Merged openstack/magnum stable/train: ng-8: APIs for nodegroup CRUD operations https://review.opendev.org/685919 | 09:26 |
| openstackgerrit | Merged openstack/magnum stable/train: ng-9: Driver for nodegroup operations https://review.opendev.org/685920 | 09:26 |
| ttsiouts | brtknr: conductor is running fine? | 09:32 |
| ttsiouts | brtknr: any weird logs? | 09:33 |
| brtknr | ttsiouts: checking | 09:33 |
| brtknr | so if i upgrade the cluster, i need to upgrade the nodegroup separately as a second step right? | 09:34 |
| strigazi | brtknr: can you leave a quick comment here: https://review.opendev.org/#/c/678458/14..15 | 09:34 |
| ttsiouts | brtknr: the non-default nodegroup has to be upgraded separately | 09:35 |
| brtknr | ttsiouts: that makes sense now | 09:40 |
| ttsiouts | brtknr: can you share the output of openstack coe nodegroup show <cluster> <your worker> | 09:42 |
| brtknr | ttsiouts: wait i am repeating everything just to make sure | 09:43 |
| brtknr | ttsiouts: just realised that deleting a cluster template is allowed when a non-default nodegroup is using it | 09:49 |
| brtknr | not sure if this is a bug? | 09:49 |
| *** ttsiouts has quit IRC | 09:57 | |
| *** ttsiouts has joined #openstack-containers | 09:58 | |
| brtknr | ttsiouts: false alarm | 10:01 |
| brtknr | all working well | 10:01 |
| brtknr | but i noticed that when i created a nodegroup, the node wasnt able to join the cluster first time | 10:02 |
| brtknr | not sure how repeatable it is | 10:02 |
| brtknr | i deleted the nodegroup and tried creating it again and it joined fine second time | 10:02 |
| *** ttsiouts has quit IRC | 10:02 | |
| brtknr | http://paste.openstack.org/show/782923/ | 10:04 |
| *** rcernin has joined #openstack-containers | 10:09 | |
| *** rcernin has quit IRC | 10:21 | |
| *** jmlowe has quit IRC | 10:24 | |
| *** jmlowe has joined #openstack-containers | 10:27 | |
| openstackgerrit | Merged openstack/magnum stable/train: Update TOX/UPPER_CONSTRAINTS_FILE for stable/train https://review.opendev.org/685413 | 10:28 |
| *** ivve has quit IRC | 10:28 | |
| brtknr | strigazi: what else is required for FCOS driver? | 10:36 |
| *** ivve has joined #openstack-containers | 10:44 | |
| *** dave-mccowan has joined #openstack-containers | 10:59 | |
| *** ttsiouts has joined #openstack-containers | 11:01 | |
| *** ttsiouts has quit IRC | 11:06 | |
| *** ttsiouts has joined #openstack-containers | 11:06 | |
| strigazi | brtknr: nothing, I'm writing the UTs and we take it. | 11:07 |
| strigazi | brtknr: Can you run conformance for calico? I will run for flannel. | 11:07 |
| brtknr | strigazi: full? | 11:08 |
| strigazi | brtknr: just 'sonobuoy run' with https://github.com/vmware-tanzu/sonobuoy/releases/tag/v0.16.1 | 11:08 |
| strigazi | yes. I think it is 200 tests or something | 11:08 |
| strigazi | I have only one dev machine I can't run them in parallel | 11:09 |
| ttsiouts | brtknr: the NGs do not directly reference the CT. So in order to have some control we allow only upgrades using the CT that the cluster uses | 11:09 |
| strigazi | brtknr: create a cluster with 2 worker nodes (some tests need it) | 11:09 |
| strigazi | ttsiouts: kubecluster.yaml is ok for the FCOS driver? | 11:10 |
| strigazi | ttsiouts: you will patch it with the NG patch? | 11:11 |
| brtknr | strigazi: happy to run tests for both flannel and calico in parallel | 11:11 |
| brtknr | i'm | 11:11 |
| ttsiouts | strigazi: I will rebase ng-series on top of fcos | 11:11 |
| strigazi | ttsiouts: pl | 11:11 |
| strigazi | ttsiouts: ok | 11:11 |
| ttsiouts | strigazi: doing it now | 11:11 |
| strigazi | brtknr: run first one test, this one: | 11:12 |
| strigazi | brtknr: sonobuoy_0.16.1_linux_amd64 run --e2e-focus "should provide DNS for services" | 11:12 |
| strigazi | brtknr: if this passes the rest will pass | 11:12 |
| brtknr | strigazi: cool will do that | 11:13 |
| strigazi | brtknr: sounds reasonable? you have any question? | 11:13 |
| brtknr | strigazi: all understoond | 11:13 |
| strigazi | brtknr: I was annoyed by it a couple of time so I'm just saying again the a test or two need two workers :) | 11:14 |
| primeministerp | brtknr: using openstack/magnum deployed with kolla-ansible using Stein. | 11:14 |
| brtknr | strigazi: understood, my default cluster always have 2 workers | 11:15 |
| brtknr | strigazi: https://github.com/brtknr/magnum-cluster/blob/master/site.sh | 11:15 |
| primeministerp | brtknr, it's a basic fedora-atomic template | 11:15 |
| primeministerp | brtknr: i've gotten to point now where I get this error: 'Failed to create certificates for Cluster' | 11:23 |
| brtknr | primeministerp: are you using 8.1.0? | 11:27 |
| brtknr | or 8.0.0? | 11:27 |
| brtknr | ttsiouts: wouldnt it be better to rebase fcos on top of ngs? | 11:37 |
| brtknr | ttsiouts: also i tried to say this to you earlier: | 11:38 |
| brtknr | ttsiouts: false alarm | 11:38 |
| brtknr | all working well | 11:38 |
| brtknr | but i noticed that when i created a nodegroup, the node wasnt able to join the cluster first time | 11:38 |
| brtknr | not sure how repeatable it is | 11:38 |
| brtknr | i deleted the nodegroup and tried creating it again and it joined fine second time | 11:38 |
| brtknr | http://paste.openstack.org/show/782923/ | 11:38 |
| ttsiouts | brtknr: we could rebase fcos on top of ngs. | 11:41 |
| ttsiouts | brtknr: I just wanted to quickly test it | 11:42 |
| ttsiouts | brtknr: I have no problem. | 11:42 |
| ttsiouts | strigazi: thoughts? ^ | 11:42 |
| brtknr | ultimately we need both so guess it doesnt matter too much | 11:43 |
| ttsiouts | brtknr: regarding the node that failed to join, I cannot say much without looking at logs | 11:43 |
| brtknr | if fcos is on top, we need to make sure that whats underneath works with it afterwards | 11:44 |
| ttsiouts | brtknr: just noticed this in the logs: | 11:46 |
| ttsiouts | 1 09:53:49.550890 2940 reflector.go:123] k8s.io/kubernetes/pkg/kubelet/kubelet.go:450: Failed to list *v1.Service: Get https://:6443/api/v1/services?limit=500&re | 11:46 |
| ttsiouts | https://:6443/ | 11:47 |
| ttsiouts | brtknr: was the cluster not created when you created the NG? | 11:47 |
| ttsiouts | brtknr: seems that the api address was missing at the time | 11:48 |
| brtknr | strigazi: both minimal tests passed | 12:00 |
| brtknr | shall I do the full run? | 12:00 |
| *** lsimngar has quit IRC | 12:04 | |
| *** lsimngar has joined #openstack-containers | 12:08 | |
| strigazi | brtknr: yeap | 12:22 |
| brtknr | strigazi: already running! | 12:25 |
| *** lsimngar has quit IRC | 12:31 | |
| ttsiouts | brtknr: I just upgraded my kube-fcos cluster from v1.16.0 to v1.16.1 (both default and non-default NGs)!!! | 12:32 |
| ttsiouts | brtknr: everything looks good! | 12:32 |
| openstackgerrit | Spyros Trigazis proposed openstack/magnum master: Support Fedora CoreOS 30 https://review.opendev.org/678458 | 12:32 |
| strigazi | ttsiouts: no code changes here ^^ | 12:33 |
| strigazi | this was under NGs or on top when you tested? | 12:33 |
| ttsiouts | strigazi: under | 12:34 |
| *** ramishra has quit IRC | 12:42 | |
| openstackgerrit | Bharat Kunwar proposed openstack/magnum master: Convert fixed_subnet name to uuid for OCCM https://review.opendev.org/629514 | 12:43 |
| openstackgerrit | Bharat Kunwar proposed openstack/magnum master: Convert fixed_subnet name to uuid for OCCM https://review.opendev.org/629514 | 12:49 |
| primeministerp | brtknr: yes | 12:52 |
| primeministerp | brtknr: it's 8.1.0 | 12:52 |
| brtknr | primeministerp: where do you see 'Failed to create certificates for Cluster' | 12:53 |
| primeministerp | brtknr: here's my magnum.conf http://paste.openstack.org/show/783004/ | 12:53 |
| brtknr | primeministerp: do you have dns nameserver configured correctly for your cluster subnet? | 12:53 |
| brtknr | i.e. can you curl google.com from one of your nodes? | 12:54 |
| primeministerp | brtknr: let me check, I don't think it get's that far | 12:55 |
| strigazi | brtknr: how we take the patches in? | 12:56 |
| strigazi | brtknr: what do you prefer? | 12:56 |
| strigazi | fcos first or after | 12:56 |
| strigazi | ? | 12:56 |
| strigazi | cc ttsiouts | 12:56 |
| ttsiouts | strigazi: I'm ok either way. I am ready to push if we choose to rebase ng on top of fcos | 12:58 |
| primeministerp | brtknr: here's the magnum-conductor.log w/ the error http://paste.openstack.org/show/783005/ | 12:58 |
| brtknr | might be easier to do ng after fcos first if its ready? | 12:59 |
| strigazi | brtknr: as you want | 12:59 |
| brtknr | might be easier to do ng before fcos if its ready? | 12:59 |
| brtknr | ^^ | 13:00 |
| strigazi | brtknr: ng first? | 13:00 |
| strigazi | both are ready | 13:00 |
| brtknr | ng first yes | 13:01 |
| openstackgerrit | Theodoros Tsioutsias proposed openstack/magnum master: ng-10: Fix cluster template conditions https://review.opendev.org/685620 | 13:02 |
| openstackgerrit | Theodoros Tsioutsias proposed openstack/magnum master: ng-11: API microversion 1.9 https://review.opendev.org/686089 | 13:02 |
| openstackgerrit | Theodoros Tsioutsias proposed openstack/magnum master: ng-12: Label nodegroup nodes https://review.opendev.org/686362 | 13:02 |
| openstackgerrit | Theodoros Tsioutsias proposed openstack/magnum master: ng-13: Support nodegroup upgrade https://review.opendev.org/686733 | 13:02 |
| ttsiouts | brtknr: strigazi: ng-13 supports fcos also! | 13:03 |
| ttsiouts | ng-10 and ng-12 take fcos under consideration too | 13:04 |
| brtknr | ttsiouts: are you saying fcos first? | 13:04 |
| ttsiouts | brtknr: strigazi: sorry I misunderstood | 13:06 |
| brtknr | okay... ng-10->13 depend on fcos but fcos doesnt depend on ng | 13:06 |
| brtknr | so lets do fcos first then ng | 13:06 |
| brtknr | strigazi: ttsiouts ^ | 13:07 |
| brtknr | we're overthinking this lol | 13:07 |
| ttsiouts | brtknr: strigazi: I'm ok either way. sorry guys.. friday afternoon hit me hard | 13:09 |
| *** damiandabrowski[ has joined #openstack-containers | 13:12 | |
| primeministerp | brtknr: should I downgrade to 8.0.0? | 13:14 |
| primeministerp | brtknr: not sure I mentioned too this is on aarch64 | 13:16 |
| primeministerp | brtknr: dns is properly configured standalone vms work just fine in the same network | 13:18 |
| brtknr | ive never tried with aarch64 but i know that flannel driver only works with amd64 | 13:20 |
| brtknr | primeministerp: you should try the calico driver | 13:20 |
| primeministerp | brtknr can I do that w/ kolla-ansible easily? | 13:20 |
| primeministerp | brtknr: also are there docs for the calico driver? | 13:20 |
| brtknr | primeministerp: you just change driver to calico when you deploy cluster | 13:21 |
| primeministerp | i'll switch that and try | 13:21 |
| primeministerp | brtknr: and why wouldn't the flannel driver work w/ aarch64, are there container images used to deploy it? | 13:23 |
| primeministerp | brtknr: same thing w/ the certs btw | 13:23 |
| primeministerp | brtknr: the instances never get started | 13:24 |
| *** ttsiouts has quit IRC | 13:25 | |
| *** ttsiouts has joined #openstack-containers | 13:26 | |
| brtknr | primeministerp: check journalctl -u heat-container-agent log for errors | 13:26 |
| *** ttsiouts has quit IRC | 13:31 | |
| *** goldyfruit has joined #openstack-containers | 13:31 | |
| primeministerp | brtknr: will that log only be in the container? | 13:41 |
| primeministerp | brtknr: i'm using kolla to run everything | 13:42 |
| brtknr | yes, the master | 13:42 |
| brtknr | the master node | 13:42 |
| brtknr | the workers only start after the master sends a signal to heat | 13:42 |
| primeministerp | I don't have that specific log | 13:42 |
| primeministerp | o you mean in the vm of the k8s cluster | 13:42 |
| primeministerp | It never get's that far | 13:42 |
| primeministerp | fails before that | 13:42 |
| *** ramishra has joined #openstack-containers | 13:50 | |
| *** ttsiouts has joined #openstack-containers | 13:55 | |
| brtknr | primeministerp: check /var/log/cloud-init.log | 13:56 |
| brtknr | primeministerp: check /var/log/cloud-init-output.log | 13:56 |
| brtknr | both | 13:56 |
| primeministerp | brtknr, it never gets that far | 13:57 |
| primeministerp | brtknr, no vms get started | 13:57 |
| *** udesale has quit IRC | 13:57 | |
| brtknr | primeministerp: oh | 13:58 |
| *** udesale has joined #openstack-containers | 13:58 | |
| brtknr | did it ever work before? | 13:58 |
| primeministerp | meaning launching vms? | 13:59 |
| primeministerp | yes | 13:59 |
| primeministerp | I can launch them standalone | 13:59 |
| primeministerp | it's an AIO | 13:59 |
| primeministerp | I'm returning my config to the default kolla-ansible config for magnum now | 13:59 |
| primeministerp | brtknr: do you know if the default config should work? | 14:00 |
| primeministerp | brtknr: can i deploy it w/o barbician? | 14:00 |
| brtknr | primeministerp: yes default magnum config should work | 14:01 |
| primeministerp | brtknr: do I leave barbician or should I remove it? | 14:04 |
| brtknr | primeministerp: not sure where barbican comes in tbh | 14:05 |
| primeministerp | brtknr: removing it then | 14:06 |
| primeministerp | brtknr: i'm trying to get to enough to work | 14:06 |
| brtknr | primeministerp: you should ask in the #openstack-kolla channel about how to get magnum deployed via kolla | 14:07 |
| brtknr | primeministerp: they will point you to the minimum reference configuration you will need | 14:07 |
| primeministerp | it deploys, I just need to configured so this will launch | 14:07 |
| primeministerp | brtkr: i'll keep you posted. btw, I think that was you that gave the +2 on the typo patch I sent in... thx | 14:09 |
| primeministerp | brtknr: will you be in shanghai? | 14:10 |
| brtknr | primeministerp: no :( flwang1 and strigazi might be though | 14:10 |
| primeministerp | flwang1, strigazi come by the ampere booth if you're going to shanghai | 14:11 |
| strigazi | primeministerp +1 | 14:11 |
| primeministerp | strigazi, you're at CERN? | 14:13 |
| strigazi | primeministerp: yes | 14:14 |
| primeministerp | strigazi, are Tim or Jan going to the summit? | 14:14 |
| strigazi | primeministerp: Jan no | 14:14 |
| primeministerp | strigazi, still running hyper-v in yoru openstack? | 14:14 |
| strigazi | primeministerp: no | 14:14 |
| primeministerp | strigazi, we go way back, to 2011 | 14:14 |
| primeministerp | strigazi, you finally removed it | 14:15 |
| strigazi | primeministerp: I think 18months ago | 14:15 |
| primeministerp | strigazi, I left ms in 2017 | 14:15 |
| primeministerp | strigazi, I remember when you guys were using SCVMM in the pre openstack days | 14:17 |
| strigazi | primeministerp: I wasn't at CERN then :) | 14:17 |
| primeministerp | strigazi, I'm old | 14:17 |
| strigazi | primeministerp: it wasn't that long ago | 14:18 |
| primeministerp | strigazi, well I'll tell you when I worked on openstack full time, it felt like every 6 mos was a year | 14:19 |
| primeministerp | esp at MS | 14:19 |
| primeministerp | they didn't make it easy | 14:20 |
| primeministerp | is anyone else around that can give a +2 to this: https://review.opendev.org/#/c/688017/ | 14:20 |
| primeministerp | also has anyone looked into creating a driving using K3os? | 14:21 |
| primeministerp | er driver | 14:22 |
| strigazi | I don't think so | 14:22 |
| strigazi | it is the racher os right? | 14:22 |
| strigazi | rancher | 14:22 |
| primeministerp | well it's by rancherlabs, but it's not rancheros | 14:22 |
| primeministerp | which I also adore | 14:22 |
| primeministerp | it's a kubernetes distro, you feed in essentially a cloud-init to configure | 14:23 |
| primeministerp | I was looking at the driver examples last night wondering how much work it would be | 14:23 |
| strigazi | not much | 14:23 |
| primeministerp | I wouldn't think so | 14:23 |
| primeministerp | I may try to start trowing something together once I can get this deployment to work properly w/ magnum | 14:24 |
| strigazi | what is its license? | 14:24 |
| primeministerp | it's all open source parts | 14:24 |
| primeministerp | i'm assuming GPLv2 and Apache | 14:24 |
| primeministerp | depending on the bits | 14:24 |
| primeministerp | Apache | 14:25 |
| primeministerp | https://github.com/rancher/k3os/blob/master/LICENSE | 14:25 |
| strigazi | this is the os? https://github.com/rancher/os/releases | 14:25 |
| primeministerp | no | 14:25 |
| primeministerp | this is | 14:25 |
| primeministerp | https://github.com/rancher/k3os | 14:25 |
| primeministerp | they have multiple oses | 14:25 |
| primeministerp | racheros is docker | 14:25 |
| primeministerp | k3os there's no docker | 14:26 |
| primeministerp | containerd+k3s | 14:26 |
| primeministerp | boots right into a cluster | 14:26 |
| primeministerp | essentially | 14:26 |
| primeministerp | you can add nodes kinda like using docker swarm, but it's kubernetes | 14:26 |
| strigazi | they don't seem to have openstack images? | 14:26 |
| primeministerp | it's built in | 14:26 |
| primeministerp | the iso is a raw image | 14:26 |
| primeministerp | i've converted that to qcow2 and booted on openstack | 14:27 |
| primeministerp | the one I'm having problems w/ magnum on | 14:27 |
| primeministerp | i've booted the iso as well | 14:27 |
| strigazi | ok | 14:27 |
| primeministerp | both aarch64 | 14:27 |
| primeministerp | i've also used k3s which is k3os w/o the OS | 14:27 |
| primeministerp | and they have a micro paas for itsio and envoy called rio | 14:27 |
| primeministerp | brtknr, default config now give me this: "cinder" volume driver needs "cloud_provider_enabled" label to be true or unset. | 14:29 |
| brtknr | ah you didnt meantion you were using cinder volumes | 14:30 |
| strigazi | i saw rio is using likerd now | 14:30 |
| primeministerp | brtknr, it forced me to | 14:30 |
| primeministerp | brtknr: it wouldn't work when I tried w/o cinder either | 14:30 |
| *** lsimngar has joined #openstack-containers | 14:30 | |
| brtknr | primeministerp: thats weird, it should work without cinder | 14:31 |
| primeministerp | brtknr, I try to remove them too | 14:31 |
| primeministerp | I wouldn't mind just getting to a bareminmal | 14:31 |
| brtknr | you dont need to remove cinder frm kolla, you should be able to create a k8s cluster without creating cinder volumes | 14:32 |
| primeministerp | now i get this when I set cloud_provider_enabled = true: cluster_user_trust" must be set to True in magnum.conf when "cloud_provider_enabled" label is set to true. | 14:32 |
| brtknr | do you have boot_volume_size>0? | 14:32 |
| primeministerp | I tried that and it would fail | 14:32 |
| brtknr | can you show me your cluster template? | 14:33 |
| primeministerp | it's in terraform, is that ok? | 14:33 |
| *** ykarel is now known as ykarel|away | 14:33 | |
| brtknr | primeministerp: no probs | 14:33 |
| primeministerp | https://github.com/ppouliot/terraform-openstack-magnum-clusters | 14:33 |
| primeministerp | main.tf | 14:34 |
| primeministerp | need to push my current stuff | 14:34 |
| primeministerp | so the whole thing is, i've been adding and removing the cloud_provider enabled and the cluster_user_trust to try to get it work | 14:35 |
| primeministerp | renabling cluster_user_trust = true | 14:36 |
| *** ykarel|away has quit IRC | 14:38 | |
| *** munimeha1 has joined #openstack-containers | 14:40 | |
| primeministerp | brtknr, if it's any consolation that terraform worked in a previous life | 14:41 |
| primeministerp | (back when I was at rakuten) | 14:41 |
| *** goldyfruit_ has joined #openstack-containers | 14:43 | |
| brtknr | primeministerp: do you need to use cinder volume? | 14:44 |
| primeministerp | brtknr: do you have an example of the simplest template? | 14:44 |
| primeministerp | brtknr: i can remove it | 14:44 |
| brtknr | primeministerp: https://github.com/stackhpc/kubeflow-demo/blob/master/k8s.tf | 14:44 |
| brtknr | primeministerp: main diff is docker_storage_driver = "overlay2" | 14:45 |
| primeministerp | I'll try it | 14:45 |
| primeministerp | i tried overlay two yesterday when i removed the cinder bits and I was still getting failures | 14:45 |
| *** goldyfruit has quit IRC | 14:46 | |
| primeministerp | it seemed to come down to the cloud_provider_enabled = true in the labels | 14:46 |
| primeministerp | and the cluster_user_trust = true in the magnum.conf | 14:46 |
| primeministerp | which from the docs are required to use fedora-atomic | 14:46 |
| brtknr | you need cloud_provider_enabled=true for cinder volumes | 14:47 |
| primeministerp | it still tried using volumes it seemed even when I turned them off in the template | 14:47 |
| primeministerp | i'll try again w/ those changes and see | 14:48 |
| primeministerp | it would be better if I didn't have to use cinder for my current use case | 14:48 |
| *** ivve has quit IRC | 14:50 | |
| brtknr | primeministerp: use a more up to date version of k8s | 14:51 |
| brtknr | 1.9.1 is way too old | 14:51 |
| *** KeithMnemonic has joined #openstack-containers | 14:57 | |
| *** lsimngar has quit IRC | 15:01 | |
| *** lsimngar has joined #openstack-containers | 15:01 | |
| *** lsimngar has quit IRC | 15:03 | |
| *** FlorianFa has quit IRC | 15:03 | |
| *** lpetrut has joined #openstack-containers | 15:04 | |
| *** ttsiouts has quit IRC | 15:08 | |
| *** ykarel|away has joined #openstack-containers | 15:14 | |
| openstackgerrit | Merged openstack/magnum master: Fixing typos and spelling errors in driver template files https://review.opendev.org/688017 | 15:18 |
| *** ykarel|away is now known as ykarel | 15:21 | |
| *** udesale has quit IRC | 15:23 | |
| *** ykarel is now known as ykarel|afk | 15:55 | |
| *** lpetrut has quit IRC | 16:05 | |
| *** ivve has joined #openstack-containers | 16:11 | |
| *** ramishra has quit IRC | 16:23 | |
| *** ivve has quit IRC | 18:09 | |
| *** ivve has joined #openstack-containers | 18:10 | |
| *** pcaruana has quit IRC | 18:55 | |
| *** goldyfruit___ has joined #openstack-containers | 19:15 | |
| *** goldyfruit_ has quit IRC | 19:17 | |
| primeministerp | brknr: what's latest supported version? | 19:32 |
| primeministerp | brtknr: it's been a while | 19:33 |
| primeministerp | brtknr: default config of kolla-ansible with setup to use cinder gives me this as the error: 'NoneType' object has no attribute 'is_admin' | 19:34 |
| primeministerp | brtknr: using magnum to start the cluster | 19:35 |
| primeministerp | brtknr: I see that under 'Status Reason' in horizion | 19:36 |
| primeministerp | under the OS::Magnum::Cluster/<myclusteruid> link | 19:36 |
| *** ykarel|afk has quit IRC | 19:50 | |
| brtknr | Should try 1.14.6 with Stein | 19:55 |
| openstackgerrit | Merged openstack/magnum stable/train: Drop deprecated APIs for kube v1.16 support https://review.opendev.org/688056 | 20:16 |
| openstackgerrit | Merged openstack/magnum stable/train: Change the order of resource creation https://review.opendev.org/688057 | 20:16 |
| openstackgerrit | Merged openstack/magnum stable/train: Improve log of k8s health status check https://review.opendev.org/688058 | 20:16 |
| openstackgerrit | Merged openstack/magnum stable/train: Set cniVersion for flannel https://review.opendev.org/688059 | 20:17 |
| openstackgerrit | Merged openstack/magnum stable/train: Add hostname-override to kube-proxy https://review.opendev.org/688060 | 20:17 |
| openstackgerrit | Merged openstack/magnum stable/train: k8s_fedora: Label master nodes with kubectl https://review.opendev.org/688061 | 20:18 |
| openstackgerrit | Merged openstack/magnum stable/train: k8s_fedora: Move rp_filter=1 for calico up https://review.opendev.org/688062 | 20:18 |
| openstackgerrit | Merged openstack/magnum stable/train: Delete the ca-rotate api-ref https://review.opendev.org/688063 | 20:18 |
| openstackgerrit | Merged openstack/magnum stable/train: Pass ssh public key as string https://review.opendev.org/688064 | 20:18 |
| openstackgerrit | Merged openstack/magnum stable/train: k8s_atomic: Run all syscontainer with podman https://review.opendev.org/688065 | 20:18 |
| primeministerp | brtknr: looks like dropping cinder at least started spawing the master | 20:25 |
| primeministerp | brtknr: i used the terraform template you sent | 20:26 |
| primeministerp | taking a while | 20:26 |
| primeministerp | at 31min for deploy of master | 20:26 |
| *** munimeha1 has quit IRC | 21:22 | |
| brtknr | primeministerp: sounds like something is terribly misconfigured | 22:05 |
| brtknr | Even our baremetal setup doesn’t take that long | 22:05 |
| *** goldyfruit___ has quit IRC | 22:08 | |
| *** KeithMnemonic has quit IRC | 22:38 | |
| *** rcernin has joined #openstack-containers | 22:42 | |
| *** ivve has quit IRC | 23:07 | |
Generated by irclog2html.py 2.15.3 by Marius Gedminas - find it at mg.pov.lt!