2654 Members
talos.dev API-driven Kubernetes OS44 Servers

Load older messages

20 Jun 2024
@_slack_taloscommunity_U06TNE3HFMW:matrix.orgRoché Compaan The attached config fails with:
│ Error: Error applying configuration
│   with talos_machine_configuration_apply.controlplane[1],
│   on talos_cluster.tf line 182, in resource "talos_machine_configuration_apply" "controlplane":
│  182: resource "talos_machine_configuration_apply" "controlplane" {
│ rpc error: code = Unknown desc = failed to parse config: unknown keys found during decoding:
│ machine:
│     features:
│         hostDNS:
│             enabled: true
│     kubelet:
│         credentialProviderConfig:
│             apiVersion: kubelet.config.k8s.io/v1
│             kind: CredentialProviderConfig
│             providers:
│                 - apiVersion: credentialprovider.kubelet.k8s.io/v1
│                   defaultCacheDuration: 12h
│                   matchImages:
│                     - '.dkr.ecr..amazonaws.com'
│                     - '.dkr.ecr..amazonaws.com.cn'
│                     - '.dkr.ecr-fips..amazonaws.com'
│                     - '.dkr.ecr.us-iso-east-1.c2s.ic.gov'
│                     - '.dkr.ecr.us-isob-east-1.sc2s.sgov.gov'
│                   name: ecr-credential-provider
"Unknown keys" strikes me as an error that relates to the particular version of Talos that is used to bootstrap the ec2 instances.
@_slack_taloscommunity_U06TNE3HFMW:matrix.orgRoché Compaan I'm doing this annoying dance where I comment config available in v1.7, provision the cluster, upgrade the nodes and then uncomment the config to apply it. 10:09:11
@_slack_taloscommunity_U078L6DQM6K:matrix.orgZackeus Bengtsson joined the room.14:48:52
@_slack_taloscommunity_U078L6DQM6K:matrix.orgZackeus Bengtsson set a profile picture.14:49:00
@_slack_taloscommunity_U07931W70NN:matrix.orgManish Kalra joined the room.21:14:23
@_slack_taloscommunity_U07931W70NN:matrix.orgManish Kalra changed their display name from _slack_taloscommunity_U07931W70NN to Manish Kalra.21:14:24
@_slack_taloscommunity_U07931W70NN:matrix.orgManish Kalra set a profile picture.21:14:26
@charllopinkk:matrix.orgcharllopinkk joined the room.23:18:14
21 Jun 2024
@_slack_taloscommunity_U03SK7M9FGS:matrix.orgBoran Car Does anyone have a Rock 4 SE to do additional testing of this PR? https://github.com/siderolabs/sbc-rockchip/pull/18 00:21:26
@_slack_taloscommunity_U03SK7M9FGS:matrix.orgBoran Car I've tested it on multiple Rock 4 SE V1.53, but that's the only version of the boards I have. 00:21:52
@_slack_taloscommunity_U0793H8JC2F:matrix.orgKyle Huggins joined the room.00:25:29
@_slack_taloscommunity_U0793H8JC2F:matrix.orgKyle Huggins changed their display name from _slack_taloscommunity_U0793H8JC2F to Kyle Huggins.00:25:30
@charllopinkk:matrix.orgcharllopinkk left the room.00:54:39
@_slack_taloscommunity_U078Z3DC4NN:matrix.orgThomas changed their profile picture.01:37:16
@_slack_taloscommunity_U05URSU8X1Q:matrix.orgAdri Shahri In my k8s cluster I have 2 workers in which both of them are in different locations the workload from one node unable to communicate (ping) with workload from another node. In my case, i have an ingress on location-a and my apps scheduled in location-b and I want to use ingress in location-a for my apps in location-b. p/s: I have enabled kubespan across all nodes and it seem the kubespan status from kubespanpeerstatuses is unknown from worker in location-a to location-b, vice-versa. anyone has any ideas on how can I solve this issue ? 🙂 03:56:38
@_slack_taloscommunity_U06QW32Q44V:matrix.orgKevin Clinthorne changed their profile picture.04:05:28
@_slack_taloscommunity_U05URSU8X1Q:matrix.orgAdri Shahri Do we still need to configure wireguard manually if we have already enabled the kubespan? https://www.talos.dev/v1.7/talos-guides/network/wireguard-network/#manual-configuration 04:05:52
@_slack_taloscommunity_UG8G8UMMG:matrix.orgTim Jones No, KubeSpan will automatically create the mesh, but you need at least one node to be publicly accessible to bring the mesh up. 06:15:30
@_slack_taloscommunity_U05URSU8X1Q:matrix.orgAdri Shahri By publicly accessible you mean to assign a static public IP to that one node right? What if both nodes at both locations doesnt have static public IPs ? any workarounds ? 07:12:41
@_slack_taloscommunity_UG8G8UMMG:matrix.orgTim Jones Not really, since KubeSpan relies on at least one node being reachable on the UDP port announced to the discovery service. 07:43:01
@ascalon:matrix.orgascalon joined the room.07:49:30
@_slack_taloscommunity_U05URSU8X1Q:matrix.orgAdri Shahri Got it 08:05:45
@_slack_taloscommunity_U06QWUJ2WTC:matrix.orgThibault (enix.io) Seems like this was introduced on 1.15 because I could not reproduce the issue with Cilium 1.14.12 However 1.16.0-rc.0 is affected just like 1.15 12:32:34
@_slack_taloscommunity_U0797HBC09G:matrix.orgKristian Kostecky joined the room.15:53:00
@_slack_taloscommunity_U0797HBC09G:matrix.orgKristian Kostecky changed their display name from _slack_taloscommunity_U0797HBC09G to Kristian Kostecky.15:53:03
@_slack_taloscommunity_U04PR92T6MD:matrix.orgRomain Degez Thanks for the confirmation and workaround Thibault (enix.io) cc frezbo, maybe a warning on https://www.talos.dev/v1.7/kubernetes-guides/network/deploying-cilium/ would avoid people falling into the same trap and be careful with cilium > 1.14 when using Talos with a VIP ? (Tell me if this info should be broadcasted somewhere else ? 16:00:44
@_slack_taloscommunity_U01URQANXC0:matrix.orgfrezbo yeh, we could add a note 16:01:39
@_slack_taloscommunity_U04PR92T6MD:matrix.orgRomain Degez Would you prefer us to send a PR to update https://github.com/siderolabs/talos/blob/main/website/content/v1.7/kubernetes-guides/network/deploying-cilium.md ? or do you have another workflow to update the website ? 16:03:44
@_slack_taloscommunity_U01URQANXC0:matrix.orgfrezbo yes, that would be it, PR are super welcome ❤️ 16:04:05
@_slack_taloscommunity_U04PR92T6MD:matrix.orgRomain Degez Thibault (enix.io) when you get a chance... 🙂 16:04:27

There are no newer messages yet.

Back to Room ListRoom Version: 5