Open reganmcdonalds4 opened 3 months ago
Can you rerun the command with debug logging enabled, by passing -v4
?
Here is the output when passing -v4
. The command exits while the update is still in progress:
2024-07-29 11:20:00 [▶] Setting credentials expiry window to 30 minutes
2024-07-29 11:20:01 [▶] role ARN for the current session is "arn:aws:iam:::user/"
2024-07-29 11:20:01 [▶] cluster = &types.Cluster{AccessConfig:(*types.AccessConfigResponse)(0x14000645960), Arn:(*string)(0x14000656a80), CertificateAuthority:(*types.Certificate)(0x14000656a30), ClientRequestToken:(*string)(nil), ConnectorConfig:(*types.ConnectorConfigResponse)(nil), CreatedAt:time.Date(2019, time.June, 18, 21, 51, 50, 163000000, time.UTC), EncryptionConfig:[]types.EncryptionConfig(nil), Endpoint:(*string)(0x14000656a60), Health:(*types.ClusterHealth)(0x14000645900), Id:(*string)(nil), Identity:(*types.Identity)(0x14000656aa0), KubernetesNetworkConfig:(*types.KubernetesNetworkConfigResponse)(0x14000b821e0), Logging:(*types.Logging)(0x14000645920), Name:(*string)(0x14000656a10), OutpostConfig:(*types.OutpostConfigResponse)(nil), PlatformVersion:(*string)(0x14000656a50), ResourcesVpcConfig:(*types.VpcConfigResponse)(0x1400004b420), RoleArn:(*string)(0x14000656a20), Status:"ACTIVE", Tags:map[string]string{}, Version:(*string)(0x14000656ad0), noSmithyDocumentSerde:document.NoSerde{}}
2024-07-29 11:20:02 [ℹ] Kubernetes version "1.30" in use by cluster "cluster-name"
2024-07-29 11:20:02 [▶] addon: &{coredns latest [] map[] {false false false false false false false} map[] <nil> false true [] [] []}
2024-07-29 11:20:02 [▶] resolve conflicts set to OVERWRITE
2024-07-29 11:20:02 [▶] addon: &{coredns latest [] map[] {false false false false false false false} map[] <nil> false true [] [] []}
2024-07-29 11:20:02 [▶] could not parse version "latest", skipping finding newer versions: No Major.Minor.Patch elements found
2024-07-29 11:20:03 [ℹ] updating addon
2024-07-29 11:20:03 [▶] &{AddonName:0x140008862d0 ClusterName:0x140002aea50 AddonVersion:0x140006d8c70 ClientRequestToken:<nil> ConfigurationValues:<nil> PodIdentityAssociations:[] ResolveConflicts:OVERWRITE ServiceAccountRoleArn:<nil> noSmithyDocumentSerde:{}}
2024-07-29 11:20:03 [▶] &{CreatedAt:2024-07-29 15:20:03.331 +0000 UTC Errors:[] Id:0x14000c3a1f0 Params:[{Type:AddonVersion Value:0x14000c3a200 noSmithyDocumentSerde:{}} {Type:ResolveConflicts Value:0x14000c3a210 noSmithyDocumentSerde:{}}] Status:InProgress Type:AddonUpdate noSmithyDocumentSerde:{}}
@cPu1 can I provide any more information?
@cPu1 can I provide any more information?
@reganmcdonalds4 I think that's sufficient. We are occupied with other priorities right now, so please give us some time.
@cPu1 can I provide any more information?
@reganmcdonalds4 I think that's sufficient. We are occupied with other priorities right now, so please give us some time.
Understood. Thank you.
What were you trying to accomplish?
Update EKS addon coredns with eksctl.
What happened?
When updating coredns with the --wait argument, the command exits immediately after beginning the update. We are left unaware if the update was successful or not. In our case, the update failed silently from the perspective of eksctl.
How to reproduce it?
Set minAvailable: 1 for the coredns pod disruption budget and run:
Logs 2024-07-25 13:39:34 [ℹ] Kubernetes version "1.30" in use by cluster "cluster-name" 2024-07-25 13:39:34 [ℹ] updating addon
Versions eksctl version: 0.183.0 kubectl version: v1.30.2 OS: darwin