You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
I found this issue while upgrading the @pulumi/eks package on our cluster project. We use providerCredentialOpts.profileName on our Cluster resource to set the profile at create time (we have profiles managed inside the project with names based on stack configuration, this could just be consistent profile names we use in our org though). The provider used to create the initial resources (suffix -eks-k8s) has a Kubeconfig with the profile included but an additional provider is created (suffix -provider) that does not have the profile available and tries to use the default credential chain. We advise users not to use default internally so the error from that provider is: Unable to locate credentials. You can configure credentials by running "aws configure".
I create a provider on the outside of the Cluster resource for use later but needed to switch from using eksCluster.kubeconfig to eksCluster.getKubeconfig({ profileName: awsProfile }) so that internally created provider is unlikely to cause us issues but it is still there and will spook users when they see the error.
I think this is the change where this was introduced: #1082. My comment on the topic is here: #1082 (comment). I do not understand the logic of having a single stack where a profile is needed for the cluster and another profile is set by the user.
Contributing
Vote on this issue by adding a 👍 reaction.
To contribute a fix for this issue, leave a comment (and link to your pull request, if you've opened one already).
The text was updated successfully, but these errors were encountered:
I am going to close this now as the resource in question has been quietly removed from the codebase in this PR: #1326 (seems unrelated to the resources in the title, like a cleanup was just snuck in there).
What happened?
I found this issue while upgrading the
@pulumi/eks
package on our cluster project. We useproviderCredentialOpts.profileName
on ourCluster
resource to set the profile at create time (we have profiles managed inside the project with names based on stack configuration, this could just be consistent profile names we use in our org though). The provider used to create the initial resources (suffix-eks-k8s
) has a Kubeconfig with the profile included but an additional provider is created (suffix-provider
) that does not have the profile available and tries to use the default credential chain. We advise users not to use default internally so the error from that provider is:Unable to locate credentials. You can configure credentials by running "aws configure".
I create a provider on the outside of the
Cluster
resource for use later but needed to switch from usingeksCluster.kubeconfig
toeksCluster.getKubeconfig({ profileName: awsProfile })
so that internally created provider is unlikely to cause us issues but it is still there and will spook users when they see the error.Example
Output of
pulumi about
Additional context
I think this is the change where this was introduced: #1082. My comment on the topic is here: #1082 (comment). I do not understand the logic of having a single stack where a profile is needed for the cluster and another profile is set by the user.
Contributing
Vote on this issue by adding a 👍 reaction.
To contribute a fix for this issue, leave a comment (and link to your pull request, if you've opened one already).
The text was updated successfully, but these errors were encountered: