-
Notifications
You must be signed in to change notification settings - Fork 456
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Support reading environment from secret/configmap #2295
Conversation
d8aebf1
to
5c02e7b
Compare
5c02e7b
to
c244afe
Compare
well... if tests are broken, somenthing is broken, please take a look @ramondeklein |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
So, If I am understanding the code right, sidecar will no longer watch the .spec.config.name
secret and will solely get the env variables from a configMap or a secret mentioned in spec.env
(on a side note: my suggestions is to use spec.envFrom
)
In that case shouldn't the field .spec.configuration
be completelly removed form TenantSpec?
// *Optional* +
//
// Specify a secret that contains additional environment variable configurations to be used for the MinIO pools.
// The secret is expected to have a key named config.env containing all exported environment variables for MinIO+
// +optional
Configuration *corev1.LocalObjectReference `json:"configuration,omitempty"`
type EnvVarSource struct { | ||
// Selects a key of a ConfigMap. | ||
// +optional | ||
ConfigMapKeyRef *ConfigMapKeySelector `json:"configMapKeyRef,omitempty" protobuf:"bytes,3,opt,name=configMapKeyRef"` |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
This could be corev1.ConfigMapKeySelector
, I don't see why we would not honor the Optional
flag that is being removed
ConfigMapKeyRef *ConfigMapKeySelector `json:"configMapKeyRef,omitempty" protobuf:"bytes,3,opt,name=configMapKeyRef"` | |
ConfigMapKeyRef *corev1.ConfigMapKeySelector `json:"configMapKeyRef,omitempty" protobuf:"bytes,3,opt,name=configMapKeyRef"` |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
I did that at first, but code is more complicated, because when the value is not optional we need to generate an error and abort the creation. It adds a lot of code. The current implementation just skips the environment variable (so optional is always enabled) if it cannot find the source.
There is no valid use-case to allow optional/non-optional support (AFAIK), so I prefered to use the simple code.
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
How do we notify that a env variable could not be mounted to the user?, we just silently remove it?
optional
could solve this problem, even if means more code, if optional: false
then we should error out somewhere, if optional: true
we can silently drop the env variable, as the user stated that it is fine to run MinIO without it.
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Yes, I just skipped it. I'll rework to support Optional
and use corev1.ConfigMapKeySelector
(same for secret).
cli.StringFlag{ | ||
Name: "config-name", | ||
Value: "", | ||
Usage: "secret being watched", | ||
}, |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
This is now read from the tenant specification.
type EnvVarSource struct { | ||
// Selects a key of a ConfigMap. | ||
// +optional | ||
ConfigMapKeyRef *ConfigMapKeySelector `json:"configMapKeyRef,omitempty" protobuf:"bytes,3,opt,name=configMapKeyRef"` |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
I did that at first, but code is more complicated, because when the value is not optional we need to generate an error and abort the creation. It adds a lot of code. The current implementation just skips the environment variable (so optional is always enabled) if it cannot find the source.
There is no valid use-case to allow optional/non-optional support (AFAIK), so I prefered to use the simple code.
@pjuarezd I've looked into this and all tests seem to run fine locally. Could it be that they time out when running in GitHub actions? I don't really get why this PR results in issues and some others seem to work fine... |
c244afe
to
9225dc8
Compare
Thank you for fixing the bug breaking the test @ramondeklein! |
Paused implementation, because it looks like #2253 is trying to restart pods anyway. This needs further discussion... |
Operator v7 will revert to old behaviour, so not needed anymore. |
This PR implements the following changes:
DEV_NAMESPACE
to the namespace where the tenant is located).Fixes #2279.