[Enhancement Tracking]: Improve & Fix some of the Ksctl internal code 🛣️ #334
Comments
dipankardas011
changed the title
dipankardas011
changed the title
|
Caution I am force all the context.TODO() and context.Background() already present to use the package specific context cc @Horiodino |
dipankardas011
changed the title
dipankardas011
changed the title
|
Improve the loading screen like some go routine that we are waiting for some information or wait operation package main
import (
"fmt"
"time"
"github.com/briandowns/spinner"
)
func main() {
for i := 0; i <= 43; i++ {
fmt.Println("i:", i)
do(i)
}
}
func do(i int) {
s := spinner.New(
spinner.CharSets[i],
100*time.Millisecond,
spinner.WithSuffix(" Waiting for Data to arrive"),
spinner.WithFinalMSG("Done"),
spinner.WithColor("green"))
s.Start()
time.Sleep(2 * time.Second)
s.Stop()
} |
dipankardas011
added
kind/feature
kind/bug
Some discussion on ksctl component release cycleFor example a new tag is created and a relase for ksctl core Sol 1in the package formed it will use this v1.2.0 version of the image only Sol 2if the ksctl apps manifests (deployment, services, pods, ...) are there in a specific loc then we don't need to wait for them to be func installApplication(struct{ Name: "ksctl-stateimport", Version: "v1.1.4" })then we don't need to maintain these manifest specific to version as it will be in speicif clocation |
|
we would also need to think on for each release there has to be one file or something which will contain the latest stable release version for these components and in predetermined location like the above mentioned so proposal is and then we can provide tags to reach in specific branch Warning constraint: all the release must have this directory |
|
we can learn from this |
Improved the generalLogger |
|
also try to make the implementations for mocks in the test cases only aka inside |
|
Need to investagate on this error
|
dipankardas011
changed the title
Enhancement Title
Remove Unused components and fix some of them
Summary of the Enhancement
os.Getenv("ksctl_fake_...")rather use some better methods #363)ssh executor is failing with something that is weird(Update and check the state document is completely utilized aka no information is missed. add a array of workernode VMSize along with the count of workernodes if we are allowing dynamic size of workernodes to be able to add (feature: add nodeSizes to the ksctl-state #354) ([ACTION-TRACKING] store instance type data for both managed and self-managed #246)
example
we can do is stop using the storageImporter and instead run a temporary pod which hosts a ksctl storage server and when we want it we deploy it and tell it to use the storageDriver.Kubernetes to import the state by this we can use the suggested method from the above PR to use kubectl api server sirect call to send the data in raw json. once done we can delete the pod after that only we will deploy the ksctl agent. by this we can elemiate the http server and we can stick to grpc and we can eliminate the over complication of storageImporter controller and crds... (enhancement: replace storage.Import controller with kubectl proxy possible http temp server #343)
If possible try to use context.Context (aka a better use of them) (feature: context-aware logging #342)
Handle panics (feature: context-aware logging #342)
SSH client failing (quite unique issue)
(enhancement: replace storage.Import controller with kubectl proxy possible http temp server #343)
Ha cluster cleanup fails , make the noOfCP, noOfWp, setter functions to invoke just after state is init is done so that we can be saved from data loss even if the graceful termination is added as the DeleVM needs them to be there for us to work in deletion otherwise it will fail (refactor: Resolving TODOs and UnExport of functions and variable which are unnecessary #351)
Security vulnerablity detected on mysql database when stateimport http server recives the exposrted data
make sure if all the newInterface assignment functions are returning concerete type as return type and let the user handle the var <var> interface = ??? (feature: refactored all the interface return to concerete types #350)
add context based cli logging (exploit the context) https://github.com/vito/progrockfailed to install istio in macos (helm repo update is the error message which is needed) (bug-fix: failing helm client in macos #355)
Motivation and Goals
It will help enable to remove unwanted codebases
Detailed Design
Note
Document the changes and changelog as well
Alternatives
No response
Scope of changes
Agreement
The text was updated successfully, but these errors were encountered: