https://flyte.org logo
#flyte-deployment
Title
# flyte-deployment
d

Dhruv Malik

11/29/2023, 2:16 PM
Hi team , i wanted to ask regarding the tutorial regarding how to deploy the single cluster inputs and i wanted to ask how i can define the parameters concerning the storage for the configuration ? thanks:
Copy code
configuration:
  storage:
    metadataContainer: <BUCKET_NAME>
    userDataContainer: <USER_DATA_BUCKET_NAME>
    provider: s3
    providerConfig:
      s3:
        region: "<AWS_REGION>"
        authType: "iam"
c

Chris Grass

11/29/2023, 2:36 PM
Hi Dhruv, welcome. I'm not sure I understand the question -
<BUCKET_NAME>
and the other vars are strings that can be defined in the values.yaml file, one of the overrides, or reference a predefined value in another helm chart. are you having an issue defining them?
d

Dhruv Malik

11/29/2023, 2:41 PM
thanks for response . its fine i defined the vars of the resources by creating the instances . i was just curious regarding whether USER_DATA_BUCKET_NAME and BUCKET_NAME corresponds to the diffferent S3 object storages or they can be set to unique . but then i am trying to run the command
Copy code
helm install flyte-backend flyteorg/flyte-binary --dry-run --namespace flyte --values eks_test.yaml
and its giving the issue :
Error: INSTALLATION FAILED: Kubernetes cluster unreachable: Get "<https://127.0.0.1:30086/version>": dial tcp 127.0.0.1:30086: connect: connection refused
sorry for the naive queries. i am new to the k8s and cloud native ecosystem and thanks a ton
c

Chris Grass

11/29/2023, 2:44 PM
i was just curious regarding whether USER_DATA_BUCKET_NAME and BUCKET_NAME corresponds to the diffferent S3 object storages or they can be set to unique .
I set them to the same bucket for simplicity but you should be able to make them unique if that is what your use case calls for.
sorry for the naive queries. i am new to the k8s and cloud native ecosystem and thanks a ton
no worries! i'm new to kubernetes and flyte as well
you are trying to run the local sandbox?
d

Dhruv Malik

11/29/2023, 2:47 PM
i am trying to setup the single dedicated cluster , so as to then refer the cluster using :
Copy code
flytectl config init --host={FLYTEADMIN_URL} --storage
c

Chris Grass

11/29/2023, 2:51 PM
what do you see when you run the
kubectl cluster-info
?
d

Dhruv Malik

11/29/2023, 2:53 PM
I got this :
Copy code
E1129 14:52:36.629697 1264538 memcache.go:265] couldn't get current server API group list: Get "<https://127.0.0.1:30086/api?timeout=32s>": dial tcp 127.0.0.1:30086: connect: connection refused
E1129 14:52:36.630035 1264538 memcache.go:265] couldn't get current server API group list: Get "<https://127.0.0.1:30086/api?timeout=32s>": dial tcp 127.0.0.1:30086: connect: connection refused
seems that its clashing with the port that is already running or issue of sudo ?
c

Chris Grass

11/29/2023, 2:54 PM
do you have kubectx installed?
d

Dhruv Malik

11/29/2023, 2:57 PM
ah well i didnt know kubectx (and was confusing that with ctl )
c

Chris Grass

11/29/2023, 2:58 PM
yeah, it's a confusing ecosystem to jump into because of how many tools are helpful (required?). kubectx is useful if you want to bounce around between different k8s clusters
or managing different contexts
did you initially try to setup the local sandbox? i'm wondering if maybe kubectl is still pointing to that context
d

Dhruv Malik

11/29/2023, 3:02 PM
yeah thats helpful (kind of similar with the nvm in the nodejs ) . indeed i tested the sandbox using
flytectl demo start
and then FLYTECTL_CONFIG set up with that yaml file
c

Chris Grass

11/29/2023, 3:02 PM
my guess is that you need to switch the context kubectl is using
kubectx is a relatively easy tool to help with that
if you would prefer to use
kubectl
directly, here is a link to updating the config
d

Dhruv Malik

11/29/2023, 3:10 PM
thanksa lot yeah i installed it and seems creating separate env's will be better to separate the concerns . that was quick learning curve for either 2 situations
Hey just the followup about the progress i had and indeed i wanted to request PR for having more detailed roadmap for those who are even new to the kubernetes actually even if we set the kubectl enviornment correct, still the error was prevelant . i.e
Error:
1 error(s) decoding:
* '' has invalid keys: console
strict mode is on but received keys [map[level:{}]] to decode with no config assigned to receive them: failed strict mode check
ERRO[0000]
1 error(s) decoding:
* '' has invalid keys: console
strict mode is on but received keys [map[level:{}]] to decode with no config assigned to receive them: failed strict mode check  src="main.go:13"
then after just going through the documentation and concepts regarding kubernetes , i saw about
kind
and its being the cli specifically for deploying the clusters specifically . ( i normally thought that helm is gonna do this naturally but maybe in my case ), , so finally after doing
kind create cluster --config eks_config_file.yaml
and then the helm command i was able to get the deployment chart . but i do appreciate the references you've shared that kind of get me to the learning curve in understanding kubernetes.
2 Views