How to deploy

Requirements

• Kubernetes

• Helm 3+

• AWS CLI for getting permissions to download Helm chart

Deployment Requirements

• To deploy ODM StorageClass and IngressClass have to be configured in Kubernetes Cluster.

• To deploy an ODM-OPS helm chart, you must have administrator rights in the Kubernetes cluster.

• ODM-OPS must be installed before starting ODM updates every time.

• ODM-OPS must be 1 per Kubernetes cluster.

• To deploy ODM, you need the following:

  • Administrator rights in the ODM namespace

  • Permissions to list all Custom Resource Definitions (CRDs)

  • In addition to standard Kubernetes resources, the following custom resources will be used:

    • clickhouseinstallations API group clickhouse.altinity.com

    • clickhouseinstallationtemplates API group clickhouse.altinity.com

    • clickhousekeeperinstallations API group clickhouse-keeper.altinity.com

    • clickhouseoperatorconfigurations API group clickhouse.altinity.com

Deployment Information

• All main ODM settings are presented in the Deployment Helm Chart.

• To enable automatic metrics collection when using Prometheus Operator, configure the serviceMonitor parameter in the Helm Chart.

• We provide a tool that automates the process of updating credentials for ECR, where our container images are hosted. This tool ensures seamless access to ECR by handling credential rotation.

  • It's important to note that the ECR password is valid for a duration of 12 hours.

  • If you are using AWS EKS, you do not need this tool. Access to our ECR will be provided upon request.

• To utilize external storage solutions that are compatible with Distributed File System (DFS) within a Kubernetes environment, it is necessary to first establish a Persistent Volume Claim (PVC).

Deployment process

1. Preparing to log in to ECR.

   • If you are not hosting on AWS, Configure AWS CLI (see Long-term credentials) with the credentials provided by our support team (this is necessary to access the ECR).

   • If you are hosting on AWS, ensure that you have the AmazonEC2ContainerRegistryReadOnly policy or a similar one enabled for access to ECR.

2. Login to ECR

   aws ecr get-login-password --region us-east-1 | helm registry login --username AWS \
   --password-stdin 091468197733.dkr.ecr.us-east-1.amazonaws.com
   

3. Deploy ODM-OPS

   a) Download helm-chart:

   helm pull oci://091468197733.dkr.ecr.us-east-1.amazonaws.com/genestack/chart/odm-ops
   

   b) Untar the archive:

   tar xvf odm-ops-*.tgz
   

   c) Run ODM-OPS deployment, excluding crds directory:

   helm upgrade -i odm-ops --create-namespace --skip-crds -n odm-ops odm-ops
   

4. Deploy ODM

   a) Download helm-chart:

   helm pull oci://091468197733.dkr.ecr.us-east-1.amazonaws.com/genestack/chart/odm
   

   b) Untar the archive:

   tar xvf odm-*.tgz
   

   c) Take a look to a default values.yaml and to a odm/examples directory, the most general file in there is recomendations.yaml. Please, make changes according to your infrastructure and requirements in separate file, e.g. custom-values.yaml.

   cat odm/examples/*.yaml
   

   d) Run ODM deployment:

   helm upgrade -i odm --create-namespace -n odm -f custom-values.yaml odm