Upgrade to v1.5

This guide covers upgrading Orchestrator deployments from either v1.3 or v1.4 to v1.5.

Prerequisites

Terraform Version

Use Terraform version 0.14.5 for this Orc8r upgrade. Current Terraform scripts are not compatible with Terraform 0.15.0.

Helm Charts and Application Containers

Have all application containers, and Helm charts built and published. These should be from the head of the release branch. You can follow the Build Orchestrator guide for instructions on this process.

Terraform State

If you are using local Terraform state (the default), ensure all Terraform state files are located in your working directory before proceeding. As a check of this, terraform show should list existing state, rather than outputting No state.

Kubernetes Cluster Access

You should be able to access your Kubernetes cluster from the machine you are using to initiate the upgrade process through the kubectl CLI.

If you cannot access your Kubernetes cluster through kubectl, set your KUBECONFIG environment variable to point to your Orc8r cluster's kubeconfig file.

Kubernetes Cluster Version

Have the Kubernetes version running on your cluster ready. This version number will be used to set cluster_version later on during the upgrade process.

If you are using AWS, and the AWS CLI is configured locally, run

aws eks describe-cluster --name orc8r

and note the version listed under cluster.

Otherwise, navigate to the AWS UI. Enter service EKS and select clusters on the left hand side of the screen. Note the Kubernetes version for the orc8r cluster.

Upgrade Process

1. NMS DB Data Migration

This process prepares your cluster for Orc8r-NMS DB unification. (In release 1.5 and going forward, the default DB type is Postgres for both NMS and Orc8r.)

Paste in shell prompt:

wget https://raw.githubusercontent.com/magma/magma/v1.5/nms/app/packages/magmalte/scripts/fuji-upgrade/pre-upgrade-migration.sh && chmod +x pre-upgrade-migration.sh && ./pre-upgrade-migration.sh

The defaults options should likely work, unless the script cannot find the right pods, or you have your cluster namespaced differently.

2. Terraform main.tf Sanitization

The following variables should be removed from the main.tf Terraform file:

- nms_db_host
- nms_db_name
- nms_db_user
- nms_db_pass

If you are using a non-Postgres DB instance for your Orc8r setup, you will need to modify orc8r_db_dialect. This variable was added in 1.5, and defaults to postgres. Common dialects will be mysql, postgres, and mariadb. See the sequelize docs for a list.

3. Upgrade Terraform Modules

Set the source values for each of the Orchestrator modules in your root Terraform module to point to this release's modules and bump chart and container versions

# This will likely be found in main.tf

module orc8r {
  source = "github.com/magma/magma//orc8r/cloud/deploy/terraform/orc8r-aws?ref=v1.5"
  # ...
  cluster_version = "1.17"   # set to Kubernetes version found above. 1.17 is used as an example here
}

module orc8r-app {
  source = "github.com/magma/magma//orc8r/cloud/deploy/terraform/orc8r-helm-aws?ref=v1.5"
  # ...
  orc8r_chart_version   = "1.5.20"
  orc8r_tag             = "MAGMA_TAG"  # from build step, e.g. v1.5.0
  orc8r_deployment_type = "fwa"        # valid options: ["fwa", "federated_fwa", "all"]
}

Set cluster_version to the Kubernetes version found during the Prerequisites section. Bump your chart version to 1.5.20 and orc8r_tag to the semver tag you published your new Orchestrator container images as. You also need to set the orc8r_deployment_type variable to the deployment type that you intend to deploy. This type sets which Orc8r modules will run.

Then, prepare Terraform for the upgrade

terraform init --upgrade
terraform refresh

4. Terraform Apply

Terraform the upgrade.

Apply the upgrade

terraform apply     # Check this output VERY carefully!

After the Terraform command completes, all application components should be successfully upgraded.

5. (Optional) Post-Upgrade: Sync Pre-defined Alerts

A new Duplicate Request Alert has been added, which triggers when more than 100 alerts within the last 5 minutes have been raised.

To enable this alert, you must "Sync Pre-Defined Alerts" for every network defined in the Orc8r. See the NMS Alerts Guide for guidance on this process.

Additional Details

NMS DB Data Migration

Before upgrading Orc8r to 1.5, the NMS DB migration needs to be completed. This migration copies all NMS data to the same database housing Orc8r data.

These NMS tables are named

- Users
- Organizations
- FeatureFlags
- AuditLogEntry
- SequelizeMeta

A single script is provided to perform this migration for you, and should tell you whether or not it was successful. After migration, the tables listed above should be present in the Orc8r database.

The script will ask for various inputs, and the defaults can be chosen unless the script fails to find certain information, such as pod names. It uses kubectl internally to run commands inside the pods.

New Pre-defined Alerts

"Duplicate request alert" is raised as a critical alert, when we receive more than 100 alerts within last 5 minutes. If you would like to customize this alert, refer to NMS alerts guide - custom alert rules.

In order to start receiving these alarms, it is necessary to click on the “Sync Pre-Defined alerts” in the NMS window for every network defined in the Orc8r. For more details on this process, see the NMS alerts guide Syncing the alerts once on the NMS does not replicate the behavior to other networks managed by the same NMS.