After upgrading ArcGIS Enterprise on Kubernetes, a new backup store needs to be registered.

[cameronkroeker]

Backup and restores were successful with ArcGIS Enterprise on Kubernetes 11.2:

However, after upgrading to 11.3 the backup store is removed from Manager UI:

"We can't upgrade the Ozone cluster in-place since it wouldn't play nice with previous versions, the alternative we put in place is forcing an admin to register a new store via the Manager UI and we documented it in the upgrade considerations."

I re-ran the enterprise-k8s-aws-backup workflow, which completed successfully. The backup is in the bucket, and even shows up in the admin api:

It does not however show up in the Manager UI. I tried changing this value thinking it would register a new backup store but it failed.

https://github.com/ArcGIS/arcgis-gitops/blob/cf779c2f7acdbf86e27b9d25a9fa6e53744e7853/config/aws/arcgis-enterprise-k8s/backup.vars.json#L8

[cameronkroeker]

We might need to check versionCreated and compare it with current version, if they dont match then register a new backup store

https://ckroeker-aws-arcgis-gitops-test.dev.geocloud.com/arcgis/admin/system/disasterrecovery/stores/s3-backup-store?f=pjson

{
    "bucketName": "arcgis-enterprise-backup-028606780797-ap-east-1",
    "default": false,
    "rootFolderName": "arcgis-enterprise-k8s",
    "providerID": "tvkkzu2gj1",
    "name": "s3-backup-store",
    "versionCreated": "11.2.0",
    "type": "S3",
    "region": "ap-east-1",
    "serviceID": "lboe7lsfub"
}

[pbobov]

This fix:

1. Makes enterprise-k8s-aws-organization workflow to register new S3 backup stores on upgrades.
2. Adds version number to backup store root directories.
3. Makes enterprise-k8s-aws-backup and enterprise-k8s-aws-restore workflows use default backup store if "store" input variable is null.
4. Adds "The admin_password value must contain at least one special character." requirement validation in enterprise-k8s-aws-organization workflow.