Add multi-storage support with managed upload workflow

[mgajek-cern]

Summary

Extend k8s-tutorial to demonstrate Rucio data management across multiple storage technologies (S3, dCache) using the managed upload pattern.

Motivation

Current tutorial only demonstrates XRootD storage. Production Rucio deployments typically integrate with diverse storage backends (object stores, grid storage, distributed filesystems).

Proposed Changes

• Add S3/MinIO storage (2 instances) - cloud-native object storage
• Add dCache storage (2 instances) - grid storage with WebDAV
• Update workflow script demonstrating managed upload pattern
• Update deployment scripts to handle all storage types preserving previous behavior and opt-in feature flags for multi-storage demonstration

Workflow demonstrated:

1. Upload files to source RSEs (XRD1, S3_1, DCACHE_1)
2. Create datasets and containers
3. Create replication rules to destination RSEs
4. FTS3 orchestrates third-party copy transfers
5. Monitor transfer completion

Testing

./scripts/deploy-rucio.sh --enable-all-storage

# Auto-detects and configures enabled storages
./scripts/use-rucio.sh

Future Work

• Manual registration workflow (separate issue/branch for pre-existing data sources), e.g. see how PandaWMS handles this https://github.com/PanDAWMS/panda-server/blob/118b89651b6fbafeebd3bacf368cfb313142b991/pandaserver/dataservice/ddm.py#L17
• Cross-technology transfers using agreed upon protocol
• Performance benchmarking