feat: Add InnoDB Cluster topology

[renecannao]

Summary

Add --topology=innodb-cluster to deploy a full InnoDB Cluster stack: Group Replication + MySQL Shell + MySQL Router.

Motivation

InnoDB Cluster is the standard MySQL high-availability solution. dbdeployer already supports group replication as a topology, but doesn't orchestrate the full InnoDB Cluster stack (MySQL Shell for cluster management, MySQL Router for connection routing).

Requirements

• New topology: dbdeployer deploy replication <version> --topology=innodb-cluster
• MySQL Shell integration for cluster bootstrap (mysqlsh creates the cluster)
• --mysqlsh-path flag to specify MySQL Shell binary location
• Dedicated templates for InnoDB Cluster initialization (8.4+ compatible)
• check_nodes_cluster monitoring script
• Requires MySQL 8.4+ (InnoDB Cluster is mature from 8.0 but 8.4 is the LTS target)

Implementation Notes

• New file: sandbox/innodb_cluster.go
• New templates: sandbox/templates/cluster/
• New topology type in globals/globals.go
• MySQL Shell needs to be available (either in PATH or via --mysqlsh-path)
• Should work with --with-proxysql (ProxySQL in front of InnoDB Cluster)

[coderabbitai]

🔗 Related PRs

#22 - Add MySQL integration tests (8.0, 8.4, 9.1) [merged]
#23 - Fix remaining templates for MySQL 8.4+ replication syntax [merged]

---

📝 Issue Planner

_{Check the box below or use the @coderabbitai plan command to generate an implementation plan and prompts that you can use with your favorite coding assistant.}

• Create Plan

---

🧪 Issue enrichment is currently in open beta.

You can configure auto-planning by selecting labels in the issue_enrichment configuration.

To disable automatic issue enrichment, add the following to your .coderabbit.yaml:

issue_enrichment:
  auto_enrich:
    enabled: false

💬 Have feedback or questions? Drop into our discord!

[renecannao]

Research Summary

Deployment Flow

1. Allocate ports: N MySQL + N GR communication + N MySQLX + 4 Router ports (R/W 6446, R/O 6447, X R/W 6448, X R/O 6449)
2. Create N MySQL node sandboxes (reuse existing GR CreateChildSandbox)
3. Start all nodes in parallel
4. Bootstrap cluster via MySQL Shell AdminAPI: dba.createCluster() + cluster.addInstance() per node
5. Bootstrap MySQL Router: mysqlrouter --bootstrap root@node1:port --directory <sandbox>/router
6. Start Router
7. Write topology scripts (start_all, stop_all, check_cluster, router_start/stop/status)

New Files

• sandbox/innodb_cluster.go — main CreateInnoDBCluster() function
• sandbox/innodb_cluster_templates.go — template embeds
• 7 template files in sandbox/templates/innodb_cluster/

Modified Files

• globals/globals.go — InnoDBClusterLabel constant
• globals/template_names.go — template name constants
• sandbox/replication.go — dispatch case
• defaults/defaults.go — prefix/port defaults
• cmd/replication.go — optional --skip-router flag

Dependencies

• MySQL Shell — already partially supported (detection exists at basedir/bin/mysqlsh or PATH)
• MySQL Router — NEW dependency, needs detection logic. Often bundled in MySQL 8.0+ tarballs.
• Optional --skip-router flag for ProxySQL-only testing

Estimated Complexity: Medium

~400-600 lines new Go + 7 templates. Key risk: cluster.addInstance() with clone recovery can be slow/flaky.