Remove IBM Db2 support from RocketAdmin Agent

- Updated main.ts to remove Db2 from the list of supported database types.
- Deleted wait-for-db2.js and wait-for-db2.sh scripts as they are no longer needed.
- Removed optional dependency on ibm_db from package.json.
- Cleaned up caching constants and removed Db2 related cache options.
- Eliminated Db2 related code from caching and data access layers.
- Removed DataAccessObjectIbmDb2 and its references throughout the codebase.
- Updated connection parameters and enums to exclude Db2.

Author: Artuomka

AGENTS.md

@@ -0,0 +1,131 @@
+# CLAUDE.md
+
+This file provides guidance to Claude Code (claude.ai/code) when working with code in this repository.
+
+## Project Overview
+
+RocketAdmin is a database administration panel that allows users to manage database connections, tables, and data. It consists of multiple components in a monorepo structure:
+
+- **backend/** - NestJS API server (TypeScript, ES modules)
+- **frontend/** - Angular 19 web application (standalone components)
+- **rocketadmin-agent/** - NestJS agent for connecting to databases behind firewalls
+- **autoadmin-ws-server/** - WebSocket server for agent communication
+- **shared-code/** - Shared data access layer and utilities used by backend and agent
+
+## Development Commands
+
+### Backend
+
+```bash
+cd backend
+pnpm start:dev          # Start dev server with hot reload
+pnpm build              # Build for production
+pnpm lint               # ESLint with auto-fix
+pnpm test               # Run non-saas AVA tests (serial)
+pnpm test-all           # Run all AVA tests (5min timeout, serial)
+pnpm test-saas          # Run SaaS-specific tests
+```
+
+### Frontend
+
+```bash
+cd frontend
+yarn start              # Start Angular dev server
+yarn build              # Production build
+yarn test:ci            # Run tests headlessly (CI mode)
+yarn test --browsers=ChromeHeadlessCustom --no-watch --no-progress  # Headless tests
+yarn lint               # TSLint (deprecated, needs ESLint migration)
+```
+
+### Running Backend Tests with Docker
+
+The project uses `just` for test orchestration:
+
+```bash
+just test               # Run all backend tests with Docker Compose
+just test "path/to/test.ts"  # Run specific test file
+```
+
+This spins up test databases (MySQL, PostgreSQL, MSSQL, Oracle, MongoDB, DynamoDB) via `docker-compose.tst.yml`.
+
+### Migrations
+
+```bash
+cd backend
+pnpm build              # Must build first
+pnpm migration:generate src/migrations/MigrationName  # Generate migration
+pnpm migration:run      # Run pending migrations
+pnpm migration:revert   # Revert last migration
+```
+
+## Architecture
+
+### Monorepo Structure
+
+- Uses pnpm workspaces with packages: `backend`, `rocketadmin-agent`, `shared-code`
+- `shared-code` is imported as `@rocketadmin/shared-code` workspace dependency
+- Frontend is a separate Angular project (not a workspace member)
+
+### Backend (NestJS)
+
+- **Entities pattern**: Each entity has its own directory under `src/entities/` containing:
+  - `*.entity.ts` - TypeORM entity
+  - `*.module.ts` - NestJS module
+  - `*.controller.ts` - REST endpoints
+  - `*.service.ts` - Business logic (use cases)
+  - `dto/` - Request/response DTOs with class-validator decorators
+  - `*.controller.ee.ts` - Enterprise edition controllers (SaaS features)
+- **Guards**: Authentication and authorization in `src/guards/`
+- **Data access**: Uses `shared-code` for database operations via Knex
+- **Testing**: AVA test framework with tests in `test/ava-tests/`
+  - `non-saas-tests/` - Core functionality tests
+  - `saas-tests/` - SaaS-specific feature tests
+  - `complex-table-tests/` - Complex table operation tests
+
+### Frontend (Angular 19)
+
+See `frontend/CLAUDE.md` for detailed frontend architecture.
+
+Key points:
+- Standalone components (no NgModules)
+- BehaviorSubject-based state management (no NgRx)
+- Multi-environment builds (development, production, saas, saas-production)
+- Jasmine/Karma testing with ChromeHeadless
+
+### Shared Code
+
+Located in `shared-code/src/`:
+- `data-access-layer/` - Database abstraction supporting MySQL, PostgreSQL, MSSQL, Oracle, MongoDB, DynamoDB, Cassandra, Elasticsearch
+- `knex-manager/` - Knex connection management
+- `caching/` - LRU cache utilities
+- `helpers/` - Shared utilities
+
+### Agent Architecture
+
+The rocketadmin-agent connects to databases in private networks:
+1. Agent runs inside customer's network
+2. Connects to `autoadmin-ws-server` via WebSocket
+3. Backend communicates with agent through WebSocket server
+4. Agent executes database queries and returns results
+
+## Database Support
+
+The application supports: MySQL, PostgreSQL, MongoDB, DynamoDB, Cassandra, OracleDB, MSSQL, Elasticsearch, Redis
+
+Database-specific DAOs are in `shared-code/src/data-access-layer/`.
+
+## Testing Database Connections
+
+Test databases are defined in `docker-compose.tst.yml`:
+- MySQL: `testMySQL-e2e-testing:3306`
+- PostgreSQL: `testPg-e2e-testing:5432`
+- MSSQL: `mssql-e2e-testing:1433`
+- Oracle: `test-oracle-e2e-testing:1521`
+- MongoDB: `test-mongo-e2e-testing:27017`
+- DynamoDB: `test-dynamodb-e2e-testing:8000`
+
+## Coding Conventions
+
+### Class Member Ordering
+
+- Private methods must be placed at the end of the class, after all public methods

AGENTS.md

@@ -46,7 +46,7 @@ just test               # Run all backend tests with Docker Compose
 just test "path/to/test.ts"  # Run specific test file
 ```
 
-This spins up test databases (MySQL, PostgreSQL, MSSQL, Oracle, IBM DB2, MongoDB, DynamoDB) via `docker-compose.tst.yml`.
+This spins up test databases (MySQL, PostgreSQL, MSSQL, Oracle, MongoDB, DynamoDB) via `docker-compose.tst.yml`.
 
 ### Migrations
 
@@ -95,7 +95,7 @@ Key points:
 ### Shared Code
 
 Located in `shared-code/src/`:
-- `data-access-layer/` - Database abstraction supporting MySQL, PostgreSQL, MSSQL, Oracle, MongoDB, DynamoDB, IBM DB2, Cassandra, Elasticsearch
+- `data-access-layer/` - Database abstraction supporting MySQL, PostgreSQL, MSSQL, Oracle, MongoDB, DynamoDB, Cassandra, Elasticsearch
 - `knex-manager/` - Knex connection management
 - `caching/` - LRU cache utilities
 - `helpers/` - Shared utilities
@@ -110,7 +110,7 @@ The rocketadmin-agent connects to databases in private networks:
 
 ## Database Support
 
-The application supports: MySQL, PostgreSQL, MongoDB, DynamoDB, Cassandra, OracleDB, MSSQL, IBM DB2, Elasticsearch, Redis
+The application supports: MySQL, PostgreSQL, MongoDB, DynamoDB, Cassandra, OracleDB, MSSQL, Elasticsearch, Redis
 
 Database-specific DAOs are in `shared-code/src/data-access-layer/`.
 
@@ -121,7 +121,6 @@ Test databases are defined in `docker-compose.tst.yml`:
 - PostgreSQL: `testPg-e2e-testing:5432`
 - MSSQL: `mssql-e2e-testing:1433`
 - Oracle: `test-oracle-e2e-testing:1521`
-- IBM DB2: `test-ibm-db2-e2e-testing:50000`
 - MongoDB: `test-mongo-e2e-testing:27017`
 - DynamoDB: `test-dynamodb-e2e-testing:8000`
 

CLAUDE.md

@@ -19,7 +19,7 @@
 		"test-all": "node _run-with-timing.mjs ava --timeout=5m",
 		"test-all-parallel": "AVA_CONCURRENCY=8 node _run-with-timing.mjs ava --timeout=5m",
 		"test-saas": "node _run-with-timing.mjs ava test/ava-tests/saas-tests/*",
-		"test-fast": "AVA_CONCURRENCY=6 node _run-with-timing.mjs ava --timeout=5m 'test/ava-tests/non-saas-tests/!(*oracle*|*ibmdb2*|*cassandra*|*elasticsearch*).test.ts' 'test/ava-tests/saas-tests/!(*oracle*|*ibmdb2*|*cassandra*|*elasticsearch*).test.ts'",
+		"test-fast": "AVA_CONCURRENCY=6 node _run-with-timing.mjs ava --timeout=5m 'test/ava-tests/non-saas-tests/!(*oracle*|*cassandra*|*elasticsearch*).test.ts' 'test/ava-tests/saas-tests/!(*oracle*|*cassandra*|*elasticsearch*).test.ts'",
 		"typeorm": "ts-node -r tsconfig-paths/register ../node_modules/.bin/typeorm",
 		"migration:generate": "pnpm run typeorm migration:generate -d dist/src/shared/config/datasource.config.js",
 		"migration:create": "pnpm run typeorm migration:create -d dist/src/shared/config/datasource.config.js",
@@ -109,7 +109,6 @@
 		"@types/body-parser": "^1.19.6",
 		"@types/cookie-parser": "^1.4.10",
 		"@types/express": "^5.0.6",
-		"@types/ibm_db": "^3.2.0",
 		"@types/json2csv": "^5.0.7",
 		"@types/node": "^24.10.1",
 		"@types/supertest": "^7.2.0",

backend/package.json

@@ -59,9 +59,6 @@ export function convertDbTypeToReadableString(dataType: ConnectionTypesEnum): st
 		case ConnectionTypesEnum.oracledb:
 		case ConnectionTypesEnum.agent_oracledb:
 			return 'Oracle DB';
-		case ConnectionTypesEnum.ibmdb2:
-		case ConnectionTypesEnum.agent_ibmdb2:
-			return 'IBM DB2';
 		case ConnectionTypesEnum.clickhouse:
 		case ConnectionTypesEnum.agent_clickhouse:
 			return 'ClickHouse';

backend/src/ai-core/tools/prompts.ts

@@ -103,9 +103,6 @@ export function wrapQueryWithLimit(query: string, databaseType: ConnectionTypesE
 		case ConnectionTypesEnum.mssql:
 		case ConnectionTypesEnum.agent_mssql:
 			return `SELECT * FROM (${queryWithoutSemicolon}) AS ai_query LIMIT ${limit}`;
-		case ConnectionTypesEnum.ibmdb2:
-		case ConnectionTypesEnum.agent_ibmdb2:
-			return `SELECT * FROM (${queryWithoutSemicolon}) AS ai_query FETCH FIRST ${limit} ROWS ONLY`;
 		case ConnectionTypesEnum.oracledb:
 		case ConnectionTypesEnum.agent_oracledb:
 			return `SELECT * FROM (${queryWithoutSemicolon}) WHERE ROWNUM <= ${limit}`;

backend/src/ai-core/tools/query-validators.ts

@@ -52,8 +52,6 @@ export const customAgentRepositoryExtension: IAgentRepository = {
 				return 'MYSQL-TEST-AGENT-TOKEN';
 			case ConnectionTypeTestEnum.agent_postgres:
 				return 'POSTGRES-TEST-AGENT-TOKEN';
-			case ConnectionTypeTestEnum.agent_ibmdb2:
-				return 'IBMDB2-TEST-AGENT-TOKEN';
 			case ConnectionTypeTestEnum.agent_mongodb:
 				return 'MONGODB-TEST-AGENT-TOKEN';
 			case ConnectionTypeTestEnum.agent_redis:

backend/src/entities/agent/repository/custom-agent-repository-extension.ts

@@ -6,13 +6,11 @@ const SQL_CONNECTION_TYPES: ReadonlySet<string> = new Set<string>([
 	ConnectionTypesEnum.mysql2,
 	ConnectionTypesEnum.oracledb,
 	ConnectionTypesEnum.mssql,
-	ConnectionTypesEnum.ibmdb2,
 	ConnectionTypesEnum.clickhouse,
 	ConnectionTypesEnum.agent_postgres,
 	ConnectionTypesEnum.agent_mysql,
 	ConnectionTypesEnum.agent_oracledb,
 	ConnectionTypesEnum.agent_mssql,
-	ConnectionTypesEnum.agent_ibmdb2,
 	ConnectionTypesEnum.agent_clickhouse,
 ]);
 

backend/src/entities/connection/utils/is-sql-connection-type.util.ts

@@ -46,10 +46,9 @@ Multi-proposal rules:
 - For a single-change request, supply a "proposals" array of length 1 — same content as before.
 
 Rules for the generated SQL:
-- Target dialect is ${dialect}. Use the correct identifier quoting (double quotes for PostgreSQL/Oracle/DB2, backticks for MySQL/ClickHouse, square brackets or double quotes for Microsoft SQL Server) and the correct syntax for data types, autoincrement, and constraints.
+- Target dialect is ${dialect}. Use the correct identifier quoting (double quotes for PostgreSQL/Oracle, backticks for MySQL/ClickHouse, square brackets or double quotes for Microsoft SQL Server) and the correct syntax for data types, autoincrement, and constraints.
 - For Microsoft SQL Server, use IDENTITY for autoincrement, NVARCHAR/VARCHAR for strings, and name primary/foreign keys explicitly (e.g. CONSTRAINT PK_tbl PRIMARY KEY ...) so they are referenceable in rollback DROP CONSTRAINT statements.
 - For Oracle DB, use NUMBER/VARCHAR2, and GENERATED BY DEFAULT AS IDENTITY for autoincrement. ALTER TABLE ... MODIFY is the column-change syntax.
-- For IBM DB2, use BIGINT GENERATED BY DEFAULT AS IDENTITY and VARCHAR. ALTER TABLE ... ALTER COLUMN ... SET DATA TYPE is the column-change syntax.
 - For ClickHouse: every CREATE TABLE MUST include a table engine (prefer \`ENGINE = MergeTree()\`) followed by an \`ORDER BY (<column or tuple>)\` clause — ClickHouse has no conventional PRIMARY KEY; the sort/primary key is the \`ORDER BY\` tuple. Use types like \`UInt32\`, \`UInt64\`, \`Int64\`, \`String\`, \`Float64\`, \`DateTime\`, \`Date\`, \`UUID\`, \`Nullable(T)\` (wrap a type to allow NULL). There is no autoincrement — use a plain numeric type the user populates. Column additions use \`ALTER TABLE t ADD COLUMN c T\`, drops use \`ALTER TABLE t DROP COLUMN c\`, type changes use \`ALTER TABLE t MODIFY COLUMN c T\`. Do NOT emit \`ON CLUSTER\` clauses; the target is a single-node server. There are NO true transactions, so rollback is a best-effort compensating DDL (e.g. add-column forward / drop-column rollback). Do NOT propose foreign keys (ClickHouse does not enforce them). Indexes in ClickHouse are DATA SKIPPING indexes created with \`ALTER TABLE t ADD INDEX idx_name col TYPE minmax GRANULARITY 4\`; rollback is \`ALTER TABLE t DROP INDEX idx_name\`. Avoid \`DROP TABLE IF EXISTS\` unless the user asked.
 - For Cassandra (CQL): every CREATE TABLE MUST declare a \`PRIMARY KEY\` inline, either as a column-level \`PRIMARY KEY\` on one column or as a trailing \`PRIMARY KEY ((partition_key_cols), clustering_key_cols)\` clause. Use CQL types: \`UUID\`, \`TIMEUUID\`, \`TEXT\`, \`VARCHAR\`, \`ASCII\`, \`INT\`, \`BIGINT\`, \`SMALLINT\`, \`TINYINT\`, \`FLOAT\`, \`DOUBLE\`, \`DECIMAL\`, \`BOOLEAN\`, \`TIMESTAMP\`, \`DATE\`, \`TIME\`, \`BLOB\`, \`INET\`, \`LIST<T>\`, \`SET<T>\`, \`MAP<K,V>\`. There is NO autoincrement — prefer \`UUID\` partition keys. Do NOT propose foreign keys (Cassandra does not enforce them). Do NOT propose \`ALTER COLUMN\` type-change DDL — CQL only supports renaming primary-key columns and adding/dropping non-primary-key columns. Column additions use \`ALTER TABLE t ADD c T\` (no \`COLUMN\` keyword), drops use \`ALTER TABLE t DROP c\`. Indexes are \`CREATE INDEX idx_name ON t (col)\` with rollback \`DROP INDEX idx_name\`. There are NO transactions; rollback is a best-effort compensating DDL. Do NOT emit \`CREATE KEYSPACE\`, \`DROP KEYSPACE\`, \`USE\`, or materialized-view DDL.
 - Both forwardSql and rollbackSql MUST be single DDL statements. No semicolons terminating a chain. No multi-statement scripts.

backend/src/entities/table-schema/ai/schema-change-prompts.ts

@@ -7,7 +7,6 @@ const SUPPORTED_DIALECTS: ReadonlySet<ConnectionTypesEnum> = new Set([
 	ConnectionTypesEnum.mysql2,
 	ConnectionTypesEnum.mssql,
 	ConnectionTypesEnum.oracledb,
-	ConnectionTypesEnum.ibmdb2,
 	ConnectionTypesEnum.mongodb,
 	ConnectionTypesEnum.clickhouse,
 	ConnectionTypesEnum.agent_clickhouse,
@@ -24,7 +23,7 @@ export function isDialectSupported(connectionType: ConnectionTypesEnum): boolean
 export function assertDialectSupported(connectionType: ConnectionTypesEnum): void {
 	if (!isDialectSupported(connectionType)) {
 		throw new BadRequestException(
-			`Schema changes via AI are not yet supported for "${connectionType}". Supported: PostgreSQL, MySQL, Microsoft SQL Server, Oracle DB, IBM DB2, MongoDB, ClickHouse, DynamoDB, Cassandra, Elasticsearch.`,
+			`Schema changes via AI are not yet supported for "${connectionType}". Supported: PostgreSQL, MySQL, Microsoft SQL Server, Oracle DB, MongoDB, ClickHouse, DynamoDB, Cassandra, Elasticsearch.`,
 		);
 	}
 }
@@ -57,8 +56,6 @@ const SQL_PARSER_DIALECTS: Record<string, string> = {
 	[ConnectionTypesEnum.agent_mysql]: 'MySQL',
 	[ConnectionTypesEnum.mssql]: 'TransactSQL',
 	[ConnectionTypesEnum.agent_mssql]: 'TransactSQL',
-	[ConnectionTypesEnum.ibmdb2]: 'DB2',
-	[ConnectionTypesEnum.agent_ibmdb2]: 'DB2',
 };
 
 export function connectionTypeToParserDialect(connectionType: ConnectionTypesEnum): string {