Skip to main content
Glama
CDataSoftware

CData Sync MCP Server

Official

Server Configuration

Describes the environment variables required to run the server.

NameRequiredDescriptionDefault
NODE_ENVNoNode environmentproduction
LOG_LEVELNoLogging levelinfo
DISABLE_SSENoSet to true to disable Server-Sent Events (recommended for stdio/Claude Desktop transport)
MCP_HTTP_PATHNoHTTP transport base path/mcp/v1
MCP_HTTP_PORTNoHTTP transport port3000
CDATA_BASE_URLNoCData Sync API base URLhttp://localhost:8181/api.rsc
CDATA_PASSWORDNoBasic auth password (alternative to token)
CDATA_USERNAMENoBasic auth username (alternative to token)
CDATA_WORKSPACENoWorkspace UUID to scope all operations (optional)
CDATA_AUTH_TOKENNoAPI authentication token
MCP_TRANSPORT_MODENoTransport mode: stdio, http, or bothstdio

Capabilities

Features and capabilities supported by this server

CapabilityDetails
tools
{}
resources
{}

Tools

Functions exposed to the LLM to take actions

NameDescription
configure_sync_server

Configure the MCP server's connection to CData Sync. If not authenticated, you will be prompted for credentials. Use 'get' to view current configuration including auth type, connection status, and workspace context, or 'update' to modify connection settings including the default workspace. CRITICAL: Configuration changes take effect immediately and will reconnect all services. Test new credentials before applying. Cannot be used to view passwords or auth tokens for security reasons. The workspace setting determines which workspace context is used for all operations unless overridden in individual tool calls.

read_connections

Access data source/destination connections that define how CData Sync connects to databases, APIs, and files. If not authenticated with CData Sync, you will be prompted for credentials. Use 'list' to see all available connections, 'get' to retrieve details about a specific connection, 'test' to verify credentials and connectivity. Connections must be created and tested before being used in jobs. Multiple jobs can use the same connection simultaneously. Note: For counting connections, use 'list' and count the results client-side.

write_connections

Create, update, or delete data connections. If not authenticated with CData Sync, you will be prompted for credentials. Connections define how to access your data sources (databases, APIs, files) and destinations. Connection strings are provider-specific - consult CData documentation for your provider. Cannot modify or delete connections currently in use by running jobs. To change providers, delete and recreate the connection.

read_jobs

Access and monitor data replication jobs that move data from source to destination.

RETURNS:

  • list: Array of job objects with name, status, and configuration

  • get: Full job configuration details

  • status: Current execution state and progress

  • history: Past execution records

  • logs: Detailed execution logs for debugging

COMMON ERRORS:

  • "Job not found" - Verify job name/ID

  • "No logs available" - Check log retention settings

  • "Access denied" - Verify user permissions

write_jobs

Create, modify, or delete data replication jobs. If not authenticated with CData Sync, you will be prompted for credentials. Jobs orchestrate moving data from source to destination with options for scheduling, transformations, error handling, and notifications. IMPORTANT: Use table names exactly as reported by the source connection - do not modify names or extensions. Job types: Standard=Basic replication, ReplicateAll=Sync all tables, LoadFolder=File-based, ChangeDataCapture=CDC, ReverseETL=Reverse sync, EnhancedChangeDataCapture=Enhanced CDC. Cannot modify/delete running jobs. Add tasks after creating the job.

execute_job

Run a job immediately, bypassing its schedule. If not authenticated with CData Sync, you will be prompted for credentials. Executes all tasks in sequence (or parallel if configured). Use waitForResults=true to wait for completion and see results, or false to start asynchronously. Running jobs cannot be modified or deleted until complete.

cancel_job

Stop a currently running job.

RETURNS:

  • Success: Confirmation of cancellation request

  • Error: { code: -32603, message: "error details" }

COMMON ERRORS:

  • "Job not running" - Job may have already completed

  • "Job not found" - Verify job name/ID

  • "Cancellation failed" - Job may be in uncancellable state

read_tasks

Access tasks within a specific job. Each task defines specific data to replicate.

REQUIRED: jobName parameter for all actions.

RETURNS:

  • get: Array of task objects with TaskId, Query, Table, and Index

COMMON ERRORS:

  • "Job not found" - Verify job name with read_jobs

  • "No tasks defined" - Job may be empty, add tasks with write_tasks

write_tasks

Manage tasks within jobs. Tasks define what data to replicate.

IMPORTANT: Use exact table names from get_connection_tables, including file extensions.

RETURNS:

  • create: New task with assigned TaskId

  • update: Creates new task (cannot modify existing)

  • delete: Confirmation of task removal

COMMON ERRORS:

  • "Invalid table name" - Use exact name from get_connection_tables

  • "Task not found" - Verify TaskId with read_tasks

  • "Invalid query syntax" - Check SQL syntax for provider

execute_query

Execute pre-defined queries within a job context for testing or ad-hoc operations.

IMPORTANT: Can only execute queries already defined as tasks - cannot run arbitrary SQL.

RETURNS:

  • Success: Query execution results with row counts and timing

  • Async: Confirmation that queries started

  • Error: { code: -32603, message: "error details" }

COMMON ERRORS:

  • "Query not found" - Query must exist as task in job

  • "Job not found" - Verify job name/ID

  • "Execution failed" - Check query syntax and connections

get_connection_tables

Discover available tables/views in a data source. Essential before creating jobs.

RETURNS: Array of table names in exact format needed for tasks.

COMMON ERRORS:

  • "Connection not found" - Verify connection name

  • "Access denied" - Check connection permissions

  • "No tables found" - Verify schema/catalog settings

get_table_columns

Get column details for a specific table including names, data types, and keys.

RETURNS: Array of column objects with name, type, nullable, and key information.

COMMON ERRORS:

  • "Table not found" - Use exact name from get_connection_tables

  • "Connection not found" - Verify connection name

  • "Access denied" - Check table permissions

get_job_tables

List tables available to add to a specific job, considering job configuration and connection capabilities. If not authenticated with CData Sync, you will be prompted for credentials. Similar to get_connection_tables but filtered for job compatibility. Use when expanding job scope or adding new tables to existing job. Returns exact table names to use in tasks.

read_users

Access CData Sync user accounts and their permissions.

ROLES:

  • cdata_admin: Full administrative access

  • cdata_standard: Run existing jobs

  • cdata_job_creator: Create and modify jobs

  • cdata_support: Operate jobs (start/stop)

RETURNS:

  • list: Array of user objects

  • get: Detailed user information

  • count: Total number of users

COMMON ERRORS:

  • "User not found" - Check username spelling

  • "Access denied" - Admin role required

write_users

Create or update CData Sync users.

PASSWORD REQUIREMENTS: Must contain uppercase, lowercase, numbers, AND special characters.

RETURNS:

  • create: New user details with confirmation

  • update: Updated user information

COMMON ERRORS:

  • "Invalid password" - Must meet complexity requirements

  • "User already exists" - Use unique usernames

  • "Invalid role" - Use exact role values (cdata_admin, etc.)

  • "Bulk creation failed" - Check individual user errors

read_history

Access job execution history to analyze performance, troubleshoot failures, and audit data movements. If not authenticated with CData Sync, you will be prompted for credentials. Each history record shows when a job ran, its status, duration, and records affected. Use 'list' to browse history with filters/sorting. Note: Count action not supported by API - use 'list' and count results client-side. Essential for monitoring job health and SLA compliance.

read_requests

Access API request logs for auditing, debugging, and compliance.

RETURNS:

  • list: Array of request log entries

  • count: Total number of requests

  • get: Detailed request information

COMMON ERRORS:

  • "Request not found" - Check request ID

  • "Logs purged" - Older logs may be deleted per retention policy

write_requests

Delete API request log entries for privacy compliance or cleanup.

RETURNS: Confirmation of log entry deletion.

COMMON ERRORS:

  • "Request not found" - Verify request ID

  • "Access denied" - Admin permissions required

  • "Cannot delete" - Some logs may be protected

read_transformations

Access data transformations that run SQL in the destination (ELT pattern).

RETURNS:

  • list: Array of transformation objects

  • get: Detailed transformation configuration

  • count: Total number of transformations

COMMON ERRORS:

  • "Transformation not found" - Check name spelling

  • "Access denied" - Verify permissions

write_transformations

Create, update, or delete SQL transformations for ELT processing.

RETURNS:

  • create: New transformation details

  • update: Updated configuration

  • delete: Confirmation of removal

COMMON ERRORS:

  • "Invalid SQL syntax" - Check destination SQL dialect

  • "Connection not found" - Verify destination connection

  • "Trigger job not found" - Check job name for AfterJob mode

read_certificates

List SSL/TLS certificates used for secure connections.

RETURNS: Array of certificate objects with name, subject, expiration, and thumbprint.

COMMON ERRORS:

  • "No certificates found" - None uploaded yet

  • "Access denied" - Certificate access may be restricted

write_certificates

Upload SSL/TLS certificates for secure connections.

RETURNS: Certificate details with confirmation of upload.

COMMON ERRORS:

  • "Invalid certificate format" - Must be base64-encoded

  • "Certificate expired" - Check expiration date

  • "Duplicate certificate" - Certificate already exists

  • "Invalid store type" - Check provider documentation

read_workspaces

Access CData Sync workspaces for multi-tenant organization.

WORKSPACES:

  • Organize jobs, connections, and transformations

  • Isolate resources between teams/projects

  • Support multi-tenant deployments

RETURNS:

  • list: Array of workspace objects with Id and Name

  • get: Detailed workspace information

  • count: Total number of workspaces

COMMON ERRORS:

  • "Workspace not found" - Check workspace name spelling

  • "Access denied" - Requires appropriate permissions

write_workspaces

Create, update, or delete CData Sync workspaces.

WORKSPACE NAMES:

  • Must be unique across the system

  • Cannot contain special characters except _ - .

  • Case-sensitive

RETURNS:

  • create: New workspace details

  • update: Updated workspace information

  • delete: Success confirmation

COMMON ERRORS:

  • "Workspace already exists" - Use unique names

  • "Workspace in use" - Cannot delete workspace with active resources

  • "Invalid name" - Check naming requirements

Prompts

Interactive templates invoked by user choice

NameDescription

No prompts

Resources

Contextual data attached and managed by the client

NameDescription

No resources

Latest Blog Posts

MCP directory API

We provide all the information about MCP servers via our MCP API.

curl -X GET 'https://glama.ai/api/mcp/v1/servers/CDataSoftware/cdata-sync-mcp-server'

If you have feedback or need assistance with the MCP directory API, please join our Discord server