Configuring a Data Engine Proxy Server¶
Overview¶
Each type of OMAG Server is configured by creating a configuration document. For data engine proxy server, following can be configured:
Configuring the basic properties
Configure the basic properties¶
The basic properties are used in logging and events originating from the server. They help to document the purpose of the server (which helps with problem determination) and enable performance improvements by allowing the server to ignore activity or metadata that is not relevant to its operation.
The basic properties include two unique identifiers that are set up when you first create the configuration document:
Property | Description |
---|---|
localServerId |
Unique identifier for this server. By default, this is initialized to a randomly generated Universal Unique identifier (UUID). |
localServerName |
Meaningful name for the server for use in messages and UIs. Ideally this value is unique to aid administrators in understanding the source of messages and events from the server. This value is set to the server name assigned when the configuration is created. |
The other basic properties have values that can be changed through the admin services API:
Property | Description |
---|---|
localServerType |
Descriptive type name for the server. Again this is useful information for the administrator to understand the role of the server. The default value is Open Metadata and Governance Server . |
organizationName |
Descriptive name for the organization that owns the local server/repository. This is useful when the open metadata repository cluster consists of metadata servers from different organizations, or different departments of an enterprise. The default value is null . |
localServerUserId |
UserId to use for server-initiated REST calls. The default is OMAGServer . |
localServerPassword |
Password to use for server-initiated REST calls. The default is null . This means that only the userId is sent in the HTTP header. |
maxPageSize |
The maximum page size that can be set on requests to the server. The default value is 1000 . A value of zero means unlimited page size. Although supported, the zero value is not recommended because it provides no protection from a large request denial of service attack. |
The sections that follow cover how to set up these values.
Set server type name¶
The server type name should be set to something that describes the OMAG Server's role. It may be the name of a specific product that it is enabling, or a role in the metadata and governance landscape. The default value is Open Metadata and Governance Server
.
If you have no specific value to set the server type name to, we recommend that you set the server type name to null. This will cause the server start up process will derive a standard server type name based on the rest of the configuration for the server.
POST - set server type
{{platformURLRoot}}/open-metadata/admin-services/users/{{adminUserId}}/servers/{{serverName}}/server-type?typeName="{{serverTypeName}}"
Set organization name¶
The organization name may be the owning organization or department or team supported by the server.
POST - set organization name
{{platformURLRoot}}/open-metadata/admin-services/users/{{adminUserId}}/servers/{{serverName}}/organization-name?name="{{organizationName}}"
Set the server's userId and optional password¶
The server's userId is used when processing requests that do not have an end user, such as receiving an event from a topic. The default value is OMAGServer
. Ideally each server should have its own userId, so it is possible to restrict the resources that each server has access to and identify the origin of updates to the metadata elements.
If the password is specified as well, the userId and password combination are used to provide authentication information on each REST call made by the server.
POST - set server's userId
{{platformURLRoot}}/open-metadata/admin-services/users/{{adminUserId}}/servers/{{serverName}}/server-user-id?id="{{serverUserId}}"
POST - set server's password
{{platformURLRoot}}/open-metadata/admin-services/users/{{adminUserId}}/servers/{{serverName}}/server-user-password?password="{{serverUserPassword}}"
Set the maximum page size for REST API requests¶
The maximum page size value sets an upper limit on the number of results that a caller can request on any paging REST API to this server. Setting maximum page size helps to prevent a denial of service attack that uses very large requests to overwhelm the server. A value of 0
means no limit, and leaves the server open to such attacks.
POST - set maximum page size
{{platformURLRoot}}/open-metadata/admin-services/users/{{adminUserId}}/servers/{{serverName}}/max-page-size?limit={{maxPageSize}}
Configuring the audit log
Configure the audit log¶
Egeria's audit log provides a configurable set of destinations for audit records and other diagnostic logging for an OMAG Server. Some destinations also support a query interface to allow an administrator to understand how the server is running.
Each audit log record has a severity that can be used to route it to one or more specific destinations. Therefore, when an audit log destination is configured, it is optionally supplied with a list of severities to filter audit log records it should receive.
The audit log severities are as follows:
Severity | Description |
---|---|
Information |
The server is providing information about its normal operation. |
Event |
An event was received from another member of the open metadata repository cohort. |
Decision |
A decision has been made related to the interaction of the local metadata repository and the rest of the cohort. |
Action |
An Action is required by the administrator. At a minimum, the situation needs to be investigated and if necessary, corrective action taken. |
Error |
An error occurred, possibly caused by an incompatibility between the local metadata repository and one of the remote repositories. The local repository may restrict some of the metadata interchange functions as a result. |
Exception |
An unexpected exception occurred. This means that the server needs some administration attention to correct configuration or fix a logic error because it is not operating as a proper peer in the open metadata repository cohort. |
Security |
Unauthorized access to a service or metadata instance has been attempted. |
Startup |
A new component is starting up. |
Shutdown |
An existing component is shutting down. |
Asset |
An auditable action relating to an asset has been taken. |
Types |
Activity is occurring that relates to the open metadata types in use by this server. |
Cohort |
The server is exchanging registration information about an open metadata repository cohort that it is connecting to. |
Trace |
This is additional information on the operation of the server that may be of assistance in debugging a problem. It is not normally logged to any destination, but can be added when needed. |
PerfMon |
This log record contains performance monitoring timing information for specific types of processing. It is not normally logged to any destination, but can be added when needed. |
<Unknown> |
Uninitialized Severity |
The default audit log destination is the console audit log destination. This writes selected parts of each audit log record to "standard out" (stdout). It is configured to receive log records of all severities except Trace
and PerfMon
.
The default audit log destination is added automatically to Cohort Member servers when a repository, or cohort connection is configured.
For View Servers and Governance Servers, it is only added when other audit log destinations are configured. Therefore, is it necessary to issue at least one of the following configuration commands for the audit log for these types of servers.
Add audit log destinations¶
If the server is a development or test server, then the default audit log configuration is probably sufficient, and you should use the following command:
POST - set default audit log destination
{{platformURLRoot}}/open-metadata/admin-services/users/{{adminUserId}}/servers/{{serverName}}/audit-log-destinations/default
Note: Using this command overrides all previous audit log destinations configured for the server.
If this server is a production server then you will probably want to set up the audit log destinations explicitly. You can add multiple destinations and each one can be set up to receive different severities of audit log records.
There are various destinations that can be configured for the audit log:
Since the default audit log destination is also a console audit log destination, only use this option to add the Trace
and PerfMon
severities.
POST - add console audit log destination
{{platformURLRoot}}/open-metadata/admin-services/users/{{adminUserId}}/servers/{{serverName}}/audit-log-destinations/console
The body of the request should be a list of severities
If an empty list is passed as the request body then all severities are supported by the destination.
This destination writes JSON files in a shared directory. One file for each audit log record.
POST - add JSON file-based audit log destination
{{platformURLRoot}}/open-metadata/admin-services/users/{{adminUserId}}/servers/{{serverName}}/audit-log-destinations/files
The body of the request should be a list of severities
If an empty list is passed as the request body then all severities are supported by the destination.
This destination writes each log record as an event on the supplied event topic. It assumes that the event bus is set up first.
POST - add event-based audit log destination
{{platformURLRoot}}/open-metadata/admin-services/users/{{adminUserId}}/servers/{{serverName}}/audit-log-destinations/event-topic
The body of the request should be a list of severities
If an empty list is passed as the request body then all severities are supported by the destination.
This writes full log records to the slf4j ecosystem. When configuring slf4j as destination you also need to specify audit log logger category via the application properties of the OMAG Server Platform. This is described in Connecting the OMAG Audit Log Framework section of the developer logging guide.
The configuration of the slf4j ecosystem determines it ultimate destination(s).
POST - add slf4j audit log destination
{{platformURLRoot}}/open-metadata/admin-services/users/{{adminUserId}}/servers/{{serverName}}/audit-log-destinations/slf4j
The body of the request should be a list of severities
If an empty list is passed as the request body then all severities are supported by the destination.
This sets up an audit log destination that is described though a connection. In this case, the connection is passed in the request body and the supported severities are supplied in the connection's configuration properties.
POST - add connection-based audit log destination
{{platformURLRoot}}/open-metadata/admin-services/users/{{adminUserId}}/servers/{{serverName}}/audit-log-destinations/connection
It is also possible to set up all the audit log destinations in one command as a list of connections. Using this option overrides all previous audit log destinations and so can be used as the update command. The list of connections is passed in the request body and the supported severities are supplied in each connection's configuration properties.
POST - add a list of connection-based audit log destinations
{{platformURLRoot}}/open-metadata/admin-services/users/{{adminUserId}}/servers/{{serverName}}/audit-log-destinations
Retrieving audit log destinations¶
The configured list of audit log destinations can be retrieved using this command:
GET - the list of configured audit log destinations
{{platformURLRoot}}/open-metadata/admin-services/users/{{adminUserId}}/servers/{{serverName}}/audit-log-destinations
Updating audit log destinations¶
Audit log destinations can be updated individually, by qualified name using the following command:
POST - update connection-based audit log destination
{{platformURLRoot}}/open-metadata/admin-services/users/{{adminUserId}}/servers/{{serverName}}/audit-log-destinations/connection/{{qualifiedName}}
If you are not sure what the audit log connection is called, retrieve the list of configured audit log connections and the resulting list of audit log connections will include the qualified names.
Remove audit log destinations¶
The following will remove all audit log destinations, enabling you to add a new set of audit log destinations.
DELETE - clear all audit log destinations
{{platformURLRoot}}/open-metadata/admin-services/users/{{adminUserId}}/servers/{{serverName}}/audit-log-destinations
It is also possible to remove a single audit log destination using its connection's qualified name.
DELETE - clear then named audit log destination
{{platformURLRoot}}/open-metadata/admin-services/users/{{adminUserId}}/servers/{{serverName}}/audit-log-destinations/{{qualifiedName}}
Configuring the server security connector
Configure the server security connector¶
Metadata that is being aggregated from different sources is likely to need comprehensive access controls.
Egeria provides fine-grained security control for metadata access. It is implemented in a server metadata security connector that is called whenever requests are made for to the server.
Security is configured for a specific OMAG Server by adding a connection for this connector to the server's configuration documentusing the following command.
POST - configure security connector
{{platformURLRoot}}/open-metadata/admin-services/users/{{adminUserId}}/servers/{{serverName}}/security/connection
This passes in a connection used to create the server security connector in the request body.
{
"class": "Connection",
"connectorType": {
"class": "ConnectorType",
"connectorProviderClassName": "{fullyQualifiedJavaClassName}"
}
}
Example: set up the sample server security connector
For example, this is the connection that would set up the sample server security connector provided for the Coco Pharmaceuticals case study:
{
"class": "Connection",
"connectorType": {
"class": "ConnectorType",
"connectorProviderClassName": "org.odpi.openmetadata.metadatasecurity.samples.OpenMetadataServerSecurityProvider"
}
}
Determine configured security¶
GET - query the server security connector setting
{{platformURLRoot}}/open-metadata/admin-services/users/{{adminUserId}}/servers/{{serverName}}/security/connection
Response indicating no security
{
"class": "ConnectionResponse",
"relatedHTTPCode": 200
}
Response indicating a specific security connector
If the response looks more like the JSON below, a connector is configured. The connectorProviderClassName
tells you which connector is being used.
{
"class": "ConnectionResponse",
"relatedHTTPCode": 200,
"connection": {
"class": "Connection",
"connectorType": {
"class": "ConnectorType",
"connectorProviderClassName": "{fullyQualifiedJavaClassName}"
}
}
}
Remove configured security¶
DELETE - remove configured security connector
{{platformURLRoot}}/open-metadata/admin-services/users/{{adminUserId}}/servers/{{serverName}}/security/connection
This removes all authorization checking from the server.
Configuring the Data Engine Proxy Services
Configuring Data Engine Proxy Services¶
Course-grained helper command
POST - Configure Date Engine Proxy Services
{{serverURLRoot}}/open-metadata/admin-services/users/{{userId}}/servers/{{serverName}}/data-engine-proxy-service/configuration
{
"class": "DataEngineProxyConfig",
"accessServiceRootURL": "{{{metadata-server-platform-url}}}",
"accessServiceServerName": "{{metadata-server-name}}",
"dataEngineConnection": {
"class": "Connection",
"connectorType": {
"class": "ConnectorType",
"connectorProviderClassName": "org.odpi.egeria.connectors.ibm.datastage.dataengineconnector.DataStageConnectorProvider"
},
"endpoint": {
"class": "Endpoint",
"address": "{{data-engine-endpoint}}",
"protocol": "https"
},
"userId": "{{data-engine-user}}",
"clearPassword": "{{data-engine-password}}"
},
"pollIntervalInSeconds": 60
}
Configuration Reference¶
Property | Description | Is mandatory |
---|---|---|
dataEngineConnection | OCF connection configuration object that defines the connector type and its properties. Refer to specific connector for detailed reference. Example provided for IBM Data Stage connector. | Yes |
pollIntervalInSeconds | The polling interval in seconds to call the sequence extracting metadata. | Yes |
Removing the Data Engine Services from the server configuration¶
Course-grained helper command
DELETE - Remove Data Engine Configuration from the server
{{serverURLRoot}}/open-metadata/admin-services/users/{{userId}}/servers/{{serverName}}/data-engine-proxy-service/configuration
Raise an issue or comment below