Configuration for a Highly Available Domain
To minimize system downtime, configure Informatica domain components to be highly available.
You can configure the following Informatica domain components to be highly available:
- Domain
- One node in the domain acts as a gateway to receive service requests from clients and routes them to the appropriate service and node. To prevent domain shutdown when the master gateway node is unavailable, configure more than one gateway node.
- Nodes
- Informatica services are processes that run on each node. You can configure Informatica services to restart automatically if it terminates unexpectedly.
- Application Services
- The application services run on nodes in the Informatica domain.
- Based on your license, you can configure the following high availability features for application services:
- - To minimize the application service downtime, configure backup nodes for application services.
- - To specify the resilience period for application services, review default settings and configure resilience timeout periods for application services.
- - To ensure PowerCenter Integration Service failover and recovery, configure the PowerCenter Integration Service to store process state information on a POSIX compliant shared file system or in a database.
- Application Clients
- Application clients provide access to Informatica functionality, and they run on user machines. Application clients send requests to the Service Manager or application services.
- You can configure resilience timeout periods for command line programs. You cannot configure a PowerCenter client resilience timeout.
- External Systems
- Use highly available versions of external systems such as source and target databases, message queues, and FTP servers.
- Network
- Make the network highly available by configuring redundant components such as routers, cables, and network adapters.
Application Service Resilience Configuration
When a temporary network failure occurs, application services try to reconnect to other application services for the duration of the resilience timeout. You can configure the resilience timeout for application services.
When an application service connects to another application service in the domain, the service that initiates the connection is a client of the other service.
You can configure application service resilience timeouts for the following application services:
- PowerCenter Application Services
- You can configure the resilience timeout and resilience timeout limits in the advanced properties of the PowerCenter Integration Service and PowerCenter Repository Service. The resilience timeout for application services that connects to a PowerCenter Integration Service or PowerCenter Repository Service is determined by one of the following values:
- - The service Resilience Timeout property. You can configure the resilience timeout for the service in the service properties. To disable resilience for a service, set the resilience timeout to 0.
- - The domain Resilience Timeout property. To use the resilience timeout configured for the domain, set the resilience timeout for the service to blank.
- - The service Limit on Resilience Timeout property. If the service limit on resilience timeout is smaller than the resilience timeout for the connecting client, the client uses the limit as the resilience timeout. To use the limit on resilience timeout configured for the domain, set the service resilience limit to blank.
- - The domain Limit on Resilience Timeout property. To use the resilience timeout configured for the domain, set the limit on resilience timeout for the service to blank.
You can configure the resilience timeout for the SAP BW Service in the general properties for the service. The SAP BW Service resilience timeout property is called the Retry Period.
Note: A client cannot be resilient to service interruptions if you disable the service in the Administrator tool. If you disable the service process, the client is resilient to the interruption in service.
Application Service Failover Configuration
Based on your license, you can configure backup nodes so that application services can failover to another node when the primary node fails. Configure backup nodes when you create or update an application service.
When you configure a backup node, verify that the node has access to run-time files that each application service requires to process data integration tasks such as workflows and mappings. For example, a workflow might require parameter files, input files, or output files.
PowerCenter Integration Service Failover and Recovery Configuration
During failover and recovery, the PowerCenter Integration Service needs to access state of operation files and process state information.
The state of operation files store the state of each workflow and session operation. The PowerCenter Integration Service always stores the state of each workflow and session operation in files in the $PMStorageDir directory of the PowerCenter Integration Service process.
Process state information includes information about which node was running the master PowerCenter Integration Service process and which node was running each session. You can configure the PowerCenter Integration Service to store process state information on a cluster file system or in the PowerCenter repository database.
Store High Availability Persistence on a Cluster File System
By default, the PowerCenter Integration Service stores process state information along with the state of operation files in the $PMStorageDir directory of the Integration Service process. You must configure the $PMStorageDir directory for each PowerCenter Integration Service process to use the same directory on a cluster file system.
Nodes that run the PowerCenter Integration Service must be on the same cluster file system so that they can share resources. Also, nodes within a cluster must be on the cluster file system’s heartbeat network. Use a highly available cluster file system that is configured for I/O fencing. The hardware requirements and configuration of an I/O fencing solution are different for each file system.
The following cluster file systems are certified by Informatica for use for PowerCenter Integration Service failover and session recovery:
- Storage Array Network
- Veritas Cluster Files System (VxFS)
- IBM General Parallel File System (GPFS)
- Network Attached Storage using NFS v3 protocol
- EMC UxFS hosted on an EMV Celerra NAS appliance
- NetApp WAFL hosted on a NetApp NAS appliance
Contact the file system vendors directly to evaluate which file system matches your requirements.
Store High Availability Persistence in a Database
You can configure the PowerCenter Integration Service to store process state information in database tables. When you configure the PowerCenter Integration Service to store process state information in a database, the service still stores the state of each workflow and session operation in files in the $PMStorageDir directory. You can configure the $PMStorageDir directory to use a POSIX compliant shared file system. You do not need to use a cluster file system.
Configure the PowerCenter Integration Service to store process state information in database tables in the advanced properties. The PowerCenter Integration Service stores process state information in persistent database tables in the associated PowerCenter repository database.
During failover, automatic recovery of workflows resume when the service process can access the database tables.
Command Line Program Resilience Configuration
You can configure the resilience timeout that command line programs use to perform domain and service operations.
When you use the infacmd , pmcmd, or pmrep command line programs to connect to the domain or an application service the resilience timeout is determined by the command line option, an environment variable, or the default resilience timeout.
Use the following guidelines when you configure command line program resilience:
- Command line option
- You can set the resilience timeout for infacmd by using the -ResilienceTimeout command line option each time you run a command. You can set the resilience timeout for pmcmd by using the -timeout command line option each time you run a command. When you use pmrep connect to connect to a repository, you can use the -t command line option to set the resilience timeout for pmrep commands that use the connection.
- Environment variable.
- If you do not set the timeout option in the infacmd and pmcmd command line syntax, the infacmd and pmcmd command line programs use the value of the environment variable INFA_CLIENT_RESILIENCE_TIMEOUT that is configured on the client machine. If you do not set the timeout option when you use pmrep connect to connect to the repository, pmrep commands use the value of the environment variable INFA_CLIENT_RESILIENCE_TIMEOUT that is configured on the client machine.
- Default value
- If you do not use the command line option or the environment variable, the pmcmd and pmrep command line program uses the default resilience timeout of 180 seconds. If you do not use the command line option or the environment variable, the infacmd command line program uses the value of the domain Service Level Timeout property as the default resilience timeout.
- Limit on timeout
- If the limit on resilience timeout for the PowerCenter Integration Service or the PowerCenter Repository Service is smaller than the command line resilience timeout, the command line program uses the limit as the resilience timeout.
Note: PowerCenter does not provide resilience for a repository client when the PowerCenter Repository Service is running in exclusive mode.
Domain Failover Configuration
You can define multiple gateway nodes to prevent domain shutdown when the master gateway node is unavailable.
The first time that you install Informatica services, you create one gateway node. After you install Informatica, you can define additional gateway nodes. To define a gateway node, add a gateway node to the domain or configure a worker node to serve as a gateway node.
Node Restart Configuration
The Informatica services run the Service Manager on all nodes in the domain. You can configure the Informatica services to start automatically when a node terminates unexpectedly and restarts.
To restart the Informatica services when a node restarts, complete the following steps:
- •In a UNIX environment, you can create a script to automatically start the Informatica services when the node starts.
- •In a Windows environment, go to the Control Panel and configure the Informatica services to start automatically.
You can configure restart for all nodes, regardless of node type or node role.