DataStage: How To Resolve ‘Scratch Space Full’ Error
- By : Mydatahack
- Category : DataStage, ETL
- Tags: APT_CONFIG_FILE, DataStage, ETL, Scratch Disk
When the data volume is large, DataStage uses a scratch disk to process data. The default scratch disk space is usually the Scratch folder in the Server folder where the application is installed. To use a larger scratch disk space, we can create a custom configuration file.
The default configuration file is called default.apt and sits in the Configuration folder (for example /opt/IBM/InformationServer/Server/Configuration/). In this folder, we can create a new configuration file (new_2node.apt). The example uses two nodes. Change the resource disk and resource scratch disk parameters as below.
{ node "node1" { fastname "<Your DataStage Server URL>" pools "" resource disk "/scratch" {pools ""} resource scratchdisk "/scratch" {pools ""} } node "node2" { fastname "<Your DataStage Server URL>" pools "" resource disk "/scratch" {pools ""} resource scratchdisk "/scratch" {pools ""} } }
Go back to the parallel job that gave the error. In the job parameter, add the variable APT_CONFIG_FILE and set the new config file path (/opt/IBM/InformationServer/Server/Configuration/new_2node.apt) in order to override the default config file path.
Check out the further reference here.