IBM Tivoli Monitoring > Version 6.3 Fix Pack 2 > Installation Guides > Installation Guide > Tivoli Data Warehouse solutions: common procedures > Tuning the performance of the Warehouse Proxy
IBM Tivoli Monitoring, Version 6.3 Fix Pack 2
Timeout values
You can set two environment variables to control the timeout value. One variable is set on the application agent, the other on the Warehouse Proxy.
- KHD_STATUSTIMEOUT
- The time, in seconds, to wait for a status from the Warehouse Proxy before sending an export request again. The default value is 900 seconds, or 15 minutes.
- KHD_SRV_STATUSTIMEOUT
- The timeout value, in seconds, for the work queue to perform work. The default value is 600 seconds, or 10 minutes.
Export requests are rejected by the Warehouse Proxy are the following four reasons:
- The time between when an export request is sent to the work queue and when it is extracted from the queue exceeds the timeout. If you have tracing for the Warehouse Proxy set to ERROR, an error similar to the following is logged in the Warehouse Proxy log file:
REJECTED: The export for the originnode OriginNodeName, the application applicationName and the table tableName has been rejected for timeout reason in stage END_QUEUE.
The time between when an export request is sent to the work queue and when the work queue starts to do existence checking in the database exceeds the timeout. If you have tracing for the Warehouse Proxy set to ERROR, an error similar to the following is logged in the Warehouse Proxy log file and the WAREHOUSELOG table: Sample data rejected for timeout reason at stage START EXPORT
The time between when an export request is sent to the work queue and when the work queue fetches all the rows in the sample exceeds the timeout. If you have tracing for the Warehouse Proxy set to ERROR, a message similar to the following is logged in the Warehouse Proxy log file and the WAREHOUSELOG table: Sample data rejected for timeout reason at stage START SAMPLE
The time between when an export request is sent to the work queue and when the work queue commits the rows in the database exceeds the timeout. If you have tracing for the Warehouse Proxy set to ERROR, a message similar to the following is logged in the Warehouse Proxy log file and the WAREHOUSELOG table: Sample data rejected for timeout reason at stage COMMIT
The KHD_SRV_STATUSTIMEOUT variable should be set less than KHD_STATUSTIMEOUT by at least 60 seconds. Do not change these values unless there is a problem in the environment.
Parent topic:
Tuning the performance of the Warehouse Proxy