keepAliveCache.socketHealthCheck Is Causing A Severe Performance Degradation (Doc ID 2021598.1)

Last updated on FEBRUARY 21, 2024

Applies to:

Symptoms

WLS 10.3.6, NIO socket muxer is used

After setting -Dhttp.keepAliveCache.socketHealthCheckTimeout=1 (in this case, the cache is checking the health condition before returning the cached connection to the client for use), you are reporting a severe performance degradation.

Tracking the way ExecuteThread 36 evolves in the various thread dumps collected over the time we see:

From reading the source code, this implies that the socket read didn't timeout as it should. 

Changes

Cause

My Oracle Support provides customers with access to over a million knowledge articles and a vibrant support community of peers and Oracle experts.