Server/JRockit Hang due to excessive GC worker Threads (Doc ID 1326276.1)

Last updated on NOVEMBER 05, 2016

Applies to:

Oracle JRockit - Version R28 to R28 [Release JRockit]
Information in this document applies to any platform.

Symptoms

Customer was running WLS 10.3.2 running on Jrockit JDK : JRockit(R) R28.0.1-21-133393-1.6.0_20-20100512-2127-solaris-sparcv9 on a 64 bit Oracle Solaris Sparc physical machine.


Very frequently, some managed server was going to a hung state, causing jvm to freeze.

 

Before the hang, CPU utilization, memory, and GC everything is under control

 

Once the issue manifests, the JVM does not respond to any jrcmd commands (to collect thread dumps OR to perform any debugging). At this time IO does not occur. It comes to a standstill.
So after the issue manifests, GC logs are not populated as well as server logs also do not get populated.

 

The situation persists forever, the only way to recover is kill -9

 

It was also observed that the CPU utilization increases by 25% when the JVM is in a hung state. It stays the same till the JVM is killed.

There were no specific errors seen in the log filers which could give an idea on the issue. No JVM commands respond too at the time of the issue.

Cause

Sign In with your My Oracle Support account

Don't have a My Oracle Support account? Click to get started

My Oracle Support provides customers with access to over a
Million Knowledge Articles and hundreds of Community platforms