X
  • Tuesday, May 16, 2017

Java SE support for Docker CPU and memory limits

By: Donald Smith | Sr. Director of Product Management

Special thanks to Charlie Hunt, contributor to multiple OpenJDK projects for providing much of the content of this blog.

As a follow-up to our blog announcing OpenJDK Project Portola, a project to provide a port of the JDK to Alpine Linux, there are several enhancements coming to Java SE that also fit with Docker and Alpine Linux’s slogan, "Small, Simple. Secure." Two worth mentioning are:

  • Support of Docker CPU and memory limits.
  • Project Jigsaw’s jlink, which offers the ability to produce a minimal JRE based the modules the application uses or requires (we'll cover jlink in a downstream blog).

Docker image for Oracle Server JRE is now available on Docker Store


Java SE support for Docker CPU and memory limits

For those who have been using Docker’s CPU and memory limits with Java applications may have experienced some challenges. In particular CPU limits since the JVM internally and transparently sets the number of GC threads, and JIT compiler threads. These can be explicitly set via command line options, -XX:ParallelGCThreads and -XX:CICompilerCount. For memory limits, the maximum Java heap size can also be set explicitly via JVM command line option, -Xmx. However, in cases where none of the aforementioned JVM command line options are specified, when a Java application using Java SE 8u121 and earlier, when run in a Docker container, the JVM will use the underlying host configuration of CPUs and memory when transparently determining the number of GC threads, JIT compiler threads and maximum Java heap to use.

As of Java SE 8u131, and in JDK 9, the JVM is Docker-aware with respect to Docker CPU limits transparently. That means if -XX:ParalllelGCThreads, or -XX:CICompilerCount are not specified as command line options, the JVM will apply the Docker CPU limit as the number of CPUs the JVM sees on the system. The JVM will then adjust the number of GC threads and JIT compiler threads just like it would as if it were running on a bare metal system with number of CPUs set as the Docker CPU limit. If -XX:ParallelGCThreads or -XX:CICompilerCount are specified as JVM command line options, and Docker CPU limit are specified, the JVM will use the -XX:ParallelGCThreads and -XX:CICompilerCount values.

For Docker memory limits, there is a little more work for the transparent setting of a maximum Java heap. To tell the JVM to be aware of Docker memory limits in the absence of setting a maximum Java heap via -Xmx, there are two JVM command line options required, -XX:+UnlockExperimentalVMOptions -XX:+UseCGroupMemoryLimitForHeap. The -XX:+UnlockExperimentalVMOptions is required because in a future release transparent identification of Docker memory limits is the goal. When these two JVM command line options are used, and -Xmx is not specified, the JVM will look at the Linux cgroup configuration, which is what Docker containers use for setting memory limits, to transparently size a maximum Java heap size. FYI, Docker containers also use cgroups configuration for CPU limits too.

These two modifications in Java SE 8u131 and JDK 9 are expected to improve the Java running in Docker experience.

The work for both of these enhancements have been tracked in OpenJDK as follows:

Be the first to comment

Comments ( 0 )
Please enter your name.Please provide a valid email address.Please enter a comment.CAPTCHA challenge response provided was incorrect. Please try again.Captcha