This issue can be closed once the performance tests have been run comparing DS 7.2 using JDK17 with parallel GC and JDK17 with G1 in order to understand the performance impact of this garbage collector. Preliminary testing seems to show a small but negligible impact. If this is confirmed by tests then we can simplify our JVM tuning guidelines by recommending G1, which is the default GC algorithm, for all deployments.
The goal of this task is to test the server, not the client. Therefore, the JDK used for running the client, along with its parameters, should remain the same as those used for OPENDJ-8313.