There is a JIRA for using the new —release build flag available in JDK 9+ to safely build binaries meant to run on JDK8. See https://issues.apache.org/jira/browse/HBASE-25465
This issue with the ByteBuffer signature change and linkage problems is exactly what motivated me to look into it. > On May 27, 2021, at 7:34 AM, Wellington Chevreuil > <[email protected]> wrote: > > I had faced below errors when trying to run current master branch version > on jdk8 (build was also on jdk8): > > 2021-05-26T23:02:20,202 TRACE [RS-EventLoopGroup-1-1] > ipc.NettyRpcServer: Connection /192.168.56.105:44592; caught > unexpected downstream exception. >> java.lang.NoSuchMethodError: java.nio.ByteBuffer.flip()Ljava/nio/ByteBuffer; >> at >> org.apache.hadoop.hbase.ipc.NettyRpcServerPreambleHandler.channelRead0(NettyRpcServerPreambleHandler.java:48) >> ~[hbase-server-3.0.0-SNAPSHOT.jar:3.0.0-SNAPSHOT] >> >> > This specific netty related one happens when RSes try to check into Master, > leaving the cluster unusable, but the same NSME happens at other parts of > the code too, eventually causing processes to crash. > > Running on jdk11 fixes it. Anyone else seeing the same? I guess this was > not expected, and our official goal is to still support jdk8 for the > upcoming 3.0 release?
