Description of problem: The ISPN-1015 fix added a buffer to the HotRod client transport to reduce the number of syscalls. But that buffer's size is not fixed, instead it's based on Socket.getSendBufferSize(), which in turn is based on the kernel's net.core.wmem_max. JGroups needs net.core.wmem_max greater or equal than UDP.ucast_send_buf_size, which is 1MB by default. That is probably too much for the TcpTransport's buffer, because the buffer is also duplicated in the OS.
I have used the reproducer from https://bugzilla.redhat.com/show_bug.cgi?id=1448920 and CR1 fails in the second test: Read operation. It is failing for me with OOME exceptions. Apparently, one commit is missing here in CR1 and it can be found in branch JDG_6.6.0.GA-bz-1448920 in this repo: git://git.app.eng.bos.redhat.com/infinispan/infinispan.git When I apply this commit to CR1 build and rebuild the hotrod library, the test passes. The commit is called: [BZ1448920] Limit buffer sizes for reads and writes
This product has been discontinued or is no longer tracked in Red Hat Bugzilla.