Observed lots of timeouts caused by a synchronization issue in between BinaryLinkServer::terminate() that is invoked by the JCStress::run() and tries to join handlers with:
72 synchronized (handlers) {
73 for (Handler h : handlers) {
74 h.join();
75 }
76 }
in BinaryLinkServer.
At the same time Handler tries tor remove itself from the handlers = Collections.synchronizedCollection(new ArrayList<>())
205 handlers.remove(this);

It looks like in case of the failed binary connection this leads to a deadlock.
"Thread-16" #43 prio=5 os_prio=0 tid=0x00007ffbd4001800 nid=0x33db waiting for monitor entry [0x00007ffc455b6000]
   java.lang.Thread.State: BLOCKED (on object monitor)
at java.util.Collections$SynchronizedCollection.remove(java.base@9-internal/Collections.java:2039)
- waiting to lock <0x0000000101254178> (a java.util.Collections$SynchronizedCollection)
at org.openjdk.jcstress.link.BinaryLinkServer$Handler.run(BinaryLinkServer.java:205)

"main" #1 prio=5 os_prio=0 tid=0x00007ffd8800f000 nid=0x314f in Object.wait() [0x00007ffd91d0b000]
   java.lang.Thread.State: WAITING (on object monitor)
at java.lang.Object.wait(java.base@9-internal/Native Method)
- waiting on <0x00000001013ccd98> (a org.openjdk.jcstress.link.BinaryLinkServer$Handler)
at java.lang.Thread.join(java.base@9-internal/Thread.java:1352)
- waiting to re-lock in wait() <0x00000001013ccd98> (a org.openjdk.jcstress.link.BinaryLinkServer$Handler)
at java.lang.Thread.join(java.base@9-internal/Thread.java:1426)
at org.openjdk.jcstress.link.BinaryLinkServer.terminate(BinaryLinkServer.java:74)
- locked <0x0000000101254178> (a java.util.Collections$SynchronizedCollection)
at org.openjdk.jcstress.JCStress.run(JCStress.java:128)
at org.openjdk.jcstress.Main.main(Main.java:60)