I’m having issues getting a single-node Cassandra cluster to run on a Ubuntu
16.04 VM with only IPv6 available. I’m running Oracle Java 8
(8u121-1~webupd8~2), Cassandra 3.10 (installed via the Cassandra
http://www.apache.org/dist/cassandra/debian packages.)
I consistently get a “Protocol family unavailable” exception:
ERROR [main] 2017-04-11 09:54:23,991 CassandraDaemon.java:752 - Exception
encountered during startup
java.lang.RuntimeException: java.net.SocketException: Protocol family
unavailable
at
org.apache.cassandra.net.MessagingService.getServerSockets(MessagingService.java:730)
~[apache-cassandra-3.10.jar:3.10]
at
org.apache.cassandra.net.MessagingService.listen(MessagingService.java:664)
~[apache-cassandra-3.10.jar:3.10]
at
org.apache.cassandra.net.MessagingService.listen(MessagingService.java:648)
~[apache-cassandra-3.10.jar:3.10]
at
org.apache.cassandra.service.StorageService.prepareToJoin(StorageService.java:773)
~[apache-cassandra-3.10.jar:3.10]
at
org.apache.cassandra.service.StorageService.initServer(StorageService.java:666)
~[apache-cassandra-3.10.jar:3.10]
at
org.apache.cassandra.service.StorageService.initServer(StorageService.java:612)
~[apache-cassandra-3.10.jar:3.10]
at
org.apache.cassandra.service.CassandraDaemon.setup(CassandraDaemon.java:394)
[apache-cassandra-3.10.jar:3.10]
at
org.apache.cassandra.service.CassandraDaemon.activate(CassandraDaemon.java:601)
[apache-cassandra-3.10.jar:3.10]
at
org.apache.cassandra.service.CassandraDaemon.main(CassandraDaemon.java:735)
[apache-cassandra-3.10.jar:3.10]
Caused by: java.net.SocketException: Protocol family unavailable
at sun.nio.ch.Net.bind0(Native Method) ~[na:1.8.0_121]
at sun.nio.ch.Net.bind(Net.java:433) ~[na:1.8.0_121]
at sun.nio.ch.Net.bind(Net.java:425) ~[na:1.8.0_121]
at
sun.nio.ch.ServerSocketChannelImpl.bind(ServerSocketChannelImpl.java:223)
~[na:1.8.0_121]
at sun.nio.ch.ServerSocketAdaptor.bind(ServerSocketAdaptor.java:74)
~[na:1.8.0_121]
at
org.apache.cassandra.net.MessagingService.getServerSockets(MessagingService.java:714)
~[apache-cassandra-3.10.jar:3.10]
... 8 common frames omitted
`lo` (loopback) has both `inet` and `inet6` addresses, but `eth0` has no `inet`
addresses, so only inet6 addr entries (both a local and a global scope address
are configured).
My configuration changes:
listen_address: <local or global scope IPv6 address>
listen_interface_prefer_ipv6: true
Tracing through the source code the exception shows that it is the
listen_address value above that throws the exception, changing it back to
127.0.0.1 makes the server work again (but then I don’t get to use it on my
local network). I tried both the local and the global scope IPv6 address.
I tried changing the JVM configuration to prefer IPv6 by editing
/etc/cassandra/cassandra-env.sh:
--- etc/cassandra/cassandra-env.sh 2017-01-31 16:29:32.000000000 +0000
+++ /etc/cassandra/cassandra-env.sh 2017-04-11 09:52:51.456000000 +0000
@@ -290,6 +290,9 @@
# to the location of the native libraries.
JVM_OPTS="$JVM_OPTS -Djava.library.path=$CASSANDRA_HOME/lib/sigar-bin"
+#JVM_OPTS="$JVM_OPTS -Djava.net.preferIPv4Stack=true"
+JVM_OPTS="$JVM_OPTS -Djava.net.preferIPv6Addresses=true"
+
JVM_OPTS="$JVM_OPTS $MX4J_ADDRESS"
JVM_OPTS="$JVM_OPTS $MX4J_PORT"
JVM_OPTS="$JVM_OPTS $JVM_EXTRA_OPTS"
But this makes no difference
I also tried using `listen_interface` instead, but that only changes the error
message to:
ERROR [main] 2017-04-11 10:35:16,426 CassandraDaemon.java:752 - Exception
encountered during startup: Configured listen_interface "eth0" could not be
found
What else can I do?
Martijn Pieters
