Paul Chandler created CASSANDRA-21286:
-----------------------------------------
Summary: Multiple nodes on one machine does not work for 6.0
Key: CASSANDRA-21286
URL: https://issues.apache.org/jira/browse/CASSANDRA-21286
Project: Apache Cassandra
Issue Type: Bug
Components: Transactional Cluster Metadata
Reporter: Paul Chandler
I have a setup for testing which has worked for 4.x and 5.0. This is to run
multiple nodes on one machine, using multiple loopback addresses 127.0.0.1,
127.0.0.2 and 127.0.0.3
All the ports stay the same and bind to the different loopback addresses,
except for the JMX port which always binds to 127.0.0.1 so the jmx port is set
to 7199, 7198,7197
So this works for 4.x and 5.0 and I have used this setup to test upgrades in
the past.
However now when I to the same with Cassandra 6, when the second node starts,
it fails with the following error messages
{code:java}
INFO [main] 2026-04-05T16:28:52,935 Gossiper.java:2062 - No gossip backlog;
proceedingException (java.lang.IllegalStateException) encountered during
startup: Can not commit transformation: "INVALID"(New addresses
NodeAddresses{broadcastAddress=/127.0.0.2:7000, localAddress=/127.0.0.2:7000,
nativeAddress=/127.0.0.1:9042} conflicts with existing node NodeId{id=1} with
addresses NodeAddresses{broadcastAddress=localhost/127.0.0.1:7000,
localAddress=localhost/127.0.0.1:7000,
nativeAddress=localhost/127.0.0.1:9042}).java.lang.IllegalStateException: Can
not commit transformation: "INVALID"(New addresses
NodeAddresses{broadcastAddress=/127.0.0.2:7000, localAddress=/127.0.0.2:7000,
nativeAddress=/127.0.0.1:9042} conflicts with existing node NodeId{id=1} with
addresses NodeAddresses{broadcastAddress=localhost/127.0.0.1:7000,
localAddress=localhost/127.0.0.1:7000,
nativeAddress=localhost/127.0.0.1:9042}). at
org.apache.cassandra.tcm.ClusterMetadataService.lambda$commit$6(ClusterMetadataService.java:581)
at
org.apache.cassandra.tcm.ClusterMetadataService.commit(ClusterMetadataService.java:625)
at
org.apache.cassandra.tcm.ClusterMetadataService.commit(ClusterMetadataService.java:578)
at
org.apache.cassandra.tcm.transformations.Register.register(Register.java:148)
at
org.apache.cassandra.tcm.transformations.Register.register(Register.java:135)
at
org.apache.cassandra.tcm.transformations.Register.register(Register.java:123)
at
org.apache.cassandra.tcm.transformations.Register.register(Register.java:166)
at
org.apache.cassandra.tcm.transformations.Register.maybeRegister(Register.java:117)
at
org.apache.cassandra.service.StorageService.initServer(StorageService.java:854)
at
org.apache.cassandra.service.CassandraDaemon.setup(CassandraDaemon.java:396) at
org.apache.cassandra.service.CassandraDaemon.activate(CassandraDaemon.java:836)
at
org.apache.cassandra.service.CassandraDaemon.main(CassandraDaemon.java:990)
{code}
Specifically it looks like the nativeAddress is not correct here:
"INVALID"(New addresses NodeAddresses{broadcastAddress=/127.0.0.2:7000,
localAddress=/127.0.0.2:7000, nativeAddress={*}/127.0.0.1:9042{*}}
I have tested this as a new cluster pure 6.0 and also as an upgrade from 5.0,
and both fail.
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
