Cassandra startup problem: Attempt to assign id to existing column family

Question

I'm using cassandra 0.7.4 on centos5.5 x86_64 with jdk-1.6.0_24 64-Bit. When I restart it , it throw out:

---

ERROR 11:37:32,009 Exception encountered during startup.
java.io.IOError: org.apache.cassandra.config.ConfigurationException: Attempt to assign id to existing column family.
    at org.apache.cassandra.config.DatabaseDescriptor.loadSchemas(DatabaseDescriptor.java:476)
    at org.apache.cassandra.service.AbstractCassandraDaemon.setup(AbstractCassandraDaemon.java:138)
    at org.apache.cassandra.service.AbstractCassandraDaemon.activate(AbstractCassandraDaemon.java:314)
    at org.apache.cassandra.thrift.CassandraDaemon.main(CassandraDaemon.java:79)
Caused by: org.apache.cassandra.config.ConfigurationException: Attempt to assign id to existing column family.
    at org.apache.cassandra.config.CFMetaData.map(CFMetaData.java:223)
    at org.apache.cassandra.config.DatabaseDescriptor.loadSchemas(DatabaseDescriptor.java:472)
    ... 3 more    

---

I try to location the problem: when I delete the file of the system keyspace ,It can restart sucess! So I think this problem is cause by system Keyspace,even at the CF Scheam.

Then I build a new test environment, I know this proble is cause by this opeartion

update keyspace system with replication_factor=3;

But now how can i repair it ?! There are many data on this cluster,and I couldn't lose data.

I have already do update keyspace system with replication_factor=1; ,but the problem still exist. I try to use nodetool to repair after or befor flush, all no effect.

How can I restart cassandra without lose data ? Who can help me?

Answer 1

You should never modify the system keyspace unless you really, really know what you are doing. (If you have to ask, you don't. :)

So, the answer is: don't do that.

To recover, you should set initial_token in cassandra.yaml to your node's current token (which you can see with "nodetool ring"), then delete the system keyspace and restart. Then you'll need to recreate your columnfamily definitions, but your data will not be affected.