I have a 3-node cluster running Cassandra 2.0.6 on CentOS 6.5, with Jave 1.7.0_51.
I ran a CQL statement like "alter table table_name with gc_grace_seconds=864000;" on node 1 in CQLSH, and it finished instantaneously. "desc keyspace" listed the table with the new value for gc_grace_seconds, and "select * from system.schema_columnfamilies" also reveals the new value. Then I ran CQLSH on the other two nodes. Node 2 got the change right away, while node 3 still has the old value, even after 2 hours. As soon as I restarted the Cassandra service on node 3, the new value showed up. I have encountered this issue a couple of times. I searched this user group, and saw two similar issues ( http://cassandra-user-incubator-apache-org.3065146.n2.nabble.com/default-time-to-live-tt7590662.html#a7590677 <http://cassandra-user-incubator-apache-org.3065146.n2.nabble.com/default-time-to-live-tt7590662.html#a7590677> and http://cassandra-user-incubator-apache-org.3065146.n2.nabble.com/Notes-and-questions-from-performing-a-large-delete-tt7591649.html <http://cassandra-user-incubator-apache-org.3065146.n2.nabble.com/Notes-and-questions-from-performing-a-large-delete-tt7591649.html> ). However, I did not find a clear answer. Is this a bug? Is there something wrong with my settings? I don't feel this should be the normal behavior. -- View this message in context: http://cassandra-user-incubator-apache-org.3065146.n2.nabble.com/CQL-Alter-table-does-not-propagate-to-all-nodes-tp7597211.html Sent from the cassandra-u...@incubator.apache.org mailing list archive at Nabble.com.
