We have 4 node Cassandra cluster running 2.0.12.
Recently, I upgraded one node to 2.1.13, following all the steps given in datastax upgrade procedure.
After the upgrade, I created a new table named 'testing' in existing keyspace using cqlsh on the upgraded node. It gives me the following error:
However, when i created testing table in a nonupgraded node no warning occurs. But, immediately I checked all the other nodes to confirm creation of testing table. Except upgraded node, create table has propagated to other nodes. Only after 15 minutes, create table gets propagated to upgraded node instead of immediate propagation.
nodetool describecluster shows different schema version for the upgraded node
- I ran nodetool resetlocalschema on the upgraded node and it is of no use.
- I also deleted system.schema* directories from data directory location and restarted the node. Still, no use.
- After that, I deleted system.schema* directories and copied the system.schema* directories from non-upgraded node and restarted. Still the issue persists.
(PS: I am using hector jar to connect cassandra. When I set write consistency to ALL, hector throws UnAvailable Exception for that upgraded node.)
Edit 1: In addition I got the following exception in log. error: No such attribute: EndpointToHostId cassandra 2.1