Can you provide some more information on the request you are sending, if the CF is standard or super,  and if you have row caching turned on for the CF's. Also if you have started with a fresh data directory or migrated data from b3 to rc1.

There is an execution path in ColumnFamilyStore the looks like it could lead to this but it seems unlikely to execute...

- line 1430 in scan calls getColumnFamily() does not check the result for null. Passes that to ColumnFamilyStore.satisfies()
- line 1105 in getColumnFamily() calls cacheRow() if the row cache has capacity, if cacheRow returns null getColumnFamily() returns null  
- line 1082 in cacheRow() will return null if getTopLevelColumns returns null 

But it's hard to see how getTopLevelColumns() returns null 



On 08 Dec, 2010,at 07:58 AM, Wenjun Che <> wrote:

I was loading test data (just 5000 rows) into a column family in beta3 and got the following exception when .  I upgraded to rc1 and it is still happening.  I am running just one node on Linux.

ERROR [ReadStage:6] 2010-12-07 10:34:10,800 (line 90) Fatal exception in thread Thread[ReadStage:6,5,main]
java.lang.RuntimeException: java.lang.NullPointerException
        at org.apache.cassandra.service.IndexScanVerbHandler.doVerb(
        at java.util.concurrent.ThreadPoolExecutor$Worker.runTask(
        at java.util.concurrent.ThreadPoolExecutor$
        at java.lang.Threadrun(
Caused by: java.lang.NullPointerException
        at org.apache.cassandra.db.ColumnFamilyStore.satisfies(
        at org.apache.cassandra.db.ColumnFamilyStore.scan(
        at org.apache.cassandra.service.IndexScanVerbHandler.doVerb(