Having run into a recurring compaction problem due to a corrupt sstable (perceived row size was 13 petabytes or something), I sstable2json -x 'd the key and am now trying to re-import the sstable without it. However, I'm running into the following exception:
Importing 2882 keys... java.lang.ClassCastException: org.apache.cassandra.db.ExpiringColumn cannot be cast to org.apache.cassandra.db.SuperColumn at org.apache.cassandra.db.SuperColumnSerializer.serialize(SuperColumn.java:363) at org.apache.cassandra.db.SuperColumnSerializer.serialize(SuperColumn.java:347) at org.apache.cassandra.db.ColumnFamilySerializer.serializeForSSTable(ColumnFamilySerializer.java:88) at org.apache.cassandra.db.ColumnFamilySerializer.serializeWithIndexes(ColumnFamilySerializer.java:107) at org.apache.cassandra.io.sstable.SSTableWriter.append(SSTableWriter.java:147) at org.apache.cassandra.tools.SSTableImport.importUnsorted(SSTableImport.java:290) at org.apache.cassandra.tools.SSTableImport.importJson(SSTableImport.java:252) at org.apache.cassandra.tools.SSTableImport.main(SSTableImport.java:476) ERROR: org.apache.cassandra.db.ExpiringColumn cannot be cast to org.apache.cassandra.db.SuperColumn The CF is a SuperColumnFamily, if that's relevant. 1. What should I do about this problem? 2. (somewhat unrelated) Our usage of this SCF has moved away from requiring "super"-ness. Aside from missing out on the potential for future seconary indexes, are we suffering any sort of operational/performance hit from this classification?