Fix LCS splitting sstables based on uncompressed length

patch by slebresne; reviewed by jbellis for CASSANDRA-4419

CHANGES.txt

@@ -10,6 +10,7 @@
  * Oversize integer in CQL throws NumberFormatException (CASSANDRA-4291)
  * Set gc_grace on index CF to 0 (CASSANDRA-4314)
  * fix 1.0.x node join to mixed version cluster, other nodes >= 1.1 (CASSANDRA-4195)
+ * Fix LCS splitting sstable base on uncompressed size (CASSANDRA-4419)
 
 
 1.0.10

src/java/org/apache/cassandra/db/compaction/CompactionTask.java

@@ -169,7 +169,7 @@ public class CompactionTask extends AbstractCompactionTask
                         }
                     }
                 }
-                if (!nni.hasNext() || newSSTableSegmentThresholdReached(writer, position))
+                if (!nni.hasNext() || newSSTableSegmentThresholdReached(writer))
                 {
                     SSTableReader toIndex = writer.closeAndOpenReader(getMaxDataAge(toCompact));
                     cachedKeyMap.put(toIndex, cachedKeys);
@@ -229,7 +229,7 @@ public class CompactionTask extends AbstractCompactionTask
     }
 
     //extensibility point for other strategies that may want to limit the upper bounds of the sstable segment size
-    protected boolean newSSTableSegmentThresholdReached(SSTableWriter writer, long position)
+    protected boolean newSSTableSegmentThresholdReached(SSTableWriter writer) throws IOException
     {
         return false;
     }

src/java/org/apache/cassandra/db/compaction/LeveledCompactionTask.java

@@ -69,9 +69,9 @@ public class LeveledCompactionTask extends CompactionTask
     }
 
     @Override
-    protected boolean newSSTableSegmentThresholdReached(SSTableWriter writer, long position)
+    protected boolean newSSTableSegmentThresholdReached(SSTableWriter writer) throws IOException
     {
-        return position > sstableSizeInMB * 1024L * 1024L;
+        return writer.getOnDiskFilePointer() > sstableSizeInMB * 1024L * 1024L;
     }
 
     @Override

src/java/org/apache/cassandra/io/compress/CompressedSequentialWriter.java

@@ -68,6 +68,12 @@ public class CompressedSequentialWriter extends SequentialWriter
         this.sstableMetadataCollector = sstableMetadataCollector;
     }
 
+    @Override
+    public long getOnDiskFilePointer() throws IOException
+    {
+        return out.getFilePointer();
+    }
+
     @Override
     public void sync() throws IOException
     {

src/java/org/apache/cassandra/io/sstable/SSTableWriter.java

@@ -382,6 +382,11 @@ public class SSTableWriter extends SSTable
         return dataFile.getFilePointer();
     }
 
+    public long getOnDiskFilePointer() throws IOException
+    {
+        return dataFile.getOnDiskFilePointer();
+    }
+
     /**
      * Encapsulates writing the index and filter for an SSTable. The state of this object is not valid until it has been closed.
      */

src/java/org/apache/cassandra/io/util/SequentialWriter.java

@@ -211,6 +211,18 @@ public class SequentialWriter extends OutputStream
         return current;
     }
 
+    /**
+     * Return the current file pointer of the underlying on-disk file.
+     * Note that since write works by buffering data, the value of this will increase by buffer
+     * size and not every write to the writer will modify this value.
+     * Furthermore, for compressed files, this value refers to compressed data, while the
+     * writer getFilePointer() refers to uncompressedFile
+     */
+    public long getOnDiskFilePointer() throws IOException
+    {
+        return getFilePointer();
+    }
+
     public long length() throws IOException
     {
         return Math.max(Math.max(current, out.length()), bufferOffset + validBufferBytes);