[HBASE-9870] HFileDataBlockEncoderImpl#diskToCacheFormat uses wrong format

• Type: Bug
• Status: Closed
• Priority: Major
• Resolution: Fixed
• Fix Version/s: 0.98.0, 0.96.1
• Hadoop Flags: Reviewed

Release Note

The patch makes sure that data block encoding in cache is always the same as that on disk. It deprecates the column family attribute ENCODE_ON_DISK since it doesn't apply any more. HColumnDescriptor#setEncodeOnDisk(boolean) is still there for compatibility issue. However, it doesn't do anything now. HColumnDescriptor#getDataBlockEncodingOnDisk() is also deprecated because the data block encoding on disk is always the same as that in cache for a given store file.

• Assignee: Jimmy Xiang
• Reporter: Jimmy Xiang
• Created: 31/Oct/13 21:57
• Updated: 16/Dec/13 18:46
• Resolved: 13/Nov/13 17:36

Description

In this method, we have

    if (block.getBlockType() == BlockType.ENCODED_DATA) {
      if (block.getDataBlockEncodingId() == onDisk.getId()) {
        // The block is already in the desired in-cache encoding.
        return block;
      }

This assumes onDisk encoding is the same as that of inCache. This is not true when we change the encoding of a CF. This could be one of the reasons I got data loss with online encoding change?

If I make sure onDisk == inCache all the time, my ITBLL with online encoding change worked once for me.

Issue Links

• is depended upon by: HBASE-9757 Reenable fast region move in SlowDeterministicMonkey