[HBASE-18128] compaction marker could be skipped

• Type: Improvement
• Status: Resolved
• Priority: Major
• Resolution: Not A Problem
• Component/s: Compaction, regionserver
• External issue URL: https://issues.apache.org/jira/browse/HBASE-2231
• External issue ID: HBASE-2231
• Assignee: Jingyun Tian
• Reporter: Jingyun Tian
• Created: 27/May/17 03:51
• Updated: 16/Jun/17 01:50
• Resolved: 16/Jun/17 01:50

Description

The sequence for a compaction are as follows:
1. Compaction writes new files under region/.tmp directory (compaction output)
2. Compaction atomically moves the temporary file under region directory
3. Compaction appends a WAL edit containing the compaction input and output files. Forces sync on WAL.
4. Compaction deletes the input files from the region directory.

But if a flush happened between 3 and 4, then the regionserver crushed. The compaction marker will be skipped when splitting log because the sequence id of compaction marker is smaller than lastFlushedSequenceId.

        if (lastFlushedSequenceId >= entry.getKey().getLogSeqNum()) {
          editsSkipped++;
          continue;
        }