[HBASE-7495] parallel seek in StoreScanner

• Type: Improvement
• Status: Closed
• Priority: Major
• Resolution: Fixed
• Affects Version/s: 0.94.3, 0.95.2
• Fix Version/s: 0.95.0
• Component/s: Performance, Scanners
• Hadoop Flags: Reviewed

Release Note

This JIRA adds the ability of doing parallel seek in StoreScanner. It is off by default.
The following config parameter turns on this feature:
hbase.storescanner.parallel.seek.enable

hbase.storescanner.parallel.seek.threads controls the number of threads in thread pool which serves parallel seeking. Default is 10 threads.

• Assignee: Liang Xie
• Reporter: Liang Xie
• Created: 04/Jan/13 11:47
• Updated: 23/Sep/13 18:30
• Resolved: 19/Feb/13 21:17

Description

seems there's a potential improvable space before doing scanner.next:

StoreScanner.java

    if (explicitColumnQuery && lazySeekEnabledGlobally) {
      for (KeyValueScanner scanner : scanners) {
        scanner.requestSeek(matcher.getStartKey(), false, true);
      }
    } else {
      for (KeyValueScanner scanner : scanners) {
        scanner.seek(matcher.getStartKey());
      }
    }

we can do scanner.requestSeek or scanner.seek in parallel, instead of current serialization, to reduce latency for special case.

Any ideas on it ? I'll have a try if the comments/suggestions are positive：）