[HBASE-8157] Allow better control over lease timeouts on a per-scan basis

• Type: New Feature
• Status: Closed
• Priority: Minor
• Resolution: Duplicate
• Assignee: Unassigned
• Reporter: Bryan Beaudreault
• Created: 20/Mar/13 19:08
• Updated: 16/Jun/22 05:53
• Resolved: 27/Mar/15 17:02

Description

Changing the global scanner lease timeout is a heavy-handed solution for long scans on a cluster that may be used by a variety of applications. Two additions would make this easier to manage in this circumstance:

1) Allow overriding hbase.regionserver.lease.period on a per-scan basis.

2) Allow manual reporting of progress, similar to Hadoop's context.progress(). Example usage:

Scan scan = new Scan(startRow, endRow);
scan.setCaching(someVal); // based on what we expect most rows to take for processing time

ResultScanner scanner = table.getScanner(scan);

for (Result r : scanner) {
  //
  // usual processing, the time for which we accounted for in our caching and global lease timeout settings
  //

  if (someCondition) {
    //
    // More time-intensive processing necessary on this record, which is hard to account for in the caching
    //

    scanner.progress();
  }
}

Issue Links

• duplicates: HBASE-13333 Renew Scanner Lease without advancing the RegionScanner