[HBASE-13199] Some small improvements on canary tool

• Type: Improvement
• Status: Closed
• Priority: Major
• Resolution: Fixed
• Affects Version/s: 2.0.0
• Fix Version/s: 2.0.0
• Assignee: Shaohui Liu
• Reporter: Shaohui Liu
• Created: 11/Mar/15 10:29
• Updated: 23/Jun/22 20:42
• Resolved: 20/Mar/15 20:26

Description

Improvements

• Make the sniff of region and regionserver parallel to support large cluster with 10000+ region and 500+ regionservers using thread pool.
• Set cacheblock to false in get and scan to avoid influence to block cache.
• Add FirstKeyOnlyFilter to get and scan to avoid read and translate too many data from HBase. There may be many column under a column family in a flat-wide table.
• Select the region randomly when sniffing a regionserver.
• Make the sink class of canary configurable

stack

Suggestions are welcomed. Thanks~

Another question is that why to check each column family with separate requests when sniffing a region? Can we just check a column family of a region?