[HDFS-15745] Make DataNodePeerMetrics#LOW_THRESHOLD_MS and MIN_OUTLIER_DETECTION_NODES configurable

• Type: Improvement
• Status: Resolved
• Priority: Major
• Resolution: Fixed
• Affects Version/s: 3.4.0, 3.3.5
• Fix Version/s: 3.4.0, 3.3.5
• Component/s: datanode, metrics
• Labels: pull-request-available
• Target Version/s: 3.4.0, 3.3.5
• Hadoop Flags: Reviewed
• Assignee: Haibin Huang
• Reporter: Haibin Huang
• Created: 22/Dec/20 09:02
• Updated: 11/Feb/24 07:58
• Resolved: 01/Jan/21 06:14

Description

When i enable DataNodePeerMetrics to find slow slow peer in cluster, i found there is a lot of slow peer but ReportingNodes's averageDelay is very low, and these slow peer node are normal. I think the reason of why generating so many slow peer is that  the value of DataNodePeerMetrics#LOW_THRESHOLD_MS is too small (only 5ms) and it is not configurable. The default value of slow io warning log threshold is 300ms, i.e. DFSConfigKeys.DFS_DATANODE_SLOW_IO_WARNING_THRESHOLD_DEFAULT = 300, so DataNodePeerMetrics#LOW_THRESHOLD_MS should not be less than 300ms, otherwise namenode will get a lot of invalid slow peer information.

Issue Links

• links to: GitHub Pull Request #3992