[HDFS-16999] Fix wrong use of processFirstBlockReport()

• Type: Bug
• Status: Resolved
• Priority: Major
• Resolution: Fixed
• Affects Version/s: 3.4.0
• Fix Version/s: 3.4.0
• Component/s: block placement
• Labels: pull-request-available
• Target Version/s: 3.4.0
• Hadoop Flags: Reviewed
• Assignee: Shuyan Zhang
• Reporter: Shuyan Zhang
• Created: 05/May/23 02:38
• Updated: 28/Jan/24 05:36
• Resolved: 09/May/23 02:07

Description

`processFirstBlockReport()` is used to process first block report from datanode. It does not calculating `toRemove` list because it believes that there is no metadata about the datanode in the namenode. However, If a datanode is re registered after restarting, its `blockReportCount` will be updated to 0. That is to say, the first block report after a datanode restarts will be processed by `processFirstBlockReport()`.  This is unreasonable because the metadata of the datanode already exists in namenode at this time, and if redundant replica metadata is not removed in time, the blocks with insufficient replicas cannot be reconstruct in time, which increases the risk of missing block. In summary, `processFirstBlockReport()` should only be used when the namenode restarts, not when the datanode restarts. 

Issue Links

• links to: GitHub Pull Request #5622