[HBASE-24781] Clean up peer metrics when disabling peer

• Type: Bug
• Status: Resolved
• Priority: Major
• Resolution: Fixed
• Affects Version/s: 2.2.5
• Fix Version/s: 2.6.0, 3.0.0-alpha-4, 2.4.17, 2.5.4
• Component/s: Replication
• Hadoop Flags: Reviewed
• Assignee: Yuta Imazu
• Reporter: leizhang
• Created: 28/Jul/20 07:35
• Updated: 01/Mar/23 12:16
• Resolved: 28/Feb/23 10:43

Description

  Supposed that we have an peer with id 1,  when execute shell cmd disable_peer  '1'  and enable_peer '1', then i can see the SizeOfLogQueue metric of all regionservers   +1 ,  after 10 times disable_peer ops  , it will increase to 11, and  it will never decrease to 1 in fulture .

  I can see the function ReplicationSourceManager.refreshSources(peerId) is called , it will terminate the previous replication source and create a new one.  and found the note  //Do not clear metrics in the bellow code block:

ReplicationSourceInterface toRemove = this.sources.put(peerId, src);
if (toRemove != null) {
  LOG.info("Terminate replication source for " + toRemove.getPeerId());
  // Do not clear metrics
  toRemove.terminate(terminateMessage, null, false);
}

 this cause the wrong number of sizeOfLogQueue, i think it's a sub issue of (HBASE-23231)

Issue Links

• is a child of: HBASE-23231 ReplicationSource do not update metrics after refresh
• links to: GitHub Pull Request #4997