saintstack commented on a change in pull request #2413:
URL: https://github.com/apache/hbase/pull/2413#discussion_r490414833
##########
File path:
hbase-server/src/main/java/org/apache/hadoop/hbase/replication/regionserver/ReplicationSourceInterface.java
##########
@@ -191,4 +194,17 @@ void addHFileRefs(TableName tableName, byte[] family,
List<Pair<Path, Path>> pai
default boolean isRecovered() {
return false;
}
+
+ /**
+ * ReplicationSources keep their replication queues and queue state out in
the replication
+ * store which is external to the current process -- usually zookeeper -- in
case of crash so
+ * another process can pick up where the crashed process left off. A few
specialized replication
+ * sources intentionally avoid keeping state in the replication store
because they reset and
+ * start over after a crash so need of recovery. One such is the replication
of the hbase:meta
+ * table. Its ReplicationSource will return false when this method is called.
+ * @return True if this source persists its queues and queue state to the
replication store.
+ */
+ default boolean isQueueStored() {
Review comment:
That is better.
Let me change. Also need to do another pass on peerid vs queueid... They get
intermixed....Let me try and clean up some.
----------------------------------------------------------------
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
For queries about this service, please contact Infrastructure at:
[email protected]
