brfrn169 commented on a change in pull request #3150:
URL: https://github.com/apache/hbase/pull/3150#discussion_r614858866
##########
File path:
hbase-server/src/main/java/org/apache/hadoop/hbase/master/assignment/SplitTableRegionProcedure.java
##########
@@ -110,6 +111,15 @@ public SplitTableRegionProcedure(final MasterProcedureEnv
env,
// we fail-fast on construction. There it skips the split with just a
warning.
checkOnline(env, regionToSplit);
this.bestSplitRow = splitRow;
+ TableDescriptor tableDescriptor =
env.getMasterServices().getTableDescriptors()
+ .get(getTableName());
+ Configuration conf = env.getMasterConfiguration();
+ if (hasBestSplitRow()) {
Review comment:
> Finally we should get the actual split point back from region server?
No?
No, I don't think so.
Let's see we have a table that has a key prefix restriction where the prefix
length is 2 bytes.
When a user runs split command with specifying a split point `abc` in the
hbase shell, this will break the key prefix restriction if we split the region
by `abc`. So I think we can apply the restriction to the user-specified split
point, and the restriction-applied split point will be `ab`, which won't break
the restriction.
--
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]
