saintstack commented on a change in pull request #1786:
URL: https://github.com/apache/hbase/pull/1786#discussion_r434064262
##########
File path:
hbase-server/src/main/java/org/apache/hadoop/hbase/master/normalizer/SimpleRegionNormalizer.java
##########
@@ -18,126 +17,474 @@
*/
package org.apache.hadoop.hbase.master.normalizer;
+import java.io.IOException;
+import java.time.Instant;
+import java.time.Period;
import java.util.ArrayList;
import java.util.Collections;
-import java.util.Comparator;
import java.util.List;
-
-import org.apache.hadoop.hbase.HBaseIOException;
+import java.util.Objects;
+import java.util.function.BooleanSupplier;
+import org.apache.hadoop.conf.Configuration;
+import org.apache.hadoop.hbase.HBaseInterfaceAudience;
+import org.apache.hadoop.hbase.RegionMetrics;
+import org.apache.hadoop.hbase.ServerName;
+import org.apache.hadoop.hbase.Size;
import org.apache.hadoop.hbase.TableName;
+import org.apache.hadoop.hbase.client.MasterSwitchType;
import org.apache.hadoop.hbase.client.RegionInfo;
+import org.apache.hadoop.hbase.client.TableDescriptor;
+import org.apache.hadoop.hbase.master.MasterServices;
+import org.apache.hadoop.hbase.master.RegionState;
+import org.apache.hadoop.hbase.master.assignment.RegionStates;
import org.apache.hadoop.hbase.master.normalizer.NormalizationPlan.PlanType;
+import org.apache.hadoop.hbase.util.EnvironmentEdgeManager;
import org.apache.yetus.audience.InterfaceAudience;
import org.slf4j.Logger;
import org.slf4j.LoggerFactory;
+import
org.apache.hbase.thirdparty.org.apache.commons.collections4.CollectionUtils;
/**
* Simple implementation of region normalizer. Logic in use:
* <ol>
- * <li>Get all regions of a given table
- * <li>Get avg size S of each region (by total size of store files reported in
RegionMetrics)
- * <li>Seek every single region one by one. If a region R0 is bigger than S *
2, it is kindly
- * requested to split. Thereon evaluate the next region R1
- * <li>Otherwise, if R0 + R1 is smaller than S, R0 and R1 are kindly requested
to merge. Thereon
- * evaluate the next region R2
- * <li>Otherwise, R1 is evaluated
+ * <li>Get all regions of a given table</li>
+ * <li>Get avg size S of the regions in the table (by total size of store
files reported in
+ * RegionMetrics)</li>
+ * <li>For each region R0, if R0 is bigger than S * 2, it is kindly
requested to split.</li>
+ * <li>Otherwise, for the next region in the chain R1, if R0 + R1 is smaller
then S, R0 and R1
+ * are kindly requested to merge.</li>
Review comment:
I said it already that this strikes me as too aggressive. In a follow-up
we should make this sloppier? S * (2*) + (2*slop)?
##########
File path:
hbase-server/src/main/java/org/apache/hadoop/hbase/master/normalizer/SimpleRegionNormalizer.java
##########
@@ -18,126 +17,436 @@
*/
package org.apache.hadoop.hbase.master.normalizer;
+import java.io.IOException;
+import java.time.Instant;
+import java.time.Period;
import java.util.ArrayList;
import java.util.Collections;
import java.util.Comparator;
import java.util.List;
-
-import org.apache.hadoop.hbase.HBaseIOException;
+import java.util.Objects;
+import java.util.function.BooleanSupplier;
+import org.apache.hadoop.conf.Configuration;
+import org.apache.hadoop.hbase.HBaseInterfaceAudience;
+import org.apache.hadoop.hbase.RegionMetrics;
+import org.apache.hadoop.hbase.ServerName;
+import org.apache.hadoop.hbase.Size;
import org.apache.hadoop.hbase.TableName;
+import org.apache.hadoop.hbase.client.MasterSwitchType;
import org.apache.hadoop.hbase.client.RegionInfo;
+import org.apache.hadoop.hbase.client.TableDescriptor;
+import org.apache.hadoop.hbase.master.MasterServices;
+import org.apache.hadoop.hbase.master.RegionState;
+import org.apache.hadoop.hbase.master.assignment.RegionStates;
import org.apache.hadoop.hbase.master.normalizer.NormalizationPlan.PlanType;
+import org.apache.hadoop.hbase.util.EnvironmentEdgeManager;
import org.apache.yetus.audience.InterfaceAudience;
import org.slf4j.Logger;
import org.slf4j.LoggerFactory;
+import
org.apache.hbase.thirdparty.org.apache.commons.collections4.CollectionUtils;
/**
* Simple implementation of region normalizer. Logic in use:
* <ol>
- * <li>Get all regions of a given table
- * <li>Get avg size S of each region (by total size of store files reported in
RegionMetrics)
- * <li>Seek every single region one by one. If a region R0 is bigger than S *
2, it is kindly
- * requested to split. Thereon evaluate the next region R1
- * <li>Otherwise, if R0 + R1 is smaller than S, R0 and R1 are kindly requested
to merge. Thereon
- * evaluate the next region R2
- * <li>Otherwise, R1 is evaluated
+ * <li>Get all regions of a given table</li>
+ * <li>Get avg size S of the regions in the table (by total size of store
files reported in
+ * RegionMetrics)</li>
+ * <li>For each region R0, if R0 is bigger than S * 2, it is kindly
requested to split.</li>
+ * <li>Otherwise, for the next region in the chain R1, if R0 + R1 is smaller
then S, R0 and R1
+ * are kindly requested to merge.</li>
+ * </ol>
+ * <p>
+ * The following parameters are configurable:
+ * <ol>
+ * <li>Whether to split a region as part of normalization. Configuration:
+ * {@value SPLIT_ENABLED_KEY}, default: {@value
DEFAULT_SPLIT_ENABLED}.</li>
+ * <li>Whether to merge a region as part of normalization. Configuration:
+ * {@value MERGE_ENABLED_KEY}, default: {@value
DEFAULT_MERGE_ENABLED}.</li>
+ * <li>The minimum number of regions in a table to consider it for
normalization. Configuration:
+ * {@value MIN_REGION_COUNT_KEY}, default: {@value
DEFAULT_MIN_REGION_COUNT}.</li>
+ * <li>The minimum age for a region to be considered for a merge, in days.
Configuration:
+ * {@value MERGE_MIN_REGION_AGE_DAYS_KEY}, default:
+ * {@value DEFAULT_MERGE_MIN_REGION_AGE_DAYS}.</li>
+ * <li>The minimum size for a region to be considered for a merge, in whole
MBs. Configuration:
+ * {@value MERGE_MIN_REGION_SIZE_MB_KEY}, default:
+ * {@value DEFAULT_MERGE_MIN_REGION_SIZE_MB}.</li>
* </ol>
* <p>
- * Region sizes are coarse and approximate on the order of megabytes.
Additionally, "empty" regions
- * (less than 1MB, with the previous note) are not merged away. This is by
design to prevent
- * normalization from undoing the pre-splitting of a table.
+ * To see detailed logging of the application of these configuration values,
set the log level for
+ * this class to `TRACE`.
*/
-@InterfaceAudience.Private
-public class SimpleRegionNormalizer extends AbstractRegionNormalizer {
-
+@InterfaceAudience.LimitedPrivate(HBaseInterfaceAudience.CONFIG)
+public class SimpleRegionNormalizer implements RegionNormalizer {
Review comment:
So we kept the Interface?
----------------------------------------------------------------
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:
us...@infra.apache.org
