This is an automated email from the ASF dual-hosted git repository.
vjasani pushed a commit to branch branch-2.2
in repository https://gitbox.apache.org/repos/asf/hbase.git
The following commit(s) were added to refs/heads/branch-2.2 by this push:
new 1a4b474 HBASE-24511 Ability to configure timeout between RPC retry to
RS from master (#1861)
1a4b474 is described below
commit 1a4b474c2b057bdcc48a4bb2eeb17ca77b111bc8
Author: sguggilam <sandeepbit...@gmail.com>
AuthorDate: Wed Jun 10 00:23:08 2020 -0700
HBASE-24511 Ability to configure timeout between RPC retry to RS from
master (#1861)
Signed-off-by: Viraj Jasani <vjas...@apache.org>
---
.../master/procedure/RSProcedureDispatcher.java | 20 ++++++++++--
.../TestRegionServerReportForDuty.java | 38 +++++++++++++++++++++-
2 files changed, 55 insertions(+), 3 deletions(-)
diff --git
a/hbase-server/src/main/java/org/apache/hadoop/hbase/master/procedure/RSProcedureDispatcher.java
b/hbase-server/src/main/java/org/apache/hadoop/hbase/master/procedure/RSProcedureDispatcher.java
index b469cb8..ae60848 100644
---
a/hbase-server/src/main/java/org/apache/hadoop/hbase/master/procedure/RSProcedureDispatcher.java
+++
b/hbase-server/src/main/java/org/apache/hadoop/hbase/master/procedure/RSProcedureDispatcher.java
@@ -211,12 +211,19 @@ public class RSProcedureDispatcher
private int numberOfAttemptsSoFar = 0;
private long maxWaitTime = -1;
+ private final long rsRpcRetryInterval;
+ private static final String RS_RPC_RETRY_INTERVAL_CONF_KEY =
+ "hbase.regionserver.rpc.retry.interval";
+ private static final int DEFAULT_RS_RPC_RETRY_INTERVAL = 100;
+
private ExecuteProceduresRequest.Builder request = null;
public ExecuteProceduresRemoteCall(final ServerName serverName,
final Set<RemoteProcedure> remoteProcedures) {
this.serverName = serverName;
this.remoteProcedures = remoteProcedures;
+ this.rsRpcRetryInterval =
master.getConfiguration().getLong(RS_RPC_RETRY_INTERVAL_CONF_KEY,
+ DEFAULT_RS_RPC_RETRY_INTERVAL);
}
private AdminService.BlockingInterface getRsAdmin() throws IOException {
@@ -241,7 +248,8 @@ public class RSProcedureDispatcher
LOG.warn("waiting a little before trying on the same server={}," +
" try={}, can wait up to {}ms", serverName, numberOfAttemptsSoFar,
remainingTime);
numberOfAttemptsSoFar++;
- submitTask(this, 100, TimeUnit.MILLISECONDS);
+ // Retry every rsRpcRetryInterval millis up to maximum wait time.
+ submitTask(this, rsRpcRetryInterval, TimeUnit.MILLISECONDS);
return true;
}
LOG.warn("server {} is not up for a while; try a new one", serverName);
@@ -283,7 +291,15 @@ public class RSProcedureDispatcher
e.toString(), numberOfAttemptsSoFar);
}
numberOfAttemptsSoFar++;
- submitTask(this, 100, TimeUnit.MILLISECONDS);
+ // Add some backoff here as the attempts rise otherwise if a stuck
condition, will fill logs
+ // with failed attempts. None of our backoff classes -- RetryCounter or
ClientBackoffPolicy
+ // -- fit here nicely so just do something simple; increment by
rsRpcRetryInterval millis *
+ // retry^2 on each try
+ // up to max of 10 seconds (don't want to back off too much in case of
situation change).
+ submitTask(this,
+ Math.min(rsRpcRetryInterval * (this.numberOfAttemptsSoFar *
this.numberOfAttemptsSoFar),
+ 10 * 1000),
+ TimeUnit.MILLISECONDS);
return true;
}
diff --git
a/hbase-server/src/test/java/org/apache/hadoop/hbase/regionserver/TestRegionServerReportForDuty.java
b/hbase-server/src/test/java/org/apache/hadoop/hbase/regionserver/TestRegionServerReportForDuty.java
index aaf2d2e..2821e36 100644
---
a/hbase-server/src/test/java/org/apache/hadoop/hbase/regionserver/TestRegionServerReportForDuty.java
+++
b/hbase-server/src/test/java/org/apache/hadoop/hbase/regionserver/TestRegionServerReportForDuty.java
@@ -22,7 +22,8 @@ import static org.junit.Assert.assertTrue;
import java.io.IOException;
import java.io.StringWriter;
-
+import java.util.concurrent.ScheduledThreadPoolExecutor;
+import java.util.concurrent.TimeUnit;
import org.apache.commons.lang3.StringUtils;
import org.apache.hadoop.conf.Configuration;
import org.apache.hadoop.hbase.HBaseClassTestRule;
@@ -38,6 +39,7 @@ import org.apache.hadoop.hbase.master.ServerManager;
import org.apache.hadoop.hbase.testclassification.MediumTests;
import org.apache.hadoop.hbase.util.JVMClusterUtil.MasterThread;
import org.apache.hadoop.hbase.util.JVMClusterUtil.RegionServerThread;
+import org.apache.hadoop.hbase.util.Threads;
import org.apache.log4j.Appender;
import org.apache.log4j.Layout;
import org.apache.log4j.PatternLayout;
@@ -221,6 +223,40 @@ public class TestRegionServerReportForDuty {
tablesOnMaster? 3: 2);
}
+
+ /**
+ * Tests region sever reportForDuty with RS RPC retry
+ */
+ @Test
+ public void testReportForDutyWithRSRpcRetry() throws Exception {
+ ScheduledThreadPoolExecutor scheduledThreadPoolExecutor =
+ new ScheduledThreadPoolExecutor(1,
Threads.newDaemonThreadFactory("RSDelayedStart"));
+
+ // Start a master and wait for it to become the active/primary master.
+ // Use a random unique port
+ cluster.getConfiguration().setInt(HConstants.MASTER_PORT,
HBaseTestingUtility.randomFreePort());
+ // Override the default RS RPC retry interval of 100ms to 300ms
+
cluster.getConfiguration().setLong("hbase.regionserver.rpc.retry.interval",
300);
+ // master has a rs. defaultMinToStart = 2
+ boolean tablesOnMaster =
LoadBalancer.isTablesOnMaster(testUtil.getConfiguration());
+
cluster.getConfiguration().setInt(ServerManager.WAIT_ON_REGIONSERVERS_MINTOSTART,
+ tablesOnMaster ? 2 : 1);
+
cluster.getConfiguration().setInt(ServerManager.WAIT_ON_REGIONSERVERS_MAXTOSTART,
+ tablesOnMaster ? 2 : 1);
+ master = cluster.addMaster();
+ rs = cluster.addRegionServer();
+ LOG.debug("Starting master: " + master.getMaster().getServerName());
+ master.start();
+ // Delay the RS start so that the meta assignment fails in first attempt
and goes to retry block
+ scheduledThreadPoolExecutor.schedule(new Runnable() {
+ @Override
+ public void run() {
+ rs.start();
+ }
+ }, 1000, TimeUnit.MILLISECONDS);
+
+ waitForClusterOnline(master);
+ }
private void waitForClusterOnline(MasterThread master) throws
InterruptedException {
while (true) {
