HBASE-22803 Modify config value range to enable turning off of the hbck chore (#466)

Signed-off-by: Guanghao Zhang <[email protected]>

Author: the-sakthi

hbase-server/src/main/java/org/apache/hadoop/hbase/master/HbckChore.java

@@ -101,14 +101,26 @@ public class HbckChore extends ScheduledChore {
   private volatile long checkingStartTimestamp = 0;
   private volatile long checkingEndTimestamp = 0;
 
+  private boolean disabled = false;
+
   public HbckChore(MasterServices master) {
     super("HbckChore-", master,
         master.getConfiguration().getInt(HBCK_CHORE_INTERVAL, DEFAULT_HBCK_CHORE_INTERVAL));
     this.master = master;
+    int interval =
+        master.getConfiguration().getInt(HBCK_CHORE_INTERVAL, DEFAULT_HBCK_CHORE_INTERVAL);
+    if (interval <= 0) {
+      LOG.warn(HBCK_CHORE_INTERVAL + " is <=0 hence disabling hbck chore");
+      disableChore();
+    }
   }
 
   @Override
   protected synchronized void chore() {
+    if (isDisabled() || isRunning()) {
+      LOG.warn("hbckChore is either disabled or is already running. Can't run the chore");
+      return;
+    }
     running = true;
     regionInfoMap.clear();
     disabledTableRegions.clear();
@@ -127,6 +139,29 @@ protected synchronized void chore() {
     running = false;
   }
 
+  // This function does the sanity checks of making sure the chore is not run when it is
+  // disabled or when it's already running. It returns whether the chore was actually run or not.
+  protected boolean runChore() {
+    if (isDisabled() || isRunning()) {
+      if (isDisabled()) {
+        LOG.warn("hbck chore is disabled! Set " + HBCK_CHORE_INTERVAL + " > 0 to enable it.");
+      } else {
+        LOG.warn("hbck chore already running. Can't run till it finishes.");
+      }
+      return false;
+    }
+    chore();
+    return true;
+  }
+
+  private void disableChore() {
+    this.disabled = true;
+  }
+
+  public boolean isDisabled() {
+    return this.disabled;
+  }
+
   private void saveCheckResultToSnapshot() {
     // Need synchronized here, as this "snapshot" may be access by web ui.
     rwLock.writeLock().lock();

hbase-server/src/main/java/org/apache/hadoop/hbase/master/MasterRpcServices.java

@@ -2324,11 +2324,7 @@ public RunHbckChoreResponse runHbckChore(RpcController c, RunHbckChoreRequest re
     rpcPreCheck("runHbckChore");
     LOG.info("{} request HBCK chore to run", master.getClientIdAuditPrefix());
     HbckChore hbckChore = master.getHbckChore();
-    boolean ran = false;
-    if (!hbckChore.isRunning()) {
-      hbckChore.chore();
-      ran = true;
-    }
+    boolean ran = hbckChore.runChore();
     return RunHbckChoreResponse.newBuilder().setRan(ran).build();
   }
 

hbase-server/src/main/resources/hbase-webapps/master/hbck.jsp

@@ -81,7 +81,11 @@
     <div class="page-header">
       <h1>HBCK Chore Report</h1>
       <p>
+        <% if (hbckChore.isDisabled()) { %>
+        <span>HBCK chore is currently disabled. Set hbase.master.hbck.chore.interval > 0 in the config & do a rolling-restart to enable it.</span>
+        <% } else { %>
         <span>Checking started at <%= iso8601start %> and generated report at <%= iso8601end %>. Execute 'hbck_chore_run' in hbase shell to generate a new sub-report.</span>
+        <% } %>
       </p>
     </div>
   </div>

hbase-server/src/test/java/org/apache/hadoop/hbase/master/assignment/TestHbckChore.java

@@ -198,4 +198,22 @@ public void testOrphanRegionsOnFS() throws Exception {
     hbckChore.choreForTesting();
     assertEquals(0, hbckChore.getOrphanRegionsOnFS().size());
   }
+
+  @Test
+  public void testChoreDisable() {
+    // The way to disable to chore is to set hbase.master.hbck.chore.interval <= 0
+    // When the interval is > 0, the chore should run.
+    long lastRunTime = hbckChore.getCheckingEndTimestamp();
+    hbckChore.choreForTesting();
+    boolean ran = lastRunTime != hbckChore.getCheckingEndTimestamp();
+    assertTrue(ran);
+
+    // When the interval <= 0, the chore shouldn't run
+    master.getConfiguration().setInt("hbase.master.hbck.chore.interval", 0);
+    HbckChore hbckChoreWithChangedConf = new HbckChore(master);
+    lastRunTime = hbckChoreWithChangedConf.getCheckingEndTimestamp();
+    hbckChoreWithChangedConf.choreForTesting();
+    ran = lastRunTime != hbckChoreWithChangedConf.getCheckingEndTimestamp();
+    assertFalse(ran);
+  }
 }