jcamachor commented on a change in pull request #794: HIVE-21884
URL: https://github.com/apache/hive/pull/794#discussion_r335749992
 
 

 ##########
 File path: 
ql/src/java/org/apache/hadoop/hive/ql/schq/ScheduledQueryExecutionService.java
 ##########
 @@ -0,0 +1,159 @@
+package org.apache.hadoop.hive.ql.schq;
+
+import java.io.Closeable;
+import java.io.IOException;
+import java.util.concurrent.ExecutorService;
+import java.util.concurrent.Executors;
+import java.util.concurrent.TimeUnit;
+
+import org.apache.hadoop.hive.conf.HiveConf;
+import org.apache.hadoop.hive.metastore.api.QueryState;
+import org.apache.hadoop.hive.metastore.api.ScheduledQueryPollResponse;
+import org.apache.hadoop.hive.metastore.api.ScheduledQueryProgressInfo;
+import org.apache.hadoop.hive.ql.DriverFactory;
+import org.apache.hadoop.hive.ql.IDriver;
+import org.apache.hadoop.hive.ql.processors.CommandProcessorResponse;
+import org.apache.hadoop.hive.ql.session.SessionState;
+import org.slf4j.Logger;
+import org.slf4j.LoggerFactory;
+
+import com.google.common.annotations.VisibleForTesting;
+import com.google.common.util.concurrent.ThreadFactoryBuilder;
+
+public class ScheduledQueryExecutionService implements Closeable {
+
+  private static final Logger LOG = 
LoggerFactory.getLogger(ScheduledQueryExecutionService.class);
+
+  private ScheduledQueryExecutionContext context;
+  private ScheduledQueryExecutor worker;
+
+  public static ScheduledQueryExecutionService 
startScheduledQueryExecutorService(HiveConf conf) {
+    MetastoreBasedScheduledQueryService qService = new 
MetastoreBasedScheduledQueryService(conf);
+    ExecutorService executor =
+        Executors.newCachedThreadPool(
+            new 
ThreadFactoryBuilder().setDaemon(true).setNameFormat("Scheduled Query Thread 
%d").build());
+    ScheduledQueryExecutionContext ctx = new 
ScheduledQueryExecutionContext(executor, conf, qService);
+    return new ScheduledQueryExecutionService(ctx);
+  }
+
+  public ScheduledQueryExecutionService(ScheduledQueryExecutionContext ctx) {
+    context = ctx;
+    ctx.executor.submit(worker = new ScheduledQueryExecutor());
+    ctx.executor.submit(new ProgressReporter());
+  }
+
+  class ScheduledQueryExecutor implements Runnable {
+
+    private ScheduledQueryProgressInfo info;
+
+    @Override
+    public void run() {
+      while (true) {
+        ScheduledQueryPollResponse q = 
context.schedulerService.scheduledQueryPoll();
+        if (q.isSetExecutionId()) {
+          try{
+            processQuery(q);
+          } catch (Throwable t) {
+            LOG.error("Unexpected exception during scheduled query 
processing", t);
+          }
+        } else {
+          try {
+            Thread.sleep(context.getIdleSleepTime());
+          } catch (InterruptedException e) {
+            Thread.currentThread().interrupt();
+            LOG.info("interrupted");
+            break;
+          }
+        }
+      }
+    }
+
+    QueryState lastState;
+    public synchronized void reportQueryProgress() {
+      if (info != null) {
+        if(lastState == QueryState.FINISHED || lastState == 
QueryState.ERRORED) {
+          return;
+        }
+        LOG.info("Reporting query progress of {} as {} err:{}", 
info.getScheduledExecutionId(), info.getState(),
+            info.getErrorMessage());
+        context.schedulerService.scheduledQueryProgress(info);
+        lastState=info.getState();
+      }
+    }
+
+    private void processQuery(ScheduledQueryPollResponse q) {
 
 Review comment:
   We need to fix 2. One of the common use cases that we were discussing for 
scheduled queries was incrementally refreshing contents in a table in a 
transactional way, e.g., pulling data from Kafka. Currently, if query can be 
assumed dead and then come back to life, we may basically end up inserting data 
more than once in target object. You could just grab an exclusive execution 
lock on the scheduled query identifier to prevent this from happening, we 
already have mechanisms to do that (fwiw we have similar logic for incremental 
materialized view rebuild).

----------------------------------------------------------------
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]


With regards,
Apache Git Services

---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]

Reply via email to