zhengte commented on a change in pull request #8348:
URL: https://github.com/apache/incubator-doris/pull/8348#discussion_r841072346
##########
File path:
fe/fe-core/src/main/java/org/apache/doris/statistics/StatisticsJobScheduler.java
##########
@@ -18,46 +18,243 @@
package org.apache.doris.statistics;
import org.apache.doris.catalog.Catalog;
+import org.apache.doris.catalog.Column;
+import org.apache.doris.catalog.Database;
+import org.apache.doris.catalog.KeysType;
+import org.apache.doris.catalog.OlapTable;
+import org.apache.doris.catalog.Table;
+import org.apache.doris.catalog.Type;
+import org.apache.doris.common.Config;
+import org.apache.doris.common.DdlException;
import org.apache.doris.common.util.MasterDaemon;
+import org.apache.doris.statistics.StatisticsJob.JobState;
import com.google.common.collect.Queues;
-import java.util.ArrayList;
+import org.apache.logging.log4j.LogManager;
+import org.apache.logging.log4j.Logger;
+
+import java.util.Arrays;
+import java.util.Collections;
import java.util.List;
+import java.util.Map;
import java.util.Queue;
+import java.util.Set;
-/*
-Schedule statistics job.
- 1. divide job to multi task
- 2. submit all task to StatisticsTaskScheduler
-Switch job state from pending to scheduling.
+/**
+ * Schedule statistics job.
+ * 1. divide job to multi task
+ * 2. submit all task to StatisticsTaskScheduler
+ * Switch job state from pending to scheduling.
*/
public class StatisticsJobScheduler extends MasterDaemon {
+ private static final Logger LOG =
LogManager.getLogger(StatisticsJobScheduler.class);
- public Queue<StatisticsJob> pendingJobQueue =
Queues.newLinkedBlockingQueue();
+ /**
+ * Different statistics need to be collected for the jobs submitted by
users.
+ * if all statistics be collected at the same time, the cluster may be
overburdened
+ * and normal query services may be affected. Therefore, we put the jobs
into the queue
+ * and schedule them one by one, and finally divide each job to several
subtasks and execute them.
+ */
+ public final Queue<StatisticsJob> pendingJobQueue;
public StatisticsJobScheduler() {
super("Statistics job scheduler", 0);
+ this.pendingJobQueue =
Queues.newLinkedBlockingQueue(Config.cbo_max_statistics_job_num);
}
@Override
protected void runAfterCatalogReady() {
- // TODO
- StatisticsJob pendingJob = pendingJobQueue.peek();
- // step0: check job state again
- // step1: divide statistics job to task
- List<StatisticsTask> statisticsTaskList = divide(pendingJob);
- // step2: submit
-
Catalog.getCurrentCatalog().getStatisticsTaskScheduler().addTasks(statisticsTaskList);
+ StatisticsJob pendingJob = this.pendingJobQueue.peek();
+ if (pendingJob != null) {
+ // step0: check job state again
+ JobState jobState = pendingJob.getJobState();
+ if (jobState == JobState.PENDING) {
+ try {
+ // step1: divide statistics job to tasks
+ List<StatisticsTask> tasks = this.divide(pendingJob);
+ // step2: submit tasks
+
Catalog.getCurrentCatalog().getStatisticsTaskScheduler().addTasks(tasks);
Review comment:
Indeed, the statistics job should be removed after it has been scheduled.
--
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.
To unsubscribe, e-mail: [email protected]
For queries about this service, please contact Infrastructure at:
[email protected]
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
