Lobo2008 commented on code in PR #2607:
URL: https://github.com/apache/uniffle/pull/2607#discussion_r2343491009


##########
client-mr/core/src/main/java/org/apache/hadoop/mapred/RssMapOutputCollector.java:
##########
@@ -78,12 +78,31 @@ public void init(Context context) throws IOException, 
ClassNotFoundException {
       throw new IOException("Invalid  sort memory use threshold : " + 
sortThreshold);
     }
 
-    // combiner
-    final Counters.Counter combineInputCounter =
-        reporter.getCounter(TaskCounter.COMBINE_INPUT_RECORDS);
-    combinerRunner =
-        Task.CombinerRunner.create(
-            mrJobConf, mapTask.getTaskID(), combineInputCounter, reporter, 
null);
+    boolean enableCombiner =
+        RssMRUtils.getBoolean(
+            rssJobConf,
+            RssMRConfig.RSS_CLIENT_COMBINER_ENABLE,
+            RssMRConfig.RSS_CLIENT_COMBINER_ENABLE_DEFAULT);
+
+    combinerRunner = null;
+    if (enableCombiner) {
+      try {
+        if (mrJobConf.getCombinerClass() != null ||
+                mrJobConf.get("mapreduce.job.combine.class") != null ||
+                mrJobConf.get("mapreduce.combine.class") != null ||
+                mrJobConf.get("mapred.combiner.class") != null) {
+
+          final Counters.Counter combineInputCounter =
+                  reporter.getCounter(TaskCounter.COMBINE_INPUT_RECORDS);
+          combinerRunner = Task.CombinerRunner.create(
+                  mrJobConf, mapTask.getTaskID(), combineInputCounter, 
reporter, null);
+          LOG.info("Map-stage combiner enabled. Warning: This may cause GC 
issues in large jobs. " +
+                  "Considering disable it by setting {}=false", 
RssMRConfig.RSS_CLIENT_COMBINER_ENABLE);
+        }
+      } catch (Exception e) {
+        LOG.debug("Get CombinerClass failed", e);

Review Comment:
   upgraded combiner init failure to Log.error



-- 
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: issues-unsubscr...@uniffle.apache.org

For queries about this service, please contact Infrastructure at:
us...@infra.apache.org


---------------------------------------------------------------------
To unsubscribe, e-mail: issues-unsubscr...@uniffle.apache.org
For additional commands, e-mail: issues-h...@uniffle.apache.org

Reply via email to