[incubator-uniffle] branch master updated: [Log Improvment] Add more detailed debug info for MR client (#84)

[ro...@apache.org]

This is an automated email from the ASF dual-hosted git repository.

roryqi pushed a commit to branch master
in repository https://gitbox.apache.org/repos/asf/incubator-uniffle.git


The following commit(s) were added to refs/heads/master by this push:
     new 0a866f6  [Log Improvment] Add more detailed debug info for MR client (#84)
0a866f6 is described below

commit 0a866f6b9ed0f69eac34227d884bb89c72437172
Author: frankliee <fr...@tencent.com>
AuthorDate: Tue Jul 26 20:27:15 2022 +0800

    [Log Improvment] Add more detailed debug info for MR client (#84)
    
    ### What changes were proposed in this pull request?
    Add more detailed debug info for MR client
    
    ### Why are the changes needed?
    MR debug is hardy.
    
    ### Does this PR introduce _any_ user-facing change?
    No.
    
    ### How was this patch tested?
    No need.
---
 .../hadoop/mapreduce/task/reduce/RssEventFetcher.java       | 13 +++++++++++++
 1 file changed, 13 insertions(+)

diff --git a/client-mr/src/main/java/org/apache/hadoop/mapreduce/task/reduce/RssEventFetcher.java b/client-mr/src/main/java/org/apache/hadoop/mapreduce/task/reduce/RssEventFetcher.java
index 1619678..dcc1b4b 100644
--- a/client-mr/src/main/java/org/apache/hadoop/mapreduce/task/reduce/RssEventFetcher.java
+++ b/client-mr/src/main/java/org/apache/hadoop/mapreduce/task/reduce/RssEventFetcher.java
@@ -85,16 +85,29 @@ public class RssEventFetcher<K,V> {
           if (mapIndex < totalMapsCount) {
             mapIndexBitmap.addLong(mapIndex);
           } else {
+            LOG.error(taskAttemptID + " has overflowed mapIndex");
             throw new IllegalStateException(errMsg);
           }
+        } else {
+          LOG.warn(taskAttemptID + " is redundant on index: " + mapIndex);
         }
+      } else {
+        LOG.warn(taskAttemptID + " is successful but cancelled by obsolete event");
       }
     }
     // each map should have only one success attempt
     if (mapIndexBitmap.getLongCardinality() != taskIdBitmap.getLongCardinality()) {
       throw new IllegalStateException(errMsg);
     }
+    if (tipFailedCount != 0) {
+      LOG.warn("There are " + tipFailedCount + " tipFailed tasks");
+    }
     if (taskIdBitmap.getLongCardinality() + tipFailedCount != totalMapsCount) {
+      for (int index = 0; index < totalMapsCount; index++) {
+        if (!mapIndexBitmap.contains(index)) {
+          LOG.error("Fail to fetch " + " map task on index: " + index);
+        }
+      }
       throw new IllegalStateException(errMsg);
     }
     return taskIdBitmap;