Bug#32452559 ADDS A DEBUG STATEMENT TO TESTUPGRADE TEST IN 7.3 and 7.4

VarunNagaraju · VarunNagaraju · commit b54cf1a433fc · 2021-04-30T10:43:19.000+05:30
Adds a debug statement to print the nodeGroup id which bitmask receives
from the testUpgrade test case.

Upgrade/downgrade tests are sporadically crashing when nodeGroup id,
sent from the test case is used as an index(after bit shifting) to
check the status of data nodes. The nodeGroup id from test case always
seems to be zero since the testcases are restarting the first half of
the cluster. So, adding a debug statement in the test case to confirm
that the nodeGroup id is not 0 when it crashes or if it is indeed 0
and the test still crashes, it indicates that the bitmask array is
corrupted.

Reviewed by: Priyanka Sangam &lt;priyanka.sangam@oracle.com&gt;

Change-Id: I5e59cf81acc33908bf89d76de007d89ca74b3bda
diff --git a/storage/ndb/test/ndbapi/testUpgrade.cpp b/storage/ndb/test/ndbapi/testUpgrade.cpp
@@ -1,5 +1,5 @@
 /*
-   Copyright (c) 2008, 2020, Oracle and/or its affiliates. All rights reserved.
+   Copyright (c) 2008, 2021, Oracle and/or its affiliates. All rights reserved.
 
    This program is free software; you can redistribute it and/or modify
    it under the terms of the GNU General Public License, version 2.0,
@@ -758,6 +758,9 @@ runUpgrade_Half(NDBT_Context* ctx, NDBT_Step* step)
       int processId = nodes[i].processId;
       int nodeGroup= nodes[i].nodeGroup;
 
+      if (nodeGroup != 0) {
+        ndberr << "Expected nodeGroup 0, but got " << nodeGroup << endl;
+      }
       if (seen_groups.get(nodeGroup))
       {
         // One node in this node group already down
