From 1c5286f4d107f5df6abf021a10f5cf67a9fe4686 Mon Sep 17 00:00:00 2001
From: Dmitry Konstantinov <konstantinov@netcracker.com>
Date: Fri, 22 Aug 2025 14:02:27 +0100
Subject: [PATCH] Fix IndexOutOfBoundsException in sstablemetadata tool when a
 range tombstone is a max clustering value

patch by Dmitry Konstantinov; reviewed by Stefan Miklosovic for CASSANDRA-20855
---
 CHANGES.txt                                   |   1 +
 .../tools/SSTableMetadataViewer.java          |  33 ++++++-----
 .../nb-1-big-CompressionInfo.db               | Bin 0 -> 47 bytes
 .../CASSANDRA_20855/nb-1-big-Data.db          | Bin 0 -> 83 bytes
 .../CASSANDRA_20855/nb-1-big-Digest.crc32     |   1 +
 .../CASSANDRA_20855/nb-1-big-Filter.db        | Bin 0 -> 16 bytes
 .../CASSANDRA_20855/nb-1-big-Index.db         | Bin 0 -> 14 bytes
 .../CASSANDRA_20855/nb-1-big-Statistics.db    | Bin 0 -> 4858 bytes
 .../CASSANDRA_20855/nb-1-big-Summary.db       | Bin 0 -> 74 bytes
 .../CASSANDRA_20855/nb-1-big-TOC.txt          |   8 +++
 .../tools/SSTableMetadataViewerTest.java      |  55 ++++++++++++++++++
 11 files changed, 83 insertions(+), 15 deletions(-)
 create mode 100644 test/data/sstable_metadata/CASSANDRA_20855/nb-1-big-CompressionInfo.db
 create mode 100644 test/data/sstable_metadata/CASSANDRA_20855/nb-1-big-Data.db
 create mode 100644 test/data/sstable_metadata/CASSANDRA_20855/nb-1-big-Digest.crc32
 create mode 100644 test/data/sstable_metadata/CASSANDRA_20855/nb-1-big-Filter.db
 create mode 100644 test/data/sstable_metadata/CASSANDRA_20855/nb-1-big-Index.db
 create mode 100644 test/data/sstable_metadata/CASSANDRA_20855/nb-1-big-Statistics.db
 create mode 100644 test/data/sstable_metadata/CASSANDRA_20855/nb-1-big-Summary.db
 create mode 100644 test/data/sstable_metadata/CASSANDRA_20855/nb-1-big-TOC.txt

diff --git a/CHANGES.txt b/CHANGES.txt
index 6296bb94f6..72696ad020 100644
--- a/CHANGES.txt
+++ b/CHANGES.txt
@@ -1,4 +1,5 @@
 4.0.19
+ * Fix IndexOutOfBoundsException in sstablemetadata tool when a range tombstone is a max clustering value (CASSANDRA-20855)
  * Update Jackson to 2.19.2 (CASSANDRA-20848)
  * Update commons-lang3 to 3.18.0 (CASSANDRA-20849)
  * Add NativeTransportMaxConcurrentConnectionsPerIp to StorageProxyMBean (CASSANDRA-20642)
diff --git a/src/java/org/apache/cassandra/tools/SSTableMetadataViewer.java b/src/java/org/apache/cassandra/tools/SSTableMetadataViewer.java
index bad7b13150..e3c7d76633 100755
--- a/src/java/org/apache/cassandra/tools/SSTableMetadataViewer.java
+++ b/src/java/org/apache/cassandra/tools/SSTableMetadataViewer.java
@@ -355,21 +355,9 @@ public class SSTableMetadataViewer
             if (validation != null && header != null)
                 printMinMaxToken(descriptor, FBUtilities.newPartitioner(descriptor), header.getKeyType());
 
-            if (header != null && header.getClusteringTypes().size() == stats.minClusteringValues.size())
-            {
-                List<AbstractType<?>> clusteringTypes = header.getClusteringTypes();
-                List<ByteBuffer> minClusteringValues = stats.minClusteringValues;
-                List<ByteBuffer> maxClusteringValues = stats.maxClusteringValues;
-                String[] minValues = new String[clusteringTypes.size()];
-                String[] maxValues = new String[clusteringTypes.size()];
-                for (int i = 0; i < clusteringTypes.size(); i++)
-                {
-                    minValues[i] = clusteringTypes.get(i).getString(minClusteringValues.get(i));
-                    maxValues[i] = clusteringTypes.get(i).getString(maxClusteringValues.get(i));
-                }
-                field("minClusteringValues", Arrays.toString(minValues));
-                field("maxClusteringValues", Arrays.toString(maxValues));
-            }
+            printClusteringValues(header, "minClusteringValues", stats.minClusteringValues);
+            printClusteringValues(header, "maxClusteringValues", stats.maxClusteringValues);
+
             field("Estimated droppable tombstones",
                   stats.getEstimatedDroppableTombstoneRatio((int) (System.currentTimeMillis() / 1000) - this.gc));
             field("SSTable Level", stats.sstableLevel);
@@ -436,6 +424,21 @@ public class SSTableMetadataViewer
         }
     }
 
+    private void printClusteringValues(SerializationHeader.Component header, String name, List<ByteBuffer> clusteringValues)
+    {
+        if (header != null && header.getClusteringTypes().size() >= clusteringValues.size())
+        {
+            List<AbstractType<?>> clusteringTypes = header.getClusteringTypes();
+            int size = Math.min(clusteringTypes.size(), clusteringValues.size());
+            String[] values = new String[size];
+            for (int i = 0; i < size; i++)
+            {
+                values[i] = clusteringTypes.get(i).getString(clusteringValues.get(i));
+            }
+            field(name, Arrays.toString(values));
+        }
+    }
+
     private void field(String field, Object value)
     {
         field(field, value, null);
diff --git a/test/data/sstable_metadata/CASSANDRA_20855/nb-1-big-CompressionInfo.db b/test/data/sstable_metadata/CASSANDRA_20855/nb-1-big-CompressionInfo.db
new file mode 100644
index 0000000000000000000000000000000000000000..7d7289b0c5c513e7fbd27e7fba1253035b2db5ab
GIT binary patch
literal 47
ocmZSJ^@%cZ&d)6<N-ZwVFJb@#2Zs9p|NleS423|N5z1x&03F8(LjV8(

literal 0
HcmV?d00001

diff --git a/test/data/sstable_metadata/CASSANDRA_20855/nb-1-big-Data.db b/test/data/sstable_metadata/CASSANDRA_20855/nb-1-big-Data.db
new file mode 100644
index 0000000000000000000000000000000000000000..666fa7e0120d2c7c3b779bd35fe582d7836ebb64
GIT binary patch
literal 83
zcmXS9U|{&f%)nKUSX2_9omv@hSpWb3{{{v|hGG>4-sGIpVkrjoctbvhUra{vhMW=%
l9IR!DIi;ygj0}w041Go{8jCd<nAjv30wjP8#(kw>*8%LW7CZm|

literal 0
HcmV?d00001

diff --git a/test/data/sstable_metadata/CASSANDRA_20855/nb-1-big-Digest.crc32 b/test/data/sstable_metadata/CASSANDRA_20855/nb-1-big-Digest.crc32
new file mode 100644
index 0000000000..441a80dd5f
--- /dev/null
+++ b/test/data/sstable_metadata/CASSANDRA_20855/nb-1-big-Digest.crc32
@@ -0,0 +1 @@
+3617993137
\ No newline at end of file
diff --git a/test/data/sstable_metadata/CASSANDRA_20855/nb-1-big-Filter.db b/test/data/sstable_metadata/CASSANDRA_20855/nb-1-big-Filter.db
new file mode 100644
index 0000000000000000000000000000000000000000..678c0aafe072f94d546aca818837d8e4ce5bf7db
GIT binary patch
literal 16
XcmZQzU|?lnU|?iqV&D*PU|;|M0hj<{

literal 0
HcmV?d00001

diff --git a/test/data/sstable_metadata/CASSANDRA_20855/nb-1-big-Index.db b/test/data/sstable_metadata/CASSANDRA_20855/nb-1-big-Index.db
new file mode 100644
index 0000000000000000000000000000000000000000..d533e13df5e38dbb1041e53d4111dd672890ba66
GIT binary patch
literal 14
VcmZSJDo89UiO){0j5lOp0018{1Nr~}

literal 0
HcmV?d00001

diff --git a/test/data/sstable_metadata/CASSANDRA_20855/nb-1-big-Statistics.db b/test/data/sstable_metadata/CASSANDRA_20855/nb-1-big-Statistics.db
new file mode 100644
index 0000000000000000000000000000000000000000..9f1f6998b8a500a90268e52172323d85e3ff66f2
GIT binary patch
literal 4858
zcmeI$ZA=qq902fp?ylv<wv&P{Ad*c56-`w-h9*J<>q-|l-X|L~^(c+S@s_qYU>X--
zA4Uvg%NUyZGPi^n6bVyjVa$DSgUvZnvJ7U1!K?$vGTWKy@MUy&gG>MaC;Pz1edv>1
zp8MVZf9>-;_uSo+V;F`{SiMBS%#^iaMa0b}ZmhV~iW|o;+B8+{TT{%MO2=-UP$gLH
zHl0;)I)#c7hoCF5yL8W2JIbpa`fS1BDs`1sR@fYd#=2XkGf}S(j=Ig_kEr~9|Fk-q
zkiLm-%?De`c8bT^#7!#FAHqUgY6vW1gHYj-7)I`ld{`;(%+?{4Ln((-R#Dbaj-VVx
zc`@asl$TRZq?|-~Gv!RmFH+t;Yk6GvQSnyFXDJV36Y6@9XQ;S~@<nWR5#@Spj{1G>
z2#)gw*a|(iG8H?7>Tl>H91kDBR(Y|XcoSRuA@;(Tv15j?pInEXa2#7#h@Em2JEIRf
zdja;YXl!Q=c9VeJPR-u|BaVCHkV)Jh*oFDnNAdh5-XBn$&6DTd%whI2HdlBR<?}qA
z|NQkUD6aUr8CmJ;Kn~6GBZo~(53@j-Hys&(xTh(>^D)AB8^xn~79uac8H2oZI2U>O
zAimy-Q}}x8TfRW~sbjw&KbMZ%_fm`p#rOQKMSdfz3;FPajmYjV@%W6o(Rh;Z`S)Nx
zF_;b_A6Rx0`P)nDq1lxe(DRjjU-LbTb1^k&esDcI!(g1R2}^@kyfTdXtE?=of$<Qx
z=RS0paoeBJ^Wu8Z{iH6`UW4)ZoVg!5G9@0(=S7+ahhTiksdpx!pL#m;6X=yAJ2peF
zyJkNFy=m9}W6;JM>vurs9xqRZw#I+D1$ys8^*D6%R~;tkQ+bD%L4ViQ_%`(Y^F!N!
z$trbm543gbR(SuDj_eEY{wMzTH1PgsZJwK-f%Dn1(`j|UoQrtg1?E=_YNhwTT==P}
zeen7y{%UE3?~jnxk>lxdvRH?<b=RE(4$nH#?*LX=n(x5=YkrKFg#C}Wo_SUa@)rnA
zRq%ZfebUgY1M#?T#<lQzt+;D4H-Y%t0&hHA4<ze<>(_v|VR)kVHQ?<>x=zFP6nQr#
z!}(m1Yiw@;`HjWyfid7?KaKXa17G~;<1W}<&!xnnlt7$l8?HRg05<Qr_ozN#MsAz@
z45n*2m6dm9>)@IT56E2CT=H3P&Pm5w<-lwmTyvoT39h-&aFF1d3ynu`%@tg8feG>d
zUUQK{6B*kdGK~H_qf9jwWwnpY)R&39^Jm`y@xzGR6~nkCp}}_7E?Rmolk^*FY87&D
z=7UT?lGWC-YNyNcs;$<t!Lk7|Nu=~EC?iQe8@%=9)+L?zH=lu)_r;boB<GI**K^q~
zG<=?ROF5b2E2`h`>Ha(Rac6bHh)>&T_A#se`|D7#u3T_9?LwKZAphBo`L$IxcFvh|
SPGktTS16lv`oY+%xBme~<jfub

literal 0
HcmV?d00001

diff --git a/test/data/sstable_metadata/CASSANDRA_20855/nb-1-big-Summary.db b/test/data/sstable_metadata/CASSANDRA_20855/nb-1-big-Summary.db
new file mode 100644
index 0000000000000000000000000000000000000000..b2c39da688979697f89dc77b9dab060310cc8000
GIT binary patch
literal 74
scmZQzU}#`qU|<Ad5D)|61|Vi+0TKm?MJ4gssg?1DKn@IWp@<^!08k4I{r~^~

literal 0
HcmV?d00001

diff --git a/test/data/sstable_metadata/CASSANDRA_20855/nb-1-big-TOC.txt b/test/data/sstable_metadata/CASSANDRA_20855/nb-1-big-TOC.txt
new file mode 100644
index 0000000000..b5bdca06d3
--- /dev/null
+++ b/test/data/sstable_metadata/CASSANDRA_20855/nb-1-big-TOC.txt
@@ -0,0 +1,8 @@
+Data.db
+TOC.txt
+Digest.crc32
+Summary.db
+Filter.db
+Statistics.db
+CompressionInfo.db
+Index.db
diff --git a/test/unit/org/apache/cassandra/tools/SSTableMetadataViewerTest.java b/test/unit/org/apache/cassandra/tools/SSTableMetadataViewerTest.java
index 2f6e64baf1..fd03d27193 100644
--- a/test/unit/org/apache/cassandra/tools/SSTableMetadataViewerTest.java
+++ b/test/unit/org/apache/cassandra/tools/SSTableMetadataViewerTest.java
@@ -18,16 +18,23 @@
 
 package org.apache.cassandra.tools;
 
+import java.io.File;
 import java.io.IOException;
 import java.util.Arrays;
 
 import com.google.common.base.CharMatcher;
 
+import org.apache.commons.io.FileUtils;
 import org.apache.commons.lang3.tuple.Pair;
 
 import org.junit.BeforeClass;
+import org.junit.Ignore;
 import org.junit.Test;
 
+import org.apache.cassandra.SchemaLoader;
+import org.apache.cassandra.cql3.QueryProcessor;
+import org.apache.cassandra.db.Keyspace;
+import org.apache.cassandra.service.StorageService;
 import org.apache.cassandra.tools.ToolRunner.ToolResult;
 import org.assertj.core.api.Assertions;
 import org.hamcrest.CoreMatchers;
@@ -225,4 +232,52 @@ public class SSTableMetadataViewerTest extends OfflineToolUtils
         assertKeyspaceNotLoaded();
         assertServerNotLoaded();
     }
+
+    @Test
+    public void testLoadingOfSStableWithRangeTombstoneAsMaxClustering() throws IOException
+    {
+        String sstableWithTombstones = copySSTables("sstable_metadata", "CASSANDRA_20855");
+        ToolResult tool = ToolRunner.invokeClass(SSTableMetadataViewer.class, sstableWithTombstones);
+        assertTrue(tool.getStdout(), CharMatcher.ascii().matchesAllOf(tool.getStdout()));
+        Assertions.assertThat(tool.getStdout()).contains(sstableWithTombstones.replaceAll("-Data\\.db$", ""));
+        Assertions.assertThat(tool.getStdout()).contains("minClusteringValues: [clust_key_1_1, clust_key_2_1]");
+        // Range tombstone end as a max clustering value, only the first clustering key is stored in metadata
+        Assertions.assertThat(tool.getStdout()).contains("maxClusteringValues: [clust_key_1_2]");
+        assertTrue(tool.getStderr(), tool.getStderr().isEmpty());
+        assertEquals(0, tool.getExitCode());
+        assertGoodEnvPostTest();
+    }
+
+    @Test
+    @Ignore
+    // the logic is used only to generate test data for testLoadingOfSStableWithRangeTombstoneAsMaxClustering
+    // the result of the generation is committed to SCM
+    public void generateSSTableWithRangeTombstone() throws IOException
+    {
+        SchemaLoader.prepareServer();
+        StorageService.instance.initServer();
+        Keyspace.setInitialized();
+        QueryProcessor.executeInternal("CREATE KEYSPACE sstable_metadata WITH replication = {'class': 'SimpleStrategy', 'replication_factor': '1'}");
+        QueryProcessor.executeInternal("CREATE TABLE sstable_metadata.CASSANDRA_20855 (pk text, ck_1 text, ck_2 text, val text, PRIMARY KEY (pk, ck_1, ck_2))");
+
+        QueryProcessor.executeInternal(String.format("INSERT INTO sstable_metadata.CASSANDRA_20855 (pk, ck_1, ck_2, val) VALUES ('%s', '%s', '%s', '%s')",
+                                                     "part_key_1", "clust_key_1_1", "clust_key_2_1", "value"));
+
+        QueryProcessor.executeInternal(String.format("DELETE FROM sstable_metadata.CASSANDRA_20855 WHERE pk = '%s' AND ck_1 = '%s'",
+                                                     "part_key_1", "clust_key_1_2"));
+
+        StorageService.instance.forceKeyspaceFlush("sstable_metadata");
+    }
+
+
+    private static String copySSTables(String keyspace, String table) throws IOException
+    {
+        File targetKeyspaceDir = new File("build/test/cassandra/data/", keyspace);
+        targetKeyspaceDir.mkdirs();
+        File targetTableDir = new File(targetKeyspaceDir, table);
+        File srcKeyspaceDir = new File("test/data/" + keyspace);
+        FileUtils.copyDirectory(new File(srcKeyspaceDir, table), targetTableDir);
+        File[] sstableFiles = targetTableDir.listFiles((file) -> file.isFile() && file.getName().endsWith("-Data.db"));
+        return sstableFiles[0].getAbsolutePath();
+    }
 }