CNDB-12972: Enhance controller config file handling for long names (#1862)

This pull request introduces improvements to the handling of controller
configuration files in the compaction subsystem of Apache Cassandra. The
changes focus on consolidating constants, enhancing file naming logic to
support very long table/keyspace names.

In particular - controller configuration files consists of keyspace and
table names, separated by dot and a fixed suffix. This scheme is
retained as long as the while filename does not exceed 255 chars. If it
does, keyspace and table names are abbreviated and a third component is
added - table id.

In a situation when the keyspace and table names needs to be resolved
from the file name, the code examines the filename whether it has 2 or 3
components (delimited by dots). If 2, follow the legacy pattern, if 3,
ignore keyspace and table names, and read just the table id. Table id is
then used to resolve metadata from schema.

The change is forward compatible because:
- for normal keyspace and table names lengths, the 2 components scheme
is still readable by the new version
- for long keyspace and tables names lengths, the file was not created
at all

The change is backward compatible because:
- for normal keyspace and table names lengths, the 2 components scheme
is retained, thus the old version will still able to resolve it
- for long keyspace and table names lengths, the 3 components schema
will be wrongly resolved and lead to deletion of the controller file as
wrongly resolved keyspace and table names would not be found. This is
compatible with the current behaviour of the old version because in such
a case, the controller config file would not be created at all.

Author: jacek-lewandowski

src/java/org/apache/cassandra/db/compaction/CompactionManager.java

@@ -59,6 +59,7 @@
 import com.google.common.util.concurrent.Uninterruptibles;
 import org.slf4j.Logger;
 import org.slf4j.LoggerFactory;
+
 import com.codahale.metrics.Meter;
 import io.netty.util.concurrent.FastThreadLocal;
 import org.apache.cassandra.cache.AutoSavingCache;
@@ -109,6 +110,7 @@
 import org.apache.cassandra.schema.CompactionParams.TombstoneOption;
 import org.apache.cassandra.schema.Schema;
 import org.apache.cassandra.schema.SchemaConstants;
+import org.apache.cassandra.schema.TableId;
 import org.apache.cassandra.schema.TableMetadata;
 import org.apache.cassandra.service.ActiveRepairService;
 import org.apache.cassandra.service.StorageService;
@@ -165,6 +167,7 @@ protected Boolean initialValue()
         }
     };
     private static final int ACQUIRE_GRANULARITY = COMPACTION_RATE_LIMIT_GRANULARITY_IN_KB.getInt(128) * 1024;
+    private static final String CONTROLLER_CONFIG_JSON_SUFFIX = "-controller-config.JSON";
 
     static
     {
@@ -247,26 +250,40 @@ public static void storeControllerConfig()
     @VisibleForTesting
     public static void cleanupControllerConfig()
     {
-        Pattern fileNamePattern = Pattern.compile("-controller-config.JSON", Pattern.LITERAL);
+        Pattern fileNamePattern = Pattern.compile(CONTROLLER_CONFIG_JSON_SUFFIX, Pattern.LITERAL);
         Pattern keyspaceNameSeparator = Pattern.compile("\\.");
         File dir = DatabaseDescriptor.getMetadataDirectory();
         if (dir != null)
         {
             for (File file : dir.tryList())
             {
-                if (file.name().contains("-controller-config.JSON"))
+                if (file.name().contains(CONTROLLER_CONFIG_JSON_SUFFIX))
                 {
                     String[] names = keyspaceNameSeparator.split(fileNamePattern.matcher(file.name()).replaceAll(""));
-                    try
+                    if (names.length == 2)
                     {
-                        //table exists so keep the file
-                        Schema.instance.getKeyspaceInstance(names[0]).getColumnFamilyStore(names[1]);
+                        try
+                        {
+                            //table exists so keep the file
+                            Schema.instance.getKeyspaceInstance(names[0]).getColumnFamilyStore(names[1]);
+                        }
+                        catch (NullPointerException e)
+                        {
+                            //table does not exist so delete the file
+                            logger.debug("Removing " + file + " because it does not correspond to an existing table");
+                            file.delete();
+                        }
                     }
-                    catch(NullPointerException e)
+                    else if (names.length == 3) // if keyspace/table names are long, we include table id as a 3rd component while the keyspace and table names are abbreviated
                     {
-                        //table does not exist so delete the file
-                        logger.debug("Removing " + file + " because it does not correspond to an existing table");
-                        file.delete();
+                        TableId tableId = TableId.fromHexString(names[2]);
+                        //table exists so keep the file
+                        if (Schema.instance.getTableMetadata(tableId) == null)
+                        {
+                            //table does not exist so delete the file
+                            logger.debug("Removing " + file + " because it does not correspond to an existing table");
+                            file.delete();
+                        }
                     }
                 }
             }

src/java/org/apache/cassandra/db/compaction/unified/AdaptiveController.java

@@ -23,18 +23,18 @@
 import javax.annotation.Nullable;
 
 import com.google.common.annotations.VisibleForTesting;
-
-import org.apache.cassandra.io.FSError;
-import org.apache.cassandra.utils.JVMStabilityInspector;
 import org.slf4j.Logger;
 import org.slf4j.LoggerFactory;
 
 import org.apache.cassandra.db.compaction.CompactionPick;
 import org.apache.cassandra.db.compaction.UnifiedCompactionStrategy;
 import org.apache.cassandra.exceptions.ConfigurationException;
+import org.apache.cassandra.io.FSError;
 import org.apache.cassandra.io.util.File;
 import org.apache.cassandra.io.util.FileReader;
+import org.apache.cassandra.schema.TableMetadata;
 import org.apache.cassandra.utils.FBUtilities;
+import org.apache.cassandra.utils.JVMStabilityInspector;
 import org.apache.cassandra.utils.MonotonicClock;
 import org.apache.cassandra.utils.Overlaps;
 import org.json.simple.JSONArray;
@@ -132,8 +132,7 @@ public AdaptiveController(MonotonicClock clock,
                               double threshold,
                               int minCost,
                               int maxAdaptiveCompactions,
-                              String keyspaceName,
-                              String tableName)
+                              TableMetadata metadata)
     {
         super(clock,
               env,
@@ -154,7 +153,8 @@ public AdaptiveController(MonotonicClock clock,
               reservationsType,
               overlapInclusionMethod,
               parallelizeOutputShards,
-              hasVectorType);
+              hasVectorType,
+              metadata);
 
         this.scalingParameters = scalingParameters;
         this.previousScalingParameters = previousScalingParameters;
@@ -164,8 +164,6 @@ public AdaptiveController(MonotonicClock clock,
         this.threshold = threshold;
         this.minCost = minCost;
         this.maxAdaptiveCompactions = maxAdaptiveCompactions;
-        this.keyspaceName = keyspaceName;
-        this.tableName = tableName;
     }
 
     static Controller fromOptions(Environment env,
@@ -186,14 +184,13 @@ static Controller fromOptions(Environment env,
                                   Overlaps.InclusionMethod overlapInclusionMethod,
                                   boolean parallelizeOutputShards,
                                   boolean hasVectorType,
-                                  String keyspaceName,
-                                  String tableName,
+                                  TableMetadata metadata,
                                   Map<String, String> options)
     {
         int[] scalingParameters = null;
         long currentFlushSize = flushSizeOverride;
 
-        File f = getControllerConfigPath(keyspaceName, tableName);
+        File f = getControllerConfigPath(metadata);
         try
         {
             JSONParser jsonParser = new JSONParser();
@@ -301,8 +298,7 @@ else if (staticScalingFactors != null)
                                       threshold,
                                       minCost,
                                       maxAdaptiveCompactions,
-                                      keyspaceName,
-                                      tableName);
+                                      metadata);
     }
 
     private static int[] readStoredScalingParameters(JSONArray storedScalingParameters)
@@ -602,7 +598,7 @@ else if (i == scalingParameters.length-1)
     @Override
     public void storeControllerConfig()
     {
-        storeOptions(keyspaceName, tableName, scalingParameters, getFlushSizeBytes());
+        storeOptions(metadata, scalingParameters, getFlushSizeBytes());
     }
 
     @Override

src/java/org/apache/cassandra/db/compaction/unified/Controller.java

@@ -330,8 +330,7 @@ public abstract class Controller
     protected final long expiredSSTableCheckFrequency;
     protected final boolean ignoreOverlapsInExpirationCheck;
     protected final boolean parallelizeOutputShards;
-    protected String keyspaceName;
-    protected String tableName;
+    protected final TableMetadata metadata;
 
     protected final int baseShardCount;
     private final boolean isReplicaAware;
@@ -369,7 +368,8 @@ public abstract class Controller
                Reservations.Type reservationsType,
                Overlaps.InclusionMethod overlapInclusionMethod,
                boolean parallelizeOutputShards,
-               boolean hasVectorType)
+               boolean hasVectorType,
+               TableMetadata metadata)
     {
         this.clock = clock;
         this.env = env;
@@ -390,6 +390,7 @@ public abstract class Controller
         this.l0ShardsEnabled = UCS_L0_SHARDS_ENABLED.getBooleanWithLegacyFallback(false); // FIXME VECTOR-23
         this.parallelizeOutputShards = parallelizeOutputShards;
         this.hasVectorType = hasVectorType;
+        this.metadata = metadata;
 
         if (maxSSTablesToCompact <= 0)  // use half the maximum permitted compaction size as upper bound by default
             maxSSTablesToCompact = (int) (dataSetSize * this.maxSpaceOverhead * 0.5 / getMinSstableSizeBytes());
@@ -404,17 +405,26 @@ public abstract class Controller
         this.ignoreOverlapsInExpirationCheck = ALLOW_UNSAFE_AGGRESSIVE_SSTABLE_EXPIRATION && ignoreOverlapsInExpirationCheck;
     }
 
-    public static File getControllerConfigPath(String keyspaceName, String tableName)
+    public static File getControllerConfigPath(TableMetadata metadata)
     {
-        String fileName = keyspaceName + '.' + tableName + '-' + "controller-config.JSON";
+        String suffix = "-controller-config.JSON";
+        String fileName = metadata.keyspace + '.' + metadata.name + suffix;
+        if (fileName.length() > 255)
+        {
+            int spaceLeft = 255 - suffix.length() - 36 - 2; // 36 is the length of a UUID, 2 - for two separators
+            String keyspaceAbbrev = metadata.keyspace.substring(0, Math.min(metadata.keyspace.length(), spaceLeft / 2));
+            spaceLeft -= keyspaceAbbrev.length();
+            String tableAbbrev = metadata.name.substring(0, Math.min(metadata.name.length(), spaceLeft));
+            fileName = String.format("%s.%s.%s%s", keyspaceAbbrev, tableAbbrev, metadata.id.toHexString(), suffix);
+        }
         return new File(DatabaseDescriptor.getMetadataDirectory(), fileName);
     }
 
-    public static void storeOptions(String keyspaceName, String tableName, int[] scalingParameters, long flushSizeBytes)
+    public static void storeOptions(TableMetadata metadata, int[] scalingParameters, long flushSizeBytes)
     {
-        if (SchemaConstants.isSystemKeyspace(keyspaceName))
+        if (SchemaConstants.isSystemKeyspace(metadata.keyspace))
             return;
-        File f = getControllerConfigPath(keyspaceName, tableName);
+        File f = getControllerConfigPath(metadata);
         try(FileWriter fileWriter = new FileWriter(f, File.WriteMode.OVERWRITE);)
         {
             JSONArray jsonArray = new JSONArray();
@@ -1058,8 +1068,7 @@ public static Controller fromOptions(CompactionRealm realm, Map<String, String>
                                                 overlapInclusionMethod,
                                                 parallelizeOutputShards,
                                                 hasVectorType,
-                                                realm.getKeyspaceName(),
-                                                realm.getTableName(),
+                                                realm.metadata(),
                                                 options)
                : StaticController.fromOptions(env,
                                               survivalFactors,
@@ -1079,8 +1088,7 @@ public static Controller fromOptions(CompactionRealm realm, Map<String, String>
                                               overlapInclusionMethod,
                                               parallelizeOutputShards,
                                               hasVectorType,
-                                              realm.getKeyspaceName(),
-                                              realm.getTableName(),
+                                              realm.metadata(),
                                               options,
                                               useVectorOptions);
     }

src/java/org/apache/cassandra/db/compaction/unified/StaticController.java

@@ -28,6 +28,7 @@
 import org.apache.cassandra.io.FSError;
 import org.apache.cassandra.io.util.File;
 import org.apache.cassandra.io.util.FileReader;
+import org.apache.cassandra.schema.TableMetadata;
 import org.apache.cassandra.utils.JVMStabilityInspector;
 import org.apache.cassandra.utils.MonotonicClock;
 import org.apache.cassandra.utils.Overlaps;
@@ -75,8 +76,7 @@ public StaticController(Environment env,
                             Overlaps.InclusionMethod overlapInclusionMethod,
                             boolean parallelizeOutputShards,
                             boolean hasVectorType,
-                            String keyspaceName,
-                            String tableName)
+                            TableMetadata metadata)
     {
         super(MonotonicClock.Global.preciseTime,
               env,
@@ -97,10 +97,9 @@ public StaticController(Environment env,
               reservationsType,
               overlapInclusionMethod,
               parallelizeOutputShards,
-              hasVectorType);
+              hasVectorType,
+              metadata);
         this.scalingParameters = scalingParameters;
-        this.keyspaceName = keyspaceName;
-        this.tableName = tableName;
     }
 
     static Controller fromOptions(Environment env,
@@ -121,8 +120,7 @@ static Controller fromOptions(Environment env,
                                   Overlaps.InclusionMethod overlapInclusionMethod,
                                   boolean parallelizeOutputShards,
                                   boolean hasVectorType,
-                                  String keyspaceName,
-                                  String tableName,
+                                  TableMetadata metadata,
                                   Map<String, String> options,
                                   boolean useVectorOptions)
     {
@@ -136,7 +134,7 @@ static Controller fromOptions(Environment env,
 
         long currentFlushSize = flushSizeOverride;
 
-        File f = getControllerConfigPath(keyspaceName, tableName);
+        File f = getControllerConfigPath(metadata);
         try
         {
             JSONParser jsonParser = new JSONParser();
@@ -184,8 +182,7 @@ static Controller fromOptions(Environment env,
                                     overlapInclusionMethod,
                                     parallelizeOutputShards,
                                     hasVectorType,
-                                    keyspaceName,
-                                    tableName);
+                                    metadata);
     }
 
     public static Map<String, String> validateOptions(Map<String, String> options) throws ConfigurationException
@@ -232,7 +229,7 @@ public int getMaxRecentAdaptiveCompactions()
     @Override
     public void storeControllerConfig()
     {
-        storeOptions(keyspaceName, tableName, scalingParameters, getFlushSizeBytes());
+        storeOptions(metadata, scalingParameters, getFlushSizeBytes());
     }
 
     @Override

src/java/org/apache/cassandra/schema/TableId.java

@@ -66,6 +66,14 @@ public static TableId fromString(String idString)
         return new TableId(UUID.fromString(idString));
     }
 
+    public static TableId fromHexString(String nonDashUUID)
+    {
+        ByteBuffer bytes = ByteBufferUtil.hexToBytes(nonDashUUID);
+        long msb = bytes.getLong(0);
+        long lsb = bytes.getLong(8);
+        return fromUUID(new UUID(msb, lsb));
+    }
+
     @Nullable
     public static Pair<String, TableId> tableNameAndIdFromFilename(String filename)
     {
@@ -79,14 +87,6 @@ public static Pair<String, TableId> tableNameAndIdFromFilename(String filename)
         return Pair.create(tableName, id);
     }
 
-    private static TableId fromHexString(String nonDashUUID)
-    {
-        ByteBuffer bytes = ByteBufferUtil.hexToBytes(nonDashUUID);
-        long msb = bytes.getLong(0);
-        long lsb = bytes.getLong(8);
-        return fromUUID(new UUID(msb, lsb));
-    }
-
     /**
      * Creates the UUID of a system table.
      *