Improve database compaction and prune-states (#5142)

* Fix no-op state prune check

* Compact freezer DB after pruning

* Refine DB compaction

* Add blobs-db options to inspect/compact

* Better key size

* Fix compaction end key

beacon_node/store/src/hot_cold_store.rs

@@ -2376,6 +2376,9 @@ impl<E: EthSpec, Hot: ItemStore<E>, Cold: ItemStore<E>> HotColdDB<E, Hot, Cold>
             self.cold_db.do_atomically(cold_ops)?;
         }
 
+        // In order to reclaim space, we need to compact the freezer DB as well.
+        self.cold_db.compact()?;
+
         Ok(())
     }
 }

beacon_node/store/src/leveldb_store.rs

@@ -154,25 +154,15 @@ impl<E: EthSpec> KeyValueStore<E> for LevelDB<E> {
         self.transaction_mutex.lock()
     }
 
-    /// Compact all values in the states and states flag columns.
-    fn compact(&self) -> Result<(), Error> {
-        let endpoints = |column: DBColumn| {
-            (
-                BytesKey::from_vec(get_key_for_col(column.as_str(), Hash256::zero().as_bytes())),
-                BytesKey::from_vec(get_key_for_col(
-                    column.as_str(),
-                    Hash256::repeat_byte(0xff).as_bytes(),
-                )),
-            )
-        };
-
-        for (start_key, end_key) in [
-            endpoints(DBColumn::BeaconStateTemporary),
-            endpoints(DBColumn::BeaconState),
-            endpoints(DBColumn::BeaconStateSummary),
-        ] {
-            self.db.compact(&start_key, &end_key);
-        }
+    fn compact_column(&self, column: DBColumn) -> Result<(), Error> {
+        // Use key-size-agnostic keys [] and 0xff..ff with a minimum of 32 bytes to account for
+        // columns that may change size between sub-databases or schema versions.
+        let start_key = BytesKey::from_vec(get_key_for_col(column.as_str(), &[]));
+        let end_key = BytesKey::from_vec(get_key_for_col(
+            column.as_str(),
+            &vec![0xff; std::cmp::max(column.key_size(), 32)],
+        ));
+        self.db.compact(&start_key, &end_key);
         Ok(())
     }
 

beacon_node/store/src/lib.rs

@@ -80,8 +80,22 @@ pub trait KeyValueStore<E: EthSpec>: Sync + Send + Sized + 'static {
     /// this method. In future we may implement a safer mandatory locking scheme.
     fn begin_rw_transaction(&self) -> MutexGuard<()>;
 
-    /// Compact the database, freeing space used by deleted items.
-    fn compact(&self) -> Result<(), Error>;
+    /// Compact a single column in the database, freeing space used by deleted items.
+    fn compact_column(&self, column: DBColumn) -> Result<(), Error>;
+
+    /// Compact a default set of columns that are likely to free substantial space.
+    fn compact(&self) -> Result<(), Error> {
+        // Compact state and block related columns as they are likely to have the most churn,
+        // i.e. entries being created and deleted.
+        for column in [
+            DBColumn::BeaconState,
+            DBColumn::BeaconStateSummary,
+            DBColumn::BeaconBlock,
+        ] {
+            self.compact_column(column)?;
+        }
+        Ok(())
+    }
 
     /// Iterate through all keys and values in a particular column.
     fn iter_column<K: Key>(&self, column: DBColumn) -> ColumnIter<K> {

beacon_node/store/src/memory_store.rs

@@ -108,7 +108,7 @@ impl<E: EthSpec> KeyValueStore<E> for MemoryStore<E> {
         self.transaction_mutex.lock()
     }
 
-    fn compact(&self) -> Result<(), Error> {
+    fn compact_column(&self, _column: DBColumn) -> Result<(), Error> {
         Ok(())
     }
 }