consider merge width for tiered trigger + mvcc compaction

Signed-off-by: Alex Chi <iskyzh@gmail.com>

mini-lsm-mvcc/src/compact.rs

@@ -115,26 +115,22 @@ impl LsmStorageInner {
     fn compact_generate_sst_from_iter(
         &self,
         mut iter: impl for<'a> StorageIterator<KeyType<'a> = KeySlice<'a>>,
-        compact_to_bottom_level: bool,
+        _compact_to_bottom_level: bool,
     ) -> Result<Vec<Arc<SsTable>>> {
         let mut builder = None;
         let mut new_sst = Vec::new();
 
+        let mut last_key = Vec::<u8>::new();
         while iter.is_valid() {
             if builder.is_none() {
                 builder = Some(SsTableBuilder::new(self.options.block_size));
             }
             let builder_inner = builder.as_mut().unwrap();
-            if compact_to_bottom_level {
-                if !iter.value().is_empty() {
-                    builder_inner.add(iter.key(), iter.value());
-                }
-            } else {
-                builder_inner.add(iter.key(), iter.value());
-            }
-            iter.next()?;
+            builder_inner.add(iter.key(), iter.value());
 
-            if builder_inner.estimated_size() >= self.options.target_sst_size {
+            let same_as_last_key = iter.key().key_ref() == &last_key;
+
+            if builder_inner.estimated_size() >= self.options.target_sst_size && !same_as_last_key {
                 let sst_id = self.next_sst_id();
                 let builder = builder.take().unwrap();
                 let sst = Arc::new(builder.build(
@@ -144,6 +140,12 @@ impl LsmStorageInner {
                 )?);
                 new_sst.push(sst);
             }
+
+            if !same_as_last_key {
+                last_key.clear();
+                last_key.extend(iter.key().key_ref());
+            }
+            iter.next()?;
         }
         if let Some(builder) = builder {
             let sst_id = self.next_sst_id(); // lock dropped here
@@ -253,6 +255,7 @@ impl LsmStorageInner {
         let CompactionOptions::NoCompaction = self.options.compaction_options else {
             panic!("full compaction can only be called with compaction is not enabled")
         };
+
         let snapshot = {
             let state = self.state.read();
             state.clone()
@@ -264,6 +267,9 @@ impl LsmStorageInner {
             l0_sstables: l0_sstables.clone(),
             l1_sstables: l1_sstables.clone(),
         };
+
+        println!("force full compaction: {:?}", compaction_task);
+
         let sstables = self.compact(&compaction_task)?;
 
         {
@@ -294,10 +300,14 @@ impl LsmStorageInner {
         for sst in l0_sstables.iter().chain(l1_sstables.iter()) {
             std::fs::remove_file(self.path_of_sst(*sst))?;
         }
+
+        println!("force full compaction done");
+
         Ok(())
     }
 
     fn trigger_compaction(&self) -> Result<()> {
+        self.dump_structure();
         let snapshot = {
             let state = self.state.read();
             state.clone()

mini-lsm-mvcc/src/debug.rs

@@ -1,8 +1,8 @@
-use crate::lsm_storage::MiniLsm;
+use crate::lsm_storage::{LsmStorageInner, MiniLsm};
 
-impl MiniLsm {
+impl LsmStorageInner {
     pub fn dump_structure(&self) {
-        let snapshot = self.inner.state.read();
+        let snapshot = self.state.read();
         if !snapshot.l0_sstables.is_empty() {
             println!(
                 "L0 ({}): {:?}",
@@ -15,3 +15,9 @@ impl MiniLsm {
         }
     }
 }
+
+impl MiniLsm {
+    pub fn dump_structure(&self) {
+        self.inner.dump_structure()
+    }
+}

mini-lsm-mvcc/src/tests.rs

@@ -6,7 +6,7 @@ mod week1_day4;
 mod week1_day5;
 mod week1_day6;
 mod week1_day7;
-// mod week2_day1;
-// mod week2_day2;
-// mod week2_day3;
-// mod week2_day4;
+mod week2_day1;
+mod week2_day2;
+mod week2_day3;
+mod week2_day4;