src/backup/datastore.rs: generic index_mark_used_chunks implementation, improve GC stats

src/backup/chunk_store.rs

@@ -12,20 +12,24 @@ use super::DataChunk;
 #[derive(Clone, Serialize)]
 pub struct GarbageCollectionStatus {
     pub upid: Option<String>,
-    pub used_bytes: usize,
-    pub used_chunks: usize,
-    pub disk_bytes: usize,
+    pub index_file_count: usize,
+    pub index_data_bytes: u64,
+    pub disk_bytes: u64,
     pub disk_chunks: usize,
+    pub removed_bytes: u64,
+    pub removed_chunks: usize,
 }
 
 impl Default for GarbageCollectionStatus {
     fn default() -> Self {
         GarbageCollectionStatus {
             upid: None,
-            used_bytes: 0,
-            used_chunks: 0,
+            index_file_count: 0,
+            index_data_bytes: 0,
             disk_bytes: 0,
             disk_chunks: 0,
+            removed_bytes: 0,
+            removed_chunks: 0,
         }
     }
 }
@@ -143,13 +147,9 @@ impl ChunkStore {
         })
     }
 
-    pub fn touch_chunk(&self, digest:&[u8]) -> Result<(), Error> {
+    pub fn touch_chunk(&self, digest: &[u8; 32]) -> Result<(), Error> {
 
-        let mut chunk_path = self.chunk_dir.clone();
-        let prefix = digest_to_prefix(&digest);
-        chunk_path.push(&prefix);
-        let digest_str = proxmox::tools::digest_to_hex(&digest);
-        chunk_path.push(&digest_str);
+        let (chunk_path, _digest_str) = self.chunk_path(digest);
 
         const UTIME_NOW: i64 = ((1 << 30) - 1);
         const UTIME_OMIT: i64 = ((1 << 30) - 2);
@@ -172,7 +172,7 @@ impl ChunkStore {
         Ok(())
     }
 
-    pub fn read_chunk(&self, digest:&[u8; 32]) -> Result<DataChunk, Error> {
+    pub fn read_chunk(&self, digest: &[u8; 32]) -> Result<DataChunk, Error> {
 
         let (chunk_path, digest_str) = self.chunk_path(digest);
         let mut file = std::fs::File::open(&chunk_path)
@@ -302,9 +302,11 @@ impl ChunkStore {
                             err,
                         );
                     }
+                    status.removed_chunks += 1;
+                    status.removed_bytes += stat.st_size as u64;
                } else {
                     status.disk_chunks += 1;
-                    status.disk_bytes += stat.st_size as usize;
+                    status.disk_bytes += stat.st_size as u64;
                 }
             }
             drop(lock);

src/backup/datastore.rs

@@ -201,6 +201,28 @@ impl DataStore {
         Ok(list)
     }
 
+    // mark chunks  used by ``index`` as used
+    fn index_mark_used_chunks<I: IndexFile>(
+        &self,
+        index: I,
+        file_name: &Path, // only used for error reporting
+        status: &mut GarbageCollectionStatus,
+    ) -> Result<(), Error> {
+
+        status.index_file_count += 1;
+        status.index_data_bytes += index.index_bytes();
+
+        for pos in 0..index.index_count() {
+            tools::fail_on_shutdown()?;
+            let digest = index.index_digest(pos).unwrap();
+            if let Err(err) = self.chunk_store.touch_chunk(digest) {
+                bail!("unable to access chunk {}, required by {:?} - {}",
+                      proxmox::tools::digest_to_hex(digest), file_name, err);
+            }
+        }
+        Ok(())
+    }
+
     fn mark_used_chunks(&self, status: &mut GarbageCollectionStatus) -> Result<(), Error> {
 
         let image_list = self.list_images()?;
@@ -212,10 +234,10 @@ impl DataStore {
             if let Some(ext) = path.extension() {
                 if ext == "fidx" {
                     let index = self.open_fixed_reader(&path)?;
-                    index.mark_used_chunks(status)?;
+                    self.index_mark_used_chunks(index, &path, status)?;
                 } else if ext == "didx" {
                     let index = self.open_dynamic_reader(&path)?;
-                    index.mark_used_chunks(status)?;
+                    self.index_mark_used_chunks(index, &path, status)?;
                 }
             }
         }
@@ -245,10 +267,14 @@ impl DataStore {
             worker.log("Start GC phase2 (sweep unused chunks)");
             self.chunk_store.sweep_unused_chunks(oldest_writer, &mut gc_status)?;
 
-            worker.log(&format!("Used bytes: {}", gc_status.used_bytes));
-            worker.log(&format!("Used chunks: {}", gc_status.used_chunks));
-            worker.log(&format!("Disk bytes: {}", gc_status.disk_bytes));
+            worker.log(&format!("Removed bytes: {}", gc_status.removed_bytes));
+            worker.log(&format!("Removed chunks: {}", gc_status.removed_chunks));
+            worker.log(&format!("Original data bytes: {}", gc_status.index_data_bytes));
+            let comp_per = (gc_status.disk_bytes*100)/gc_status.index_data_bytes;
+            worker.log(&format!("Disk bytes: {} ({} %)", gc_status.disk_bytes, comp_per));
             worker.log(&format!("Disk chunks: {}", gc_status.disk_chunks));
+            let avg_chunk = gc_status.index_data_bytes/(gc_status.disk_chunks as u64);
+            worker.log(&format!("Average chunk size: {}", avg_chunk));
 
             *self.last_gc_status.lock().unwrap() = gc_status;
 

src/backup/dynamic_index.rs

@@ -167,25 +167,7 @@ impl DynamicIndexReader {
         slice.try_into().unwrap()
     }
 
-    pub fn mark_used_chunks(&self, _status: &mut GarbageCollectionStatus) -> Result<(), Error> {
-        unimplemented!();
-    }
     /*
-    pub fn mark_used_chunks(&self, _status: &mut GarbageCollectionStatus) -> Result<(), Error> {
-
-        for pos in 0..self.index_entries {
-
-            tools::fail_on_shutdown()?;
-
-            let digest = self.chunk_digest(pos);
-            if let Err(err) = self.store.touch_chunk(digest) {
-                bail!("unable to access chunk {}, required by {:?} - {}",
-                      proxmox::tools::digest_to_hex(digest), self.filename, err);
-            }
-        }
-        Ok(())
-    }
-
     pub fn dump_pxar(&self, mut writer: Box<dyn Write>) -> Result<(), Error> {
 
         for pos in 0..self.index_entries {
@@ -243,6 +225,14 @@ impl IndexFile for DynamicIndexReader {
             })
         }
     }
+
+    fn index_bytes(&self) -> u64 {
+        if self.index_entries == 0 {
+            0
+        } else {
+            self.chunk_end((self.index_entries - 1) as usize)
+        }
+    }
 }
 
 pub struct BufferedDynamicReader<S> {

src/backup/fixed_index.rs

@@ -138,27 +138,6 @@ impl FixedIndexReader {
         Ok(())
     }
 
-    pub fn mark_used_chunks(&self, status: &mut GarbageCollectionStatus) -> Result<(), Error> {
-
-        if self.index == std::ptr::null_mut() { bail!("detected closed index file."); }
-
-        status.used_bytes += self.index_length * self.chunk_size;
-        status.used_chunks += self.index_length;
-
-        for pos in 0..self.index_length {
-
-            tools::fail_on_shutdown()?;
-
-            let digest = self.index_digest(pos).unwrap();
-            if let Err(err) = self.store.touch_chunk(digest) {
-                bail!("unable to access chunk {}, required by {:?} - {}",
-                      proxmox::tools::digest_to_hex(digest), self.filename, err);
-            }
-        }
-
-        Ok(())
-    }
-
     pub fn print_info(&self) {
         println!("Filename: {:?}", self.filename);
         println!("Size: {}", self.size);
@@ -180,6 +159,10 @@ impl IndexFile for FixedIndexReader {
             Some(unsafe { std::mem::transmute(self.index.add(pos*32)) })
         }
     }
+
+    fn index_bytes(&self) -> u64 {
+        (self.index_length * self.chunk_size) as u64
+    }
 }
 
 pub struct FixedIndexWriter {

src/backup/index.rs

@@ -2,9 +2,13 @@ use failure::*;
 use futures::*;
 use bytes::{Bytes, BytesMut};
 
+/// Trait to get digest list from index files
+///
+/// To allow easy iteration over all used chunks.
 pub trait IndexFile: Send {
     fn index_count(&self) -> usize;
     fn index_digest(&self, pos: usize) -> Option<&[u8; 32]>;
+    fn index_bytes(&self) -> u64;
 }
 
 /// Encode digest list from an `IndexFile` into a binary stream