build(deps): Upgrade OpenDAL to 0.47 (#4224)

* catch up changes Signed-off-by: tison <wander4096@gmail.com> * fmt Signed-off-by: tison <wander4096@gmail.com> * Fix cache for 0471 (#7) * Fix cache for 0471 Signed-off-by: Xuanwo <github@xuanwo.io> * Make clippy happy Signed-off-by: Xuanwo <github@xuanwo.io> --------- Signed-off-by: Xuanwo <github@xuanwo.io> * tidy Signed-off-by: tison <wander4096@gmail.com> * use opendal's exported type Signed-off-by: tison <wander4096@gmail.com> * clippy Signed-off-by: tison <wander4096@gmail.com> * fmt Signed-off-by: tison <wander4096@gmail.com> --------- Signed-off-by: tison <wander4096@gmail.com> Signed-off-by: Xuanwo <github@xuanwo.io> Co-authored-by: Xuanwo <github@xuanwo.io>
GreptimeTeam · Jul 1, 2024 · 2665616 · 2665616
1 parent e531326
commit 2665616
Show file tree

Hide file tree

Showing 15 changed files with 200 additions and 158 deletions.
diff --git a/Cargo.lock b/Cargo.lock
diff --git a/src/common/datasource/src/file_format.rs b/src/common/datasource/src/file_format.rs
@@ -149,7 +149,9 @@ pub fn open_with_decoder<T: ArrowDecoder, F: Fn() -> DataFusionResult<T>>(
             .reader(&path)
             .await
             .map_err(|e| DataFusionError::External(Box::new(e)))?
-            .into_bytes_stream(..);
+            .into_bytes_stream(..)
+            .await
+            .map_err(|e| DataFusionError::External(Box::new(e)))?;
 
         let mut upstream = compression_type.convert_stream(reader).fuse();
 

diff --git a/src/common/datasource/src/file_format/csv.rs b/src/common/datasource/src/file_format/csv.rs
@@ -169,11 +169,14 @@ impl FileFormat for CsvFormat {
             .stat(path)
             .await
             .context(error::ReadObjectSnafu { path })?;
+
         let reader = store
             .reader(path)
             .await
             .context(error::ReadObjectSnafu { path })?
             .into_futures_async_read(0..meta.content_length())
+            .await
+            .context(error::ReadObjectSnafu { path })?
             .compat();
 
         let decoded = self.compression_type.convert_async_read(reader);

diff --git a/src/common/datasource/src/file_format/json.rs b/src/common/datasource/src/file_format/json.rs
@@ -87,11 +87,14 @@ impl FileFormat for JsonFormat {
             .stat(path)
             .await
             .context(error::ReadObjectSnafu { path })?;
+
         let reader = store
             .reader(path)
             .await
             .context(error::ReadObjectSnafu { path })?
             .into_futures_async_read(0..meta.content_length())
+            .await
+            .context(error::ReadObjectSnafu { path })?
             .compat();
 
         let decoded = self.compression_type.convert_async_read(reader);

diff --git a/src/common/datasource/src/file_format/parquet.rs b/src/common/datasource/src/file_format/parquet.rs
@@ -52,11 +52,14 @@ impl FileFormat for ParquetFormat {
             .stat(path)
             .await
             .context(error::ReadObjectSnafu { path })?;
+
         let mut reader = store
             .reader(path)
             .await
             .context(error::ReadObjectSnafu { path })?
             .into_futures_async_read(0..meta.content_length())
+            .await
+            .context(error::ReadObjectSnafu { path })?
             .compat();
 
         let metadata = reader
@@ -129,6 +132,7 @@ impl LazyParquetFileReader {
                 .reader(&self.path)
                 .await?
                 .into_futures_async_read(0..meta.content_length())
+                .await?
                 .compat();
             self.reader = Some(reader);
         }

diff --git a/src/datanode/src/store.rs b/src/datanode/src/store.rs
@@ -20,6 +20,7 @@ mod gcs;
 mod oss;
 mod s3;
 
+use std::sync::Arc;
 use std::time::Duration;
 use std::{env, path};
 
@@ -28,7 +29,7 @@ use common_telemetry::info;
 use object_store::layers::{LruCacheLayer, RetryLayer};
 use object_store::services::Fs;
 use object_store::util::{join_dir, normalize_dir, with_instrument_layers};
-use object_store::{HttpClient, ObjectStore};
+use object_store::{HttpClient, ObjectStore, ObjectStoreBuilder};
 use snafu::prelude::*;
 
 use crate::config::{ObjectStoreConfig, DEFAULT_OBJECT_STORE_CACHE_SIZE};
@@ -106,13 +107,14 @@ async fn create_object_store_with_cache(
     if let Some(path) = cache_path {
         let atomic_temp_dir = join_dir(path, ".tmp/");
         clean_temp_dir(&atomic_temp_dir)?;
-        let mut builder = Fs::default();
-        builder.root(path).atomic_write_dir(&atomic_temp_dir);
-        let cache_store = ObjectStore::new(builder)
-            .context(error::InitBackendSnafu)?
-            .finish();
 
-        let cache_layer = LruCacheLayer::new(cache_store, cache_capacity.0 as usize)
+        let cache_store = {
+            let mut builder = Fs::default();
+            builder.root(path).atomic_write_dir(&atomic_temp_dir);
+            builder.build().context(error::InitBackendSnafu)?
+        };
+
+        let cache_layer = LruCacheLayer::new(Arc::new(cache_store), cache_capacity.0 as usize)
             .await
             .context(error::InitBackendSnafu)?;
 

diff --git a/src/mito2/src/cache/write_cache.rs b/src/mito2/src/cache/write_cache.rs
@@ -188,7 +188,9 @@ impl WriteCache {
             .reader(&cache_path)
             .await
             .context(error::OpenDalSnafu)?
-            .into_futures_async_read(0..cached_value.content_length());
+            .into_futures_async_read(0..cached_value.content_length())
+            .await
+            .context(error::OpenDalSnafu)?;
 
         let mut writer = remote_store
             .writer_with(upload_path)

diff --git a/src/mito2/src/sst/index/applier.rs b/src/mito2/src/sst/index/applier.rs
@@ -28,8 +28,8 @@ use store_api::storage::RegionId;
 
 use crate::cache::file_cache::{FileCacheRef, FileType, IndexKey};
 use crate::error::{
-    ApplyIndexSnafu, PuffinBlobTypeNotFoundSnafu, PuffinReadBlobSnafu, PuffinReadMetadataSnafu,
-    Result,
+    ApplyIndexSnafu, OpenDalSnafu, PuffinBlobTypeNotFoundSnafu, PuffinReadBlobSnafu,
+    PuffinReadMetadataSnafu, Result,
 };
 use crate::metrics::{
     INDEX_APPLY_ELAPSED, INDEX_APPLY_MEMORY_USAGE, INDEX_PUFFIN_READ_BYTES_TOTAL,
@@ -128,11 +128,19 @@ impl SstIndexApplier {
             return Ok(None);
         };
 
-        Ok(file_cache
+        let Some(reader) = file_cache
             .reader(IndexKey::new(self.region_id, file_id, FileType::Puffin))
             .await
-            .map(|v| v.into_futures_async_read(0..indexed_value.file_size as u64))
-            .map(PuffinFileReader::new))
+        else {
+            return Ok(None);
+        };
+
+        let reader = reader
+            .into_futures_async_read(0..indexed_value.file_size as u64)
+            .await
+            .context(OpenDalSnafu)?;
+
+        Ok(Some(PuffinFileReader::new(reader)))
     }
 
     /// Helper function to create a [`PuffinFileReader`] from the remote index file.

diff --git a/src/mito2/src/sst/index/store.rs b/src/mito2/src/sst/index/store.rs
@@ -67,7 +67,9 @@ impl InstrumentedStore {
             .reader(path)
             .await
             .context(OpenDalSnafu)?
-            .into_futures_async_read(0..meta.content_length());
+            .into_futures_async_read(0..meta.content_length())
+            .await
+            .context(OpenDalSnafu)?;
         Ok(InstrumentedAsyncRead::new(
             reader,
             read_byte_count,

diff --git a/src/object-store/Cargo.toml b/src/object-store/Cargo.toml
@@ -17,7 +17,7 @@ futures.workspace = true
 lazy_static.workspace = true
 md5 = "0.7"
 moka = { workspace = true, features = ["future"] }
-opendal = { version = "0.46", features = [
+opendal = { version = "0.47", features = [
     "layers-tracing",
     "services-azblob",
     "services-fs",

diff --git a/src/object-store/src/layers/lru_cache.rs b/src/object-store/src/layers/lru_cache.rs
@@ -14,26 +14,26 @@
 
 use std::sync::Arc;
 
-use opendal::raw::oio::ReadDyn;
+use opendal::raw::oio::Reader;
 use opendal::raw::{
     Access, Layer, LayeredAccess, OpDelete, OpList, OpRead, OpWrite, RpDelete, RpList, RpRead,
     RpWrite,
 };
-use opendal::{Operator, Result};
+use opendal::Result;
 mod read_cache;
 use common_telemetry::info;
 use read_cache::ReadCache;
 
 /// An opendal layer with local LRU file cache supporting.
 #[derive(Clone)]
-pub struct LruCacheLayer {
+pub struct LruCacheLayer<C: Access> {
     // The read cache
-    read_cache: ReadCache,
+    read_cache: ReadCache<C>,
 }
 
-impl LruCacheLayer {
+impl<C: Access> LruCacheLayer<C> {
     /// Create a `[LruCacheLayer]` with local file cache and capacity in bytes.
-    pub async fn new(file_cache: Operator, capacity: usize) -> Result<Self> {
+    pub async fn new(file_cache: Arc<C>, capacity: usize) -> Result<Self> {
         let read_cache = ReadCache::new(file_cache, capacity);
         let (entries, bytes) = read_cache.recover_cache().await?;
 
@@ -52,12 +52,12 @@ impl LruCacheLayer {
 
     /// Returns the read cache statistics info `(EntryCount, SizeInBytes)`.
     pub async fn read_cache_stat(&self) -> (u64, u64) {
-        self.read_cache.stat().await
+        self.read_cache.cache_stat().await
     }
 }
 
-impl<I: Access> Layer<I> for LruCacheLayer {
-    type LayeredAccess = LruCacheAccess<I>;
+impl<I: Access, C: Access> Layer<I> for LruCacheLayer<C> {
+    type LayeredAccess = LruCacheAccess<I, C>;
 
     fn layer(&self, inner: I) -> Self::LayeredAccess {
         LruCacheAccess {
@@ -68,14 +68,14 @@ impl<I: Access> Layer<I> for LruCacheLayer {
 }
 
 #[derive(Debug)]
-pub struct LruCacheAccess<I> {
+pub struct LruCacheAccess<I, C> {
     inner: I,
-    read_cache: ReadCache,
+    read_cache: ReadCache<C>,
 }
 
-impl<I: Access> LayeredAccess for LruCacheAccess<I> {
+impl<I: Access, C: Access> LayeredAccess for LruCacheAccess<I, C> {
     type Inner = I;
-    type Reader = Arc<dyn ReadDyn>;
+    type Reader = Reader;
     type BlockingReader = I::BlockingReader;
     type Writer = I::Writer;
     type BlockingWriter = I::BlockingWriter;
@@ -87,7 +87,9 @@ impl<I: Access> LayeredAccess for LruCacheAccess<I> {
     }
 
     async fn read(&self, path: &str, args: OpRead) -> Result<(RpRead, Self::Reader)> {
-        self.read_cache.read(&self.inner, path, args).await
+        self.read_cache
+            .read_from_cache(&self.inner, path, args)
+            .await
     }
 
     async fn write(&self, path: &str, args: OpWrite) -> Result<(RpWrite, Self::Writer)> {