Always use blob id as the name of the filecache when use separate blob

Before we only has one blob, called a data blob, so when generating a filecache, we always used the id of this blob as the name of the filecache, and later after supporting separate blobs, we has two blobs, one is a data blob, the other is a meta blob, in order to maintain compatibility, we should always use the data blob id as the filecache name, not the meta blob id. Signed-off-by: zyfjeff <[email protected]>
dragonflyoss · Sep 25, 2023 · b777564 · b777564
1 parent 0cdf472
commit b777564
Showing 1 changed file with 1 addition and 1 deletion.
diff --git a/storage/src/cache/filecache/mod.rs b/storage/src/cache/filecache/mod.rs
@@ -229,7 +229,7 @@ impl FileCacheEntry {
                 Arc::new(BlobStateMap::from(NoopChunkMap::new(true))) as Arc<dyn ChunkMap>;
             (file, None, chunk_map, true, true, false)
         } else {
-            let blob_file_path = format!("{}/{}", mgr.work_dir, blob_meta_id);
+            let blob_file_path = format!("{}/{}", mgr.work_dir, blob_id);
             let (chunk_map, is_direct_chunkmap) =
                 Self::create_chunk_map(mgr, &blob_info, &blob_file_path)?;
             // Validation is supported by RAFS v5 (which has no meta_ci) or v6 with chunk digest array.