Changing needle_byte_cache so that it doesn't grow so big when larger files are added.

weed/storage/needle_byte_cache.go

@@ -8,6 +8,7 @@ import (
 	"github.com/hashicorp/golang-lru"
 
 	"github.com/chrislusf/seaweedfs/weed/util"
+	"github.com/chrislusf/seaweedfs/weed/glog"
 )
 
 var (
@@ -24,7 +25,7 @@ In caching, the string~[]byte mapping is cached
 */
 func init() {
 	bytesPool = util.NewBytesPool()
-	bytesCache, _ = lru.NewWithEvict(512, func(key interface{}, value interface{}) {
+	bytesCache, _ = lru.NewWithEvict(50, func(key interface{}, value interface{}) {
 		value.(*Block).decreaseReference()
 	})
 }
@@ -46,22 +47,37 @@ func (block *Block) increaseReference() {
 // get bytes from the LRU cache of []byte first, then from the bytes pool
 // when []byte in LRU cache is evicted, it will be put back to the bytes pool
 func getBytesForFileBlock(r *os.File, offset int64, readSize int) (dataSlice []byte, block *Block, err error) {
+	//Skip the cache if we are looking for a block that is too big to fit in the cache (defaulting to 10MB)
+	cacheable := readSize <= (1024*1024*10)
+	if !cacheable {
+		glog.V(4).Infoln("Block too big to keep in cache. Size:", readSize)
+	}
+	cacheKey := string("")
+	if cacheable {
 	// check cache, return if found
-	cacheKey := fmt.Sprintf("%d:%d:%d", r.Fd(), offset>>3, readSize)
+		cacheKey = fmt.Sprintf("%d:%d:%d", r.Fd(), offset >> 3, readSize)
 	if obj, found := bytesCache.Get(cacheKey); found {
+		glog.V(4).Infoln("Found block in cache. Size:", readSize)
 		block = obj.(*Block)
 		block.increaseReference()
 		dataSlice = block.Bytes[0:readSize]
 		return dataSlice, block, nil
+		}
 	}
 
 	// get the []byte from pool
 	b := bytesPool.Get(readSize)
 	// refCount = 2, one by the bytesCache, one by the actual needle object
-	block = &Block{Bytes: b, refCount: 2}
+	refCount := int32(1)
+	if cacheable {
+		refCount = 2
+	}
+	block = &Block{Bytes: b, refCount: refCount}
 	dataSlice = block.Bytes[0:readSize]
 	_, err = r.ReadAt(dataSlice, offset)
+	if cacheable {
 	bytesCache.Add(cacheKey, block)
+	}
 	return dataSlice, block, err
 }