Documentation
¶
Index ¶
Constants ¶
const EntryCacheSize = 20
Entries are 0.5MiB in size, so we do ~10MiB of caching here This cache is only useful in the edge case when entry reads are very slow and time out - this makes retried reads faster
const NoSkipCacheTTL = 3 * time.Minute
Variables ¶
var (
ErrNotFound = errors.New("not found")
)
Functions ¶
func NewEntriesChunkNode ¶
Types ¶
type InitialChunker ¶
type InitialChunker struct {
// contains filtered or unexported fields
}
InitialChunker is used for initial entry chain creation. It employs a dual strategy, where it tracks the number of entries, and: For chains with less than longChainThreshold entries, it accumulates entries in memory. which will then form a sorted chain of schema.EntryChunk nodes. This allows creation of ad chains which are read from the index database. For chains with more than longChainThreshold entries, it creates a chain of schema.EntryChunk nodes in car order, noting where each chunk starts and ends in the car file.
func NewInitialChunker ¶
func NewInitialChunker() *InitialChunker
type ServeChunker ¶
type ServeChunker struct {
// contains filtered or unexported fields
}
func NewServeChunker ¶
func NewServeChunker(db *harmonydb.DB, pieceProvider *pieceprovider.PieceProvider, indexStore *indexstore.IndexStore, cpr *cachedreader.CachedPieceReader) *ServeChunker
func (*ServeChunker) GetEntry ¶
func (p *ServeChunker) GetEntry(ctx context.Context, block cid.Cid) (b []byte, err error)
GetEntry retrieves an entry from the provider's database based on the given block CID and provider ID. It returns the entry data as a byte slice, or an error if the entry is not found or an error occurs during retrieval. If the entry is stored as a CAR file, it reconstructs the chunk from the CAR file.
Source Files