mirror of
https://github.com/seaweedfs/seaweedfs.git
synced 2026-11-12 14:58:44 +01:00
Chris Lu
d7865909ba
* feat(mount): proactive flush of idle writable chunks Add a background goroutine that periodically scans writable chunks across all open file handles and seals those that are idle and unlikely to receive further writes, submitting them for async upload. A writable chunk is proactively flushed when it has been idle for 500ms AND meets one of: nearly full (>=90%), behind the sequential write frontier by 2+ chunks, or stale for 5+ seconds. Flushing only happens when the upload pipeline has spare capacity (< half of concurrent writer slots in use). This prevents partial chunks from accumulating until fsync/close, which is particularly beneficial for bursty or small-file workloads where chunks may never reach IsComplete(). Also fixes a latent bug in ActivityScore where MarkRead/MarkWrite used value receivers, silently discarding all mutations. * refactor(mount): reuse WriterPattern instead of duplicating sequential detection Remove the isSequential atomic from UploadPipeline. The proactive flusher now reads IsSequentialMode() from the existing WriterPattern on PageWriter and passes it as a parameter to ProactiveFlush. This avoids duplicating the sequential/random detection that WriterPattern already maintains. * fix(mount): address PR review feedback - Make ActivityScore thread-safe using atomics (CAS loop for score updates, atomic load/swap for timestamp). Previously MarkRead was called under RLock while MarkWrite held a write lock, creating a data race on the shared fields. - Fix ProactiveFlush half-capacity guard: use multiplication (uploaderCount*2 >= max) instead of floor division (max/2) which misbehaves for odd or small concurrentWriterMax values. * fix(mount): review fixes for proactive flush - Fix TOCTOU race in lastWriteChunkIndex update: use CAS loop so concurrent writers cannot regress the frontier. - Remove unused UploaderCount() getter. - Reuse the caller-provided tsNs instead of calling time.Now() again in WriteDataAt for lastWriteTsNs, eliminating a redundant syscall per write.
121 lines
2.7 KiB
Go
121 lines
2.7 KiB
Go
package page_writer
|
|
|
|
import (
|
|
"sync"
|
|
"sync/atomic"
|
|
|
|
"github.com/seaweedfs/seaweedfs/weed/util"
|
|
"github.com/seaweedfs/seaweedfs/weed/util/mem"
|
|
)
|
|
|
|
var (
|
|
_ = PageChunk(&MemChunk{})
|
|
|
|
memChunkCounter int64
|
|
)
|
|
|
|
type MemChunk struct {
|
|
sync.RWMutex
|
|
buf []byte
|
|
usage *ChunkWrittenIntervalList
|
|
chunkSize int64
|
|
logicChunkIndex LogicChunkIndex
|
|
activityScore *ActivityScore
|
|
lastWriteTsNs atomic.Int64
|
|
}
|
|
|
|
func NewMemChunk(logicChunkIndex LogicChunkIndex, chunkSize int64) *MemChunk {
|
|
atomic.AddInt64(&memChunkCounter, 1)
|
|
return &MemChunk{
|
|
logicChunkIndex: logicChunkIndex,
|
|
chunkSize: chunkSize,
|
|
buf: mem.Allocate(int(chunkSize)),
|
|
usage: newChunkWrittenIntervalList(),
|
|
activityScore: NewActivityScore(),
|
|
}
|
|
}
|
|
|
|
func (mc *MemChunk) FreeResource() {
|
|
mc.Lock()
|
|
defer mc.Unlock()
|
|
|
|
atomic.AddInt64(&memChunkCounter, -1)
|
|
mem.Free(mc.buf)
|
|
}
|
|
|
|
func (mc *MemChunk) WriteDataAt(src []byte, offset int64, tsNs int64) (n int) {
|
|
mc.Lock()
|
|
defer mc.Unlock()
|
|
|
|
innerOffset := offset % mc.chunkSize
|
|
n = copy(mc.buf[innerOffset:], src)
|
|
mc.usage.MarkWritten(innerOffset, innerOffset+int64(n), tsNs)
|
|
mc.activityScore.MarkWrite()
|
|
mc.lastWriteTsNs.Store(tsNs)
|
|
|
|
return
|
|
}
|
|
|
|
func (mc *MemChunk) ReadDataAt(p []byte, off int64, tsNs int64) (maxStop int64) {
|
|
mc.RLock()
|
|
defer mc.RUnlock()
|
|
|
|
memChunkBaseOffset := int64(mc.logicChunkIndex) * mc.chunkSize
|
|
for t := mc.usage.head.next; t != mc.usage.tail; t = t.next {
|
|
logicStart := max(off, memChunkBaseOffset+t.StartOffset)
|
|
logicStop := min(off+int64(len(p)), memChunkBaseOffset+t.stopOffset)
|
|
if logicStart < logicStop {
|
|
copy(p[logicStart-off:logicStop-off], mc.buf[logicStart-memChunkBaseOffset:logicStop-memChunkBaseOffset])
|
|
maxStop = max(maxStop, logicStop)
|
|
}
|
|
}
|
|
mc.activityScore.MarkRead()
|
|
|
|
return
|
|
}
|
|
|
|
func (mc *MemChunk) IsComplete() bool {
|
|
mc.RLock()
|
|
defer mc.RUnlock()
|
|
|
|
return mc.usage.IsComplete(mc.chunkSize)
|
|
}
|
|
|
|
func (mc *MemChunk) ActivityScore() int64 {
|
|
return mc.activityScore.ActivityScore()
|
|
}
|
|
|
|
func (mc *MemChunk) WrittenSize() int64 {
|
|
mc.RLock()
|
|
defer mc.RUnlock()
|
|
|
|
return mc.usage.WrittenSize()
|
|
}
|
|
|
|
func (mc *MemChunk) LastWriteTsNs() int64 {
|
|
return mc.lastWriteTsNs.Load()
|
|
}
|
|
|
|
func (mc *MemChunk) SaveContent(saveFn SaveToStorageFunc) {
|
|
mc.RLock()
|
|
defer mc.RUnlock()
|
|
|
|
if saveFn == nil {
|
|
return
|
|
}
|
|
|
|
for t := mc.usage.head.next; t != mc.usage.tail; t = t.next {
|
|
startOffset := t.StartOffset
|
|
stopOffset := t.stopOffset
|
|
tsNs := t.TsNs
|
|
for t != mc.usage.tail && t.next.StartOffset == stopOffset {
|
|
stopOffset = t.next.stopOffset
|
|
t = t.next
|
|
tsNs = max(tsNs, t.TsNs)
|
|
}
|
|
reader := util.NewBytesReader(mc.buf[startOffset:stopOffset])
|
|
saveFn(reader, int64(mc.logicChunkIndex)*mc.chunkSize+startOffset, stopOffset-startOffset, tsNs, func() {
|
|
})
|
|
}
|
|
}
|