minio/cmd/erasure-lowlevel-heal.go
Harshavardhana ff36baeaa7
fix: attempt to drain the ReadFileStream for connection pooling (#12208)
avoid time_wait build up with getObject requests if there are
pending callers and they timeout, can lead to time_wait states

Bonus share the same buffer pool with erasure healing logic,
additionally also fixes a race where parallel readers were
never cleanup during Encode() phase, because pipe.Reader end
was never closed().

Added closer right away upon an error during Encode to make
sure to avoid racy Close() while stream was still being
Read().
2021-05-04 10:12:08 -07:00

63 lines
2.1 KiB
Go

// Copyright (c) 2015-2021 MinIO, Inc.
//
// This file is part of MinIO Object Storage stack
//
// This program is free software: you can redistribute it and/or modify
// it under the terms of the GNU Affero General Public License as published by
// the Free Software Foundation, either version 3 of the License, or
// (at your option) any later version.
//
// This program is distributed in the hope that it will be useful
// but WITHOUT ANY WARRANTY; without even the implied warranty of
// MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
// GNU Affero General Public License for more details.
//
// You should have received a copy of the GNU Affero General Public License
// along with this program. If not, see <http://www.gnu.org/licenses/>.
package cmd
import (
"context"
"io"
"sync"
"github.com/minio/minio/cmd/logger"
"github.com/minio/minio/pkg/bpool"
)
// Heal heals the shard files on non-nil writers. Note that the quorum passed is 1
// as healing should continue even if it has been successful healing only one shard file.
func (e Erasure) Heal(ctx context.Context, readers []io.ReaderAt, writers []io.Writer, size int64, bp *bpool.BytePoolCap) error {
r, w := io.Pipe()
var wg sync.WaitGroup
wg.Add(1)
go func() {
defer wg.Done()
_, err := e.Decode(ctx, w, readers, 0, size, size, nil)
w.CloseWithError(err)
}()
// Fetch buffer for I/O, returns from the pool if not allocates a new one and returns.
var buffer []byte
switch {
case size == 0:
buffer = make([]byte, 1) // Allocate atleast a byte to reach EOF
case size >= e.blockSize:
buffer = bp.Get()
defer bp.Put(buffer)
case size < e.blockSize:
// No need to allocate fully blockSizeV1 buffer if the incoming data is smaller.
buffer = make([]byte, size, 2*size+int64(e.parityBlocks+e.dataBlocks-1))
}
// quorum is 1 because CreateFile should continue writing as long as we are writing to even 1 disk.
n, err := e.Encode(ctx, r, writers, buffer, 1)
if err == nil && n != size {
logger.LogIf(ctx, errLessData)
err = errLessData
}
r.CloseWithError(err)
wg.Wait()
return err
}