What causes "read digest in range record: %w" in weaviate/weaviate?

重试摘要请求（异步复制调度程序将在下一个 tick 重试完整周期）。

read digest in range record: %w

Error message

read digest in range record: %w

What it means

Wraps io.ReadFull failure while streaming fixed-size digest records from the binary response body in readDigestsInRangeBinaryStream. EOF terminates the stream cleanly; any other error (unexpected EOF, connection reset, read timeout) means the peer's binary payload was truncated or the connection broke mid-stream. Pre-sizing is clamped to the request's own bound precisely so this path cannot over-allocate from a peer-controlled Content-Length.

Source

Thrown at adapters/clients/replication.go:209

		// Content-Length is peer-controlled; never pre-size beyond the request's own bound
		recordCount := contentLength / int64(replica.DigestObjectsInRangeRecordLength)
		if recordCount > int64(maxRecords) {
			recordCount = int64(maxRecords)
		}
		// a caller-supplied negative bound must not reach make
		if recordCount < 0 {
			recordCount = 0
		}
		results = make([]types.RepairDigest, 0, int(recordCount))
	}
	var buf [replica.DigestObjectsInRangeRecordLength]byte
	for {
		_, err := io.ReadFull(r, buf[:])
		if stderrors.Is(err, io.EOF) {
			return results, nil
		}
		if err != nil {
			return nil, fmt.Errorf("read digest in range record: %w", err)
		}
		id, err := uuid.FromBytes(buf[:16])
		if err != nil {
			return nil, fmt.Errorf("parse uuid from binary record: %w", err)
		}
		updateTime := int64(binary.BigEndian.Uint64(buf[16:]))
		results = append(results, types.RepairDigest{
			ID:         id,
			UpdateTime: updateTime,
		})
	}
}

// CompareDigests sends the source's local digests to the target node using a
// compact binary protocol (CompareDigestsRecordLength bytes per record) and
// returns the subset that requires source-side action.
//
// Wire format: CompareDigestsRecordLength bytes per record —

View on GitHub (pinned to 75aa4b6d11)

Solutions

重试摘要请求（异步复制调度程序将在下一个 tick 重试完整周期）。
检查两个节点之间的网络稳定性（数据包丢失、代理、超时）。
检查目标节点日志，查看在流式传输期间是否崩溃或重启。

Defensive patterns

Strategy: retry

Try / catch

digests, err := client.DigestObjectsInRange(ctx, host, index, shard, ...)
if err != nil {
    if strings.Contains(err.Error(), "read digest in range record") {
        // stream interrupted — retry full request
    }
    return err
}

Prevention

确保网络稳定，避免在节点之间设置激进的空闲超时或代理。
在流式传输时避免在目标节点上进行重量级的维护操作（重启/升级）。
使用 context deadlines 来防止无限挂起。

When it happens

Trigger: 网络中断，连接在传输中途被重置，或在流式传输二进制响应体时超时。

Common situations: 目标节点与该节点之间的网络不稳定，目标节点在流式传输时崩溃/重启，或代理截断连接。

Related errors

AI-assisted analysis of weaviate/weaviate@75aa4b6d11 (2026-09-04). Data as JSON: /api/errors/87deae7b6345655f. Report an issue: GitHub.