Mongoose使用withTransaction出现无限重试及事务不匹配报错问题

问题现象 #

使用Mongoose的withTransaction处理事务时出现两类报错，且触发无限重试（错误日志显示已重试1998次仍未停止），核心报错信息如下：

Mongodb Error withTransaction: Given transaction number 1 does not match any in-progress transactions. The active transaction number is -1

运行环境 #

• MongoDB Server 4.4.x
• mongoose ^5.10.0

withTransaction官方设计逻辑 #

Mongoose官方将withTransaction标注为事务最佳实践，该方法会自动处理以下逻辑：

• 创建事务
• 操作执行成功时提交事务
• 操作抛出异常时中止事务
• 出现临时事务错误时自动重试

问题代码 #

try {
  const session = await db.startSession();
  await session.withTransaction(() => {
    console.log('transaction in progress')
    return new Promise((resolve, reject) => {
      Promise.all([
        Model1().bulkWrite([...], { session }),
        Model2().bulkWrite([...], { session })
      ])
        .then(resp => {
          console.log('bulk write resolved successfully')
          resolve(resp)
        })
        .catch(error => {
          console.log('transaction error', error)
          reject(error)
        })
    })
  })
} catch (error) {
  console.error('withTransaction: ', error)
  throw error
} finally {
  session.endSession()
}

报错详情 #

错误1：NoSuchTransaction #

transaction in progress
transaction error MongoError: Given transaction number 1998 does not match any in-progress transactions. The active transaction number is 1997
    at MessageStream.messageHandler (/var/task/node_modules/mongodb/lib/cmap/connection.js:299:20)
    at MessageStream.emit (events.js:400:28)
    at processIncomingData (/var/task/node_modules/mongodb/lib/cmap/message_stream.js:144:12)
    at MessageStream._write (/var/task/node_modules/mongodb/lib/cmap/message_stream.js:42:5)
    at writeOrBuffer (internal/streams/writable.js:358:12)
    at MessageStream.Writable.write (internal/streams/writable.js:303:10)
    at TLSSocket.ondata (internal/streams/readable.js:726:22)
    at TLSSocket.emit (events.js:400:28)
    at addChunk (internal/streams/readable.js:290:12)
    at readableAddChunk (internal/streams/readable.js:265:9)
    at TLSSocket.Readable.push (internal/streams/readable.js:204:10)
    at TLSWrap.onStreamRead (internal/stream_base_commons.js:188:23) {
  operationTime: Timestamp { _bsontype: 'Timestamp', low_: 1, high_: 1634234124 },
  ok: 0,
  code: 251,
  codeName: 'NoSuchTransaction',
  '$clusterTime': {
    clusterTime: Timestamp { _bsontype: 'Timestamp', low_: 1, high_: 1634234124 },
    signature: { hash: [Binary], keyId: [Long] }
  },
  [Symbol(errorLabels)]: Set(1) { 'TransientTransactionError' }
}

错误2：WriteConflict #

transaction in progress
transaction error MongoError: WriteConflict error: this operation conflicted with another operation. Please retry your operation or multi-document transaction.
    at MessageStream.messageHandler (/var/task/node_modules/mongodb/lib/cmap/connection.js:299:20)
    at MessageStream.emit (events.js:400:28)
    at processIncomingData (/var/task/node_modules/mongodb/lib/cmap/message_stream.js:144:12)
    at MessageStream._write (/var/task/node_modules/mongodb/lib/cmap/message_stream.js:42:5)
    at writeOrBuffer (internal/streams/writable.js:358:12)
    at MessageStream.Writable.write (internal/streams/writable.js:303:10)
    at TLSSocket.ondata (internal/streams/readable.js:726:22)
    at TLSSocket.emit (events.js:400:28)
    at addChunk (internal/streams/readable.js:290:12)
    at readableAddChunk (internal/streams/readable.js:265:9)
    at TLSSocket.Readable.push (internal/streams/readable.js:204:10)
    at TLSWrap.onStreamRead (internal/stream_base_commons.js:188:23) {
  operationTime: Timestamp { _bsontype: 'Timestamp', low_: 1, high_: 1634234124 },
  ok: 0,
  code: 112,
  codeName: 'WriteConflict',
  '$clusterTime': {
    clusterTime: Timestamp { _bsontype: 'Timestamp', low_: 1, high_: 1634234124 },
    signature: { hash: [Binary], keyId: [Long] }
  },
  [Symbol(errorLabels)]: Set(1) { 'TransientTransactionError' }
}

根因分析与解决方案 #

核心错误点 #

bulkWrite是Mongoose Model的静态方法，不能在Model实例上调用。问题代码中Model1().bulkWrite()的写法是先创建了Model1的实例，再调用实例上不存在的bulkWrite方法，导致session没有正确绑定到操作上，操作未加入当前事务，触发事务编号不匹配的错误。

解决方案 #

1. 修正bulkWrite调用方式，直接调用Model的静态方法，确保session正确传递：

// 修正前（错误）：Model1()返回实例，无静态bulkWrite方法
// Model1().bulkWrite([...], { session })
// 修正后（正确）：直接调用Model的静态bulkWrite
Model1.bulkWrite([...], { session })

2. 限制事务重试次数，避免无限重试。withTransaction默认对标记为TransientTransactionError的错误无重试上限，可通过配置maxRetryTimeMS参数限制最大重试时间：

await session.withTransaction(async () => {
  // 事务执行逻辑
}, {
  maxRetryTimeMS: 5000 // 最多重试5秒，可根据业务调整
})

3. 排查并发操作冲突。WriteConflict报错说明有其他并发事务同时修改同一批数据，可通过调整业务逻辑避免同时操作相同资源，或增加分布式锁控制操作时序减少冲突概率。
4. 若业务需要在事务内创建并保存单条文档，使用实例的save方法并传递session即可：

const newDoc = new Model1({ /* 文档字段 */ });
await newDoc.save({ session });

内容的提问来源于stack exchange，提问作者MTA