Question

我刚开始使用mongodb。我使用ssd stoarge处理38GB数据集（6800万个文档）。但性能是通过索引完成的，也没有索引。它使用如此多的ram进行简单的查找查询，有两个字段，没有cpu使用。

花了18分钟才能获得160万条记录。有什么因素有助于提高单节点的mongodb性能？

我的文档看起来像这样：

{ "_id" : ObjectId("55e7eec02756dd0f1e693b72"), 
  "categorieId" : 2, 
  "title" : "AntiMalware",  
  "messageValue" : " #\"Antimalware: \"Windows Defender\" is Not Updated and Running\"#",  
  "timestamp" : "8/19/2015 11:06:24 AM",  
  "resultStatusId" : 2,  
  "messageFormat" : "Text",  
  "titleId" : 1,  
  "resultStatus" : "Warning",  
  "antiMalwareName" : "Comodo Antivirus",  
  "categories" : "Security" }

我的索引位于titleId和resultStatusId。

我的查询是：

   db.collection.find({"titleId":21, resultStatusId:1})

解释输出是：

{
    "queryPlanner" : {
        "plannerVersion" : 1,
        "namespace" : "techHealLogAnalysis.techHealTestLogData",
        "indexFilterSet" : false,
        "parsedQuery" : {
            "$and" : [
                {
                    "resultStatusId" : {
                        "$eq" : 1
                    }
                },
                {
                    "titleId" : {
                        "$eq" : 21
                    }
                }
            ]
        },
        "winningPlan" : {
            "stage" : "FETCH",
            "inputStage" : {
                "stage" : "IXSCAN",
                "keyPattern" : {
                    "titleId" : 1,
                    "resultStatusId" : 1
                },
                "indexName" : "titleId_1_resultStatusId_1",
                "isMultiKey" : false,
                "direction" : "forward",
                "indexBounds" : {
                    "titleId" : [
                        "[21.0, 21.0]"
                    ],
                    "resultStatusId" : [
                        "[1.0, 1.0]"
                    ]
                }
            }
        },
        "rejectedPlans" : [ ]
    },
    "executionStats" : {
        "executionSuccess" : true,
        "nReturned" : 1671842,
        "executionTimeMillis" : 1108805,
        "totalKeysExamined" : 1671842,
        "totalDocsExamined" : 1671842,
        "executionStages" : {
            "stage" : "FETCH",
            "nReturned" : 1671842,
            "executionTimeMillisEstimate" : 177670,
            "works" : 2143234,
            "advanced" : 1671842,
            "needTime" : 0,
            "needFetch" : 471391,
            "saveState" : 471391,
            "restoreState" : 471391,
            "isEOF" : 1,
            "invalidates" : 0,
            "docsExamined" : 1671842,
            "alreadyHasObj" : 0,
            "inputStage" : {
                "stage" : "IXSCAN",
                "nReturned" : 1671842,
                "executionTimeMillisEstimate" : 1470,
                "works" : 1671843,
                "advanced" : 1671842,
                "needTime" : 0,
                "needFetch" : 0,
                "saveState" : 471391,
                "restoreState" : 471391,
                "isEOF" : 1,
                "invalidates" : 0,
                "keyPattern" : {
                    "titleId" : 1,
                    "resultStatusId" : 1
                },
                "indexName" : "titleId_1_resultStatusId_1",
                "isMultiKey" : false,
                "direction" : "forward",
                "indexBounds" : {
                    "titleId" : [
                        "[21.0, 21.0]"
                    ],
                    "resultStatusId" : [
                        "[1.0, 1.0]"
                    ]
                },
                "keysExamined" : 1671842,
                "dupsTested" : 0,
                "dupsDropped" : 0,
                "seenInvalidated" : 0,
                "matchTested" : 0
            }
        }
    },
    "serverInfo" : {
        "host" : "instance-7",
        "port" : 27017,
        "version" : "3.0.6",
        "gitVersion" : "1ef45a23a4c5e3480ac919b28afcba3c615488f2"
    },
    "ok" : 1
}

Answer 1

具有大型数据集和高吞吐量应用程序的数据库系统可能会挑战单个服务器的容量。较大的数据集超过了单台机器的存储容量。最后，大于系统RAM的工作集大小会强调磁盘驱动器的I / O容量。为您的案例部署分片可能非常有用。查看以下链接后。

http://docs.mongodb.org/manual/core/sharding-introduction/

如何用大数据集的单节点提高mongodb性能？

1 个答案: