MongoDB CPU占用率高的原因分析

1.通过查找Stack Overflow找到相关问题的解答

2.跟索引相关

3.使用db.currenyOp();命令得到正在执行的命令(增删改查)以及通过ns来查看是否正在使用索引'ns':'collections.system.indexes';            

4.同时,查看MongoDB log日志 MongoDB日志器会打印索引查询时间超过100ms的操作,发现当前数据库打印的数据查询超过400m

5.2018-06-28T16:51:08.275+0800 I COMMAND  [conn30] command bws.block command: find { find: "block", filter: { blockhash: "378bc0ec06bc22534f2967db5b2b1855dc9776596a7e6b46ad76805aad71d6e3" }, sort: { blockheight: -1 }, limit: 1 } planSummary: COLLSCAN keysExamined:0 docsExamined:376638 hasSortStage:1 cursorExhausted:1 numYields:2944 nreturned:1 reslen:4594 locks:{ Global: { acquireCount: { r: 5890 } }, Database: { acquireCount: { r: 2945 } }, Collection: { acquireCount: { r: 2945 } } } protocol:op_query 396ms

6.从日志得出查询时间较长的语句,单独执行db.collections.find({条件}).sort({value:-1}).limit(1).explain('executionStatus');得出查询详细信息。其中totalDocsExamined指的是MongoDB扫描的文档数量。

{

"queryPlanner" : {

"plannerVersion" : 1,

"namespace" : "bws.block",

"indexFilterSet" : false,

"parsedQuery" : {

"blockhash" : {

"$eq" : "378bc0ec06bc22534f2967db5b2b1855dc9776596a7e6b46ad76805aad71d6e3"

}

},

"winningPlan" : {

"stage" : "SORT",

"sortPattern" : {

"blockheight" : -1

},

"limitAmount" : 1,

"inputStage" : {

"stage" : "SORT_KEY_GENERATOR",

"inputStage" : {

"stage" : "COLLSCAN",

"filter" : {

"blockhash" : {

"$eq" : "378bc0ec06bc22534f2967db5b2b1855dc9776596a7e6b46ad76805aad71d6e3"

}

},

"direction" : "forward"

}

}

},

"rejectedPlans" : [ ]

},

"executionStats" : {

"executionSuccess" : true,

"nReturned" : 1,

"executionTimeMillis" : 205,

"totalKeysExamined" : 0,

"totalDocsExamined" : 403964,

"executionStages" : {

"stage" : "SORT",

"nReturned" : 1,

"executionTimeMillisEstimate" : 171,

"works" : 403969,

"advanced" : 1,

"needTime" : 403967,

"needYield" : 0,

"saveState" : 3156,

"restoreState" : 3156,

"isEOF" : 1,

"invalidates" : 0,

"sortPattern" : {

"blockheight" : -1

},

"memUsage" : 4502,

"memLimit" : 33554432,

"limitAmount" : 1,

"inputStage" : {

"stage" : "SORT_KEY_GENERATOR",

"nReturned" : 1,

"executionTimeMillisEstimate" : 161,

"works" : 403967,

"advanced" : 1,

"needTime" : 403965,

"needYield" : 0,

"saveState" : 3156,

"restoreState" : 3156,

"isEOF" : 1,

"invalidates" : 0,

"inputStage" : {

"stage" : "COLLSCAN",

"filter" : {

"blockhash" : {

"$eq" : "378bc0ec06bc22534f2967db5b2b1855dc9776596a7e6b46ad76805aad71d6e3"

}

},

"nReturned" : 1,

"executionTimeMillisEstimate" : 161,

"works" : 403966,

"advanced" : 1,

"needTime" : 403964,

"needYield" : 0,

"saveState" : 3156,

"restoreState" : 3156,

"isEOF" : 1,

"invalidates" : 0,

"direction" : "forward",

"docsExamined" : 403964

}

}

}

},

"serverInfo" : {

"host" : "MwalletDB",

"port" : 27017,

"version" : "3.4.2",

"gitVersion" : "3f76e40c105fc223b3e5aac3e20dcd026b83b38b"

},

"ok" : 1

}

7.由于是生产环境无法停止数据库的访问,所以准备后台构建索引。命令为:db.collections.createIndex({value:1},{background:true})

你可能感兴趣的:(MongoDB CPU占用率高的原因分析)