概要
由于历史原因，一些老项目还在用MongoDB V3.2版本，集群为分片模式，一个客户表数量有15亿左右，昨天监控突然报很多慢查询，如下：
db.info.count({domain_id:888,status:{$in:[2,3]}})
表结构关键字段如下：
| 字段 | 类型 | 注释 |
|---|---|---|
| domain_id | int64 | 公司id |
| customer_id | int64 | 公司下的客户id |
| status | int32 | 客户状态 |
索引如下：
db.info.createIndex({domain_id:1,status:1,modify_time:1})
一、原因
遇到慢查询，当然要explain下看下查询语句执行状况了，如下：
db.info.explain('executionStats').count({domain_id:4730422,status:{$in:[2,3]}})
{
"queryPlanner" : {
"plannerVersion" : 1,
"namespace" : "customer.info",
"indexFilterSet" : false,
"parsedQuery" : {
"$and" : [
{
"domain_id" : {
"$eq" : 4730422
}
},
{
"status" : {
"$in" : [
2,
3
]
}
}
]
},
"winningPlan" : {
"stage" : "COUNT",
"inputStage" : {
"stage" : "FETCH",
"inputStage" : {
"stage" : "IXSCAN",
"keyPattern" : {
"domain_id" : 1,
"status" : 1,
"modify_time" : -1
},
"indexName" : "domain_id_1_status_1_modify_time_-1",
"isMultiKey" : false,
"direction" : "forward",
"indexBounds" : {
"domain_id" : [
"[4730422.0, 4730422.0]"
],
"status" : [
"[2.0, 2.0]",
"[3.0, 3.0]"
],
"modify_time" : [
"[MaxKey, MinKey]"
]
}
}
}
},
"rejectedPlans" : []
},
"executionStats" : {
"executionSuccess" : true,
"nReturned" : 0,
"executionTimeMillis" : 0,
"totalKeysExamined" : 9, #索引扫描个数
"totalDocsExamined" : 9, #文档扫描个数，不等于0表示有回表
"executionStages" : {
"stage" : "COUNT",
"nReturned" : 0,
"executionTimeMillisEstimate" : 0,
"works" : 10,
"advanced" : 0,
"needTime" : 9,
"needFetch" : 0,
"saveState" : 0,
"restoreState" : 0,
"isEOF" : 1,
"invalidates" : 0,
"nCounted" : 9,
"nSkipped" : 0,
"inputStage" : {
"stage" : "FETCH",
"nReturned" : 9,
"executionTimeMillisEstimate" : 0,
"works" : 10,
"advanced" : 9,
"needTime" : 0,
"needFetch" : 0,
"saveState" : 0,
"restoreState" : 0,
"isEOF" : 1,
"invalidates" : 0,
"docsExamined" : 9,
"alreadyHasObj" : 0,
"inputStage" : {
"stage" : "IXSCAN",
"nReturned" : 9,
"executionTimeMillisEstimate" : 0,
"works" : 10,
"advanced" : 9,
"needTime" : 0,
"needFetch" : 0,
"saveState" : 0,
"restoreState" : 0,
"isEOF" : 1,
"invalidates" : 0,
"keyPattern" : {
"domain_id" : 1,
"status" : 1,
"modify_time" : -1
},
"indexName" : "domain_id_1_status_1_modify_time_-1",
"isMultiKey" : false,
"direction" : "forward",
"indexBounds" : {
"domain_id" : [
"[4730422.0, 4730422.0]"
],
"status" : [
"[2.0, 2.0]",
"[3.0, 3.0]"
],
"modify_time" : [
"[MaxKey, MinKey]"
]
},
"keysExamined" : 9,
"dupsTested" : 0,
"dupsDropped" : 0,
"seenInvalidated" : 0,
"matchTested" : 0
}
}
}
},
"ok" : 1
}
通过上面索引执行分析可以看到 totalDocsExamined不为0，所以进行了回表，理论上走了覆盖索引是不用回表的。
我们下面再试试没有$in的情况下，查询分析：
db.info.explain('executionStats').count({domain_id:4730422,status:3})
{
"queryPlanner" : {
"plannerVersion" : 1,
"namespace" : "customer.info",
"indexFilterSet" : false,
"parsedQuery" : {
"$and" : [
{
"domain_id" : {
"$eq" : 4730422
}
},
{
"status" : {
"$eq" : 3
}
}
]
},
"winningPlan" : {
"stage" : "COUNT",
"inputStage" : {
"stage" : "COUNT_SCAN",
"keyPattern" : {
"domain_id" : 1,
"status" : 1,
"modify_time" : -1
},
"indexName" : "domain_id_1_status_1_modify_time_-1",
"isMultiKey" : false
}
},
"rejectedPlans" : [ ]
},
"executionStats" : {
"executionSuccess" : true,
"nReturned" : 0,
"executionTimeMillis" : 0,
"totalKeysExamined" : 8,
"totalDocsExamined" : 0,
"executionStages" : {
"stage" : "COUNT",
"nReturned" : 0,
"executionTimeMillisEstimate" : 0,
"works" : 8,
"advanced" : 0,
"needTime" : 7,
"needFetch" : 0,
"saveState" : 0,
"restoreState" : 0,
"isEOF" : 1,
"invalidates" : 0,
"nCounted" : 6,
"nSkipped" : 0,
"inputStage" : {
"stage" : "COUNT_SCAN",
"nReturned" : 6,
"executionTimeMillisEstimate" : 0,
"works" : 7,
"advanced" : 6,
"needTime" : 1,
"needFetch" : 0,
"saveState" : 0,
"restoreState" : 0,
"isEOF" : 1,
"invalidates" : 0,
"keysExamined" : 8,
"keyPattern" : {
"domain_id" : 1,
"status" : 1,
"modify_time" : -1
},
"indexName" : "domain_id_1_status_1_modify_time_-1",
"isMultiKey" : false
}
},
"allPlansExecution" : [ ]
},
"ok" : 1
}
可以看到此时totalDocsExamined的值为0，说明不使用$in操作就不会回表了。这显然是count操作在$in时的bug。
我们用docker快速搭建一个MongoDB V4.4的副本集试试，如下：
rsm:PRIMARY> db.info.explain('executionStats').count({domain_id:15600,status:{$in:[2,3]}})
{
"queryPlanner" : {
"plannerVersion" : 1,
"namespace" : "customer.info",
"indexFilterSet" : false,
"parsedQuery" : {
"$and" : [
{
"domain_id" : {
"$eq" : 15600
}
},
{
"status" : {
"$in" : [
2,
3
]
}
}
]
},
"winningPlan" : {
"stage" : "COUNT",
"inputStage" : {
"stage" : "IXSCAN",
"keyPattern" : {
"domain_id" : 1,
"status" : 1
},
"indexName" : "domain_id_1_status_1",
"isMultiKey" : false,
"multiKeyPaths" : {
"domain_id" : [ ],
"status" : [ ]
},
"isUnique" : false,
"isSparse" : false,
"isPartial" : false,
"indexVersion" : 2,
"direction" : "forward",
"indexBounds" : {
"domain_id" : [
"[15600.0, 15600.0]"
],
"status" : [
"[2.0, 2.0]",
"[3.0, 3.0]"
]
}
}
},
"rejectedPlans" : [ ]
},
"executionStats" : {
"executionSuccess" : true,
"nReturned" : 0,
"executionTimeMillis" : 0,
"totalKeysExamined" : 5,
"totalDocsExamined" : 0,
"executionStages" : {
"stage" : "COUNT",
"nReturned" : 0,
"executionTimeMillisEstimate" : 0,
"works" : 6,
"advanced" : 0,
"needTime" : 5,
"needYield" : 0,
"saveState" : 0,
"restoreState" : 0,
"isEOF" : 1,
"nCounted" : 5,
"nSkipped" : 0,
"inputStage" : {
"stage" : "IXSCAN",
"nReturned" : 5,
"executionTimeMillisEstimate" : 0,
"works" : 6,
"advanced" : 5,
"needTime" : 0,
"needYield" : 0,
"saveState" : 0,
"restoreState" : 0,
"isEOF" : 1,
"keyPattern" : {
"domain_id" : 1,
"status" : 1
},
"indexName" : "domain_id_1_status_1",
"isMultiKey" : false,
"multiKeyPaths" : {
"domain_id" : [ ],
"status" : [ ]
},
"isUnique" : false,
"isSparse" : false,
"isPartial" : false,
"indexVersion" : 2,
"direction" : "forward",
"indexBounds" : {
"domain_id" : [
"[15600.0, 15600.0]"
],
"status" : [
"[2.0, 2.0]",
"[3.0, 3.0]"
]
},
"keysExamined" : 5,
"seeks" : 1,
"dupsTested" : 0,
"dupsDropped" : 0
}
}
},
"serverInfo" : {
"host" : "b8b178041a4e",
"port" : 27021,
"version" : "4.4.28",
"gitVersion" : "61c2baf63a060f7c12bd76e779044800ae18710b"
},
"ok" : 1,
"$clusterTime" : {
"clusterTime" : Timestamp(1711857956, 1),
"signature" : {
"hash" : BinData(0,"AAAAAAAAAAAAAAAAAAAAAAAAAAA="),
"keyId" : NumberLong(0)
}
},
"operationTime" : Timestamp(1711857956, 1)
}
可以看到此时totalDocsExamined的值为0，说明至少MongoDB V4.4以后该bug修复了。
二、解决方法
1：升级MongoDB版本
2：既然不使用$in就可以，那就count两次，程序中累加起来。
由此可以看出，有条件的话，项目的基础组件及时升级也是很有必要的，新的版本肯定是修复了以往bug、做了性能优化以及功能新增的，好处多多。否则项目必然会慢慢腐化。
5363
被折叠的 条评论
为什么被折叠?
到【灌水乐园】发言
