Mongodb语句性能诊断、调优tuning

线上发现查询mongodb中的用户积分变慢（超过2s），mongodb有和mysql类似的慢查询诊断方法，先开启慢查询：

连上mongodb控制台，开启慢查询功能**：
[root@mongo ~]# mongo
mncg:PRIMARY> use admin
mncg:PRIMARY> db.auth('admin','*****')
mncg:PRIMARY> db.getProfilingLevel()
0
mncg:PRIMARY> db.setProfilingLevel(1,1000)
{ "was" : 0, "slowms" : 1000, "ok" : 1 }
mncg:PRIMARY> db.getProfilingLevel()
1 (已开启)

命令说明：
-- 查看跟踪级别（0为关闭，1是定义慢操作及时间，2是跟踪所有操作）
db.getProfilingLevel()
-- 设置慢操作级别和慢时间定义，单位毫秒（默认100ms）
db.setProfilingLevel(1,1000)
-- 查看当前设置
db.getProfilingStatus()
-- 禁用则设置为0
db.setProfilingLevel(0)
-- 也可以启动服务时设置
mongod --profile=1 --slowms=1000

此时我们从mongodb日志文件（看mongod.conf的中配置，我这里是/opt/mongodb-3.2.5/mongod.log

）中可以看到哪些超过1秒的慢查询，观察发现有以下慢查询：

2017-04-01T11:09:29.624+0800 I COMMAND [conn707607] query scoredb.user_score query: { orderby:score:−1,query: { orgId: 10, createDay: 20170331 } } planSummary: COLLSCAN, COLLSCAN cursorid:395774433451 ntoreturn:30 ntoskip:0 nscanned:0 nscannedObjects:231417 scanAndOrder:1 keyUpdates:0 writeConflicts:0 numYields:2092 nreturned:30 reslen:13697 locks:{ Global: { acquireCount: { r: 4186 } }, MMAPV1Journal: { acquireCount: { r: 2093 }, acquireWaitCount: { r: 1 }, timeAcquiringMicros: { r: 8825 } }, Database: { acquireCount: { r: 2093 } }, Collection: { acquireCount: { R: 2093 } } } 2284ms

发现scoredb.user_score query: { orderby:score:−1,query: { orgId: 10, createDay: 20170331 } } 这个耗时2.2s，explain分析一下这条查询:

db.user_score.find({ orgId: 10, createDay: 20170331}).sort({score: -1}).explain();

{
"queryPlanner" : {
"plannerVersion" : 1,
"namespace" : "scoredb.user_score",
"indexFilterSet" : false,
"parsedQuery" : {
"$and" : [
{
"createDay" : {
"$eq" : 20170331
}
},
{
"orgId" : {
"$eq" : 10
}
}
]
},
"winningPlan" : {
"stage" : "SORT",
"sortPattern" : {
"score" : -1
},
"inputStage" : {
"stage" : "COLLSCAN",
"filter" : {
"$and" : [
{
"createDay" : {
"$eq" : 20170331
}
},
{
"orgId" : {
"$eq" : 10
}
}
]
},
"direction" : "forward"
}
},
"rejectedPlans" : [ ]
}
"ok" : 1
}

从上面的解析结果可以看出，这条语句没有用到索引，创建索引：

db.user_score.ensureIndex({“orgId”:1, createDay: 1, score: -1},{background:1,unique:0,”name”:”idx_user_score_score_order”});

{
"queryPlanner" : {
"plannerVersion" : 1,
"namespace" : "scoredb.user_score",
"indexFilterSet" : false,
"parsedQuery" : {
"$and" : [
{
"createDay" : {
"$eq" : 20170331
}
},
{
"orgId" : {
"$eq" : 10
}
}
]
},
"winningPlan" : {
"stage" : "FETCH",
"inputStage" : {
"stage" : "IXSCAN",
"keyPattern" : {
"orgId" : 1,
"createDay" : 1,
"score" : -1
},
"indexName" : "idx_user_score_score_order",
"isMultiKey" : false,
"direction" : "forward",
"indexBounds" : {
"orgId" : [
"[10.0, 10.0]"
],
"createDay" : [
"[20170331.0, 20170331.0]"
],
"score" : [
"[MaxKey, MinKey]"
]
}
}
},
"rejectedPlans" : [ ]
},
"ok" : 1
}

再次explain后，该语句已用上索引idx_user_score_score_order，再次访问接口时已非常快（20ms）！

附：更多诊断命令

1. mongostat：

查看mongoDB实例的统计信息 字段说明:

insert: 每秒插入量

query: 每秒查询量

update: 每秒更新量

delete: 每秒删除量

locked: 锁定量

qr | qw: 客户端查询排队长度(读|写)

ar | aw: 活跃客户端量(读|写)

conn: 连接数

time: 当前时间

它每秒钟刷新一次状态值，提供良好的可读性，通过这些参数可以观察到一个整体的性能情况。

[root@mongo bin]# ./mongostat -h 127.0.0.1 -u user -p p**** –authenticationDatabase=admin

insert query update delete getmore command flushes mapped vsize res faults qr|qw ar|aw netIn netOut conn set repl time

*0 *0 *0 *0 0 1|0 0 992.0M 2.4G 166.0M 0 0|0 0|0 79b 13k 5 yongPRI 2017-04-01T13:42:27+08:00

*0    *0     *0     *0       0     3|0       0 992.0M  2.4G 166.0M      0   0|0   0|0  195b    13k    5 yongPRI 2017-04-01T13:42:28+08:00
*0    *0     *0     *0       0     1|0       0 992.0M  2.4G 166.0M      0   0|0   0|0   79b    13k    5 yongPRI 2017-04-01T13:42:29+08:00
*0    *0     *0     *0       0     3|0       0 992.0M  2.4G 166.0M      0   0|0   0|0  195b    13k    5 yongPRI 2017-04-01T13:42:30+08:00
*0    *0     *0     *0       0     1|0       0 992.0M  2.4G 166.0M      0   0|0   0|0   79b    13k    5 yongPRI 2017-04-01T13:42:31+08:00
*0    *0     *0     *0       0     3|0       0 992.0M  2.4G 166.0M      0   0|0   0|0  195b    13k    5 yongPRI 2017-04-01T13:42:32+08:00

2.db.serverStatus

查看实例运行状态

{
"host" : "mongo.xxx",
"advisoryHostFQDNs" : [
"mongo.mg.addr"
],
"version" : "3.2.5",
"process" : "mongod",
"pid" : NumberLong(2612),
"uptime" : 4507921,
"uptimeMillis" : NumberLong("4507920428"),
"uptimeEstimate" : 4400471,
"localTime" : ISODate("2017-04-01T06:27:47.470Z"),
"asserts" : {
"regular" : 0,
"warning" : 0,
"msg" : 0,
"user" : 1216,
"rollovers" : 0
},
"backgroundFlushing" : {
"flushes" : 75131,
"total_ms" : 1292618,
"average_ms" : 17.204855519026765,
"last_ms" : 29,
"last_finished" : ISODate("2017-04-01T06:27:21.165Z")
},
"connections" : {
"current" : 5,
"available" : 814,
"totalCreated" : NumberLong(169079)
}, ...
}

3.db.stats

db.stats 查看数据库状态信息：

yong:PRIMARY> db.stats()
{
"db" : "scoredb",
"collections" : 19,
"objects" : 256863,
"avgObjSize" : 366.16151022140207,
"dataSize" : 94053344,
"storageSize" : 141041664,
"numExtents" : 75,
"indexes" : 25,
"indexSize" : 8723792,
"fileSize" : 201326592,
"nsSizeMB" : 16,
"extentFreeList" : {
"num" : 0,
"totalSize" : 0
},
"dataFileVersion" : {
"major" : 4,
"minor" : 22
},
"ok" : 1
}