mongodb三节点同时创建副本集出现的bug
2014-04-04 15:11
387 查看
节点一：10.4.4.4:27017
节点二：10.4.4.5:27017
节点三：10.4.4.6:27017
创建三个节点组成的副本集，竟然产生以下结果：
./mongo 10.4.4.4:27017
查看rs
rs.status()
{
"set" : "mongodb_set",
"date" : ISODate("2014-04-04T06:44:10Z"),
"myState" : 1,
"members" : [
{
"_id" : 0,
"name" : "10.4.4.4:27017",
"health" : 1,
"state" : 1,
"stateStr" : "PRIMARY",
"uptime" : 903,
"optime" : Timestamp(1396592962, 1),
"optimeDate" : ISODate("2014-04-04T06:29:22Z"),
"self" : true
},
{
"_id" : 1,
"name" : "10.4.4.5:27017",
"health" : 1,
"state" : 1,
"stateStr" : "PRIMARY",
"uptime" : 883,
"optime" : Timestamp(1396593341, 1416),
"optimeDate" : ISODate("2014-04-04T06:35:41Z"),
"lastHeartbeat" : ISODate("2014-04-04T06:44:09Z"),
"lastHeartbeatRecv" : ISODate("1970-01-01T00:00:00Z"),
"pingMs" : 0
}
],
"ok" : 1
}
这个状态之一存在，直到关闭当前连接。
./mongo 10.4.4.5:27017
查看rs
rs.status()
{
"set" : "mongodb_set",
"date" : ISODate("2014-04-04T06:43:45Z"),
"myState" : 1,
"members" : [
{
"_id" : 0,
"name" : "10.4.4.5:27017",
"health" : 1,
"state" : 1,
"stateStr" : "PRIMARY",
"uptime" : 878,
"optime" : Timestamp(1396593341, 1416),
"optimeDate" : ISODate("2014-04-04T06:35:41Z"),
"self" : true
},
{
"_id" : 1,
"name" : "10.4.4.6:27017",
"health" : 1,
"state" : 2,
"stateStr" : "SECONDARY",
"uptime" : 858,
"optime" : Timestamp(1396593341, 1416),
"optimeDate" : ISODate("2014-04-04T06:35:41Z"),
"lastHeartbeat" : ISODate("2014-04-04T06:43:45Z"),
"lastHeartbeatRecv" : ISODate("2014-04-04T06:43:45Z"),
"pingMs" : 0,
"syncingTo" : "10.4.4.5:27017"
}
],
"ok" : 1
}
以上现象非常奇怪，副本集id都是一致的，结果产生了两个副本集，都是独立的，而且更奇怪的是，其中有个一副本集具有两个primary。
测试以下两个副本集是否相通：
在第二个副本集上执行数据插入
for(i=1;i<10000;++i){
db.person.insert({
id: i,
name: "mongodb"+i
} )
}
在第二个副本集上查询，结果：
>db.person.find()
{ "_id" : ObjectId("533e51da0fb65d4623fe3607"), "id" : 1, "name" : "mongodb1" }
{ "_id" : ObjectId("533e51da0fb65d4623fe3608"), "id" : 2, "name" : "mongodb2" }
{ "_id" : ObjectId("533e51da0fb65d4623fe3609"), "id" : 3, "name" : "mongodb3" }
{ "_id" : ObjectId("533e51da0fb65d4623fe360a"), "id" : 4, "name" : "mongodb4" }
{ "_id" : ObjectId("533e51da0fb65d4623fe360b"), "id" : 5, "name" : "mongodb5" }
{ "_id" : ObjectId("533e51da0fb65d4623fe360c"), "id" : 6, "name" : "mongodb6" }
{ "_id" : ObjectId("533e51da0fb65d4623fe360d"), "id" : 7, "name" : "mongodb7" }
{ "_id" : ObjectId("533e51da0fb65d4623fe360e"), "id" : 8, "name" : "mongodb8" }
{ "_id" : ObjectId("533e51da0fb65d4623fe360f"), "id" : 9, "name" : "mongodb9" }
{ "_id" : ObjectId("533e51da0fb65d4623fe3610"), "id" : 10, "name" : "mongodb10" }
{ "_id" : ObjectId("533e51da0fb65d4623fe3611"), "id" : 11, "name" : "mongodb11" }
{ "_id" : ObjectId("533e51da0fb65d4623fe3612"), "id" : 12, "name" : "mongodb12" }
{ "_id" : ObjectId("533e51da0fb65d4623fe3613"), "id" : 13, "name" : "mongodb13" }
{ "_id" : ObjectId("533e51da0fb65d4623fe3614"), "id" : 14, "name" : "mongodb14" }
{ "_id" : ObjectId("533e51da0fb65d4623fe3615"), "id" : 15, "name" : "mongodb15" }
{ "_id" : ObjectId("533e51da0fb65d4623fe3616"), "id" : 16, "name" : "mongodb16" }
{ "_id" : ObjectId("533e51da0fb65d4623fe3617"), "id" : 17, "name" : "mongodb17" }
{ "_id" : ObjectId("533e51da0fb65d4623fe3618"), "id" : 18, "name" : "mongodb18" }
{ "_id" : ObjectId("533e51da0fb65d4623fe3619"), "id" : 19, "name" : "mongodb19" }
{ "_id" : ObjectId("533e51da0fb65d4623fe361a"), "id" : 20, "name" : "mongodb20" }
Type "it" for more
而同时在第一个副本集上查询，结果为空，表明它与第二副本集数据不是相通的。
两个副本集同时关闭，再重启，
./mongo 10.4.4.4:27017
查看rs
>rs.status()
{
"set" : "mongodb_set",
"date" : ISODate("2014-04-04T06:58:22Z"),
"myState" : 2,
"syncingTo" : "10.4.4.5:27017",
"members" : [
{
"_id" : 0,
"name" : "10.4.4.4:27017",
"health" : 1,
"state" : 2,
"stateStr" : "SECONDARY",
"uptime" : 135,
"optime" : Timestamp(1396594084, 4268),
"optimeDate" : ISODate("2014-04-04T06:48:04Z"),
"self" : true
},
{
"_id" : 1,
"name" : "10.4.4.5:27017",
"health" : 1,
"state" : 1,
"stateStr" : "PRIMARY",
"uptime" : 133,
"optime" : Timestamp(1396594084, 4268),
"optimeDate" : ISODate("2014-04-04T06:48:04Z"),
"lastHeartbeat" : ISODate("2014-04-04T06:58:21Z"),
"lastHeartbeatRecv" : ISODate("1970-01-01T00:00:00Z"),
"pingMs" : 0
}
],
"ok" : 1
}
注意到，这时候副本集一由当初的两个primary状态转变为正常的primary和secondary。由于现在连接的secondary节点，无法正常查询。
执行
>db.getMongo().setSlaveOk();
>db.person.find()
{ "_id" : ObjectId("533e51da0fb65d4623fe3607"), "id" : 1, "name" : "mongodb1" }
{ "_id" : ObjectId("533e51da0fb65d4623fe3608"), "id" : 2, "name" : "mongodb2" }
{ "_id" : ObjectId("533e51da0fb65d4623fe3609"), "id" : 3, "name" : "mongodb3" }
{ "_id" : ObjectId("533e51da0fb65d4623fe360a"), "id" : 4, "name" : "mongodb4" }
{ "_id" : ObjectId("533e51da0fb65d4623fe360b"), "id" : 5, "name" :
4000
"mongodb5" }
{ "_id" : ObjectId("533e51da0fb65d4623fe360c"), "id" : 6, "name" : "mongodb6" }
{ "_id" : ObjectId("533e51da0fb65d4623fe360d"), "id" : 7, "name" : "mongodb7" }
{ "_id" : ObjectId("533e51da0fb65d4623fe360e"), "id" : 8, "name" : "mongodb8" }
{ "_id" : ObjectId("533e51da0fb65d4623fe360f"), "id" : 9, "name" : "mongodb9" }
{ "_id" : ObjectId("533e51da0fb65d4623fe3610"), "id" : 10, "name" : "mongodb10" }
{ "_id" : ObjectId("533e51da0fb65d4623fe3611"), "id" : 11, "name" : "mongodb11" }
{ "_id" : ObjectId("533e51da0fb65d4623fe3612"), "id" : 12, "name" : "mongodb12" }
{ "_id" : ObjectId("533e51da0fb65d4623fe3613"), "id" : 13, "name" : "mongodb13" }
{ "_id" : ObjectId("533e51da0fb65d4623fe3614"), "id" : 14, "name" : "mongodb14" }
{ "_id" : ObjectId("533e51da0fb65d4623fe3615"), "id" : 15, "name" : "mongodb15" }
{ "_id" : ObjectId("533e51da0fb65d4623fe3616"), "id" : 16, "name" : "mongodb16" }
{ "_id" : ObjectId("533e51da0fb65d4623fe3617"), "id" : 17, "name" : "mongodb17" }
{ "_id" : ObjectId("533e51da0fb65d4623fe3618"), "id" : 18, "name" : "mongodb18" }
{ "_id" : ObjectId("533e51da0fb65d4623fe3619"), "id" : 19, "name" : "mongodb19" }
{ "_id" : ObjectId("533e51da0fb65d4623fe361a"), "id" : 20, "name" : "mongodb20" }
Type "it" for more
这时候可以查找到数据。
./mongo 10.4.4.5:27017
查看rs
rs.status()
{
"set" : "mongodb_set",
"date" : ISODate("2014-04-04T06:56:30Z"),
"myState" : 1,
"members" : [
{
"_id" : 0,
"name" : "10.4.4.5:27017",
"health" : 1,
"state" : 1,
"stateStr" : "PRIMARY",
"uptime" : 22,
"optime" : Timestamp(1396594084, 4268),
"optimeDate" : ISODate("2014-04-04T06:48:04Z"),
"self" : true
},
{
"_id" : 1,
"name" : "10.4.4.6:27017",
"health" : 1,
"state" : 2,
"stateStr" : "SECONDARY",
"uptime" : 20,
"optime" : Timestamp(1396594084, 4268),
"optimeDate" : ISODate("2014-04-04T06:48:04Z"),
"lastHeartbeat" : ISODate("2014-04-04T06:56:30Z"),
"lastHeartbeatRecv" : ISODate("2014-04-04T06:56:30Z"),
"pingMs" : 0,
"lastHeartbeatMessage" : "syncing to: 10.4.4.5:27017",
"syncingTo" : "10.4.4.5:27017"
}
],
"ok" : 1
}
副本集二没有变化。
现在两个副本集的数据相通。但是三个节点共有同一个副本集id，结果实际上分属于两个不同副本集。
现在从副本集一上的10.4.4.4:27017节点退出，改为10.4.4.5:27017登录，结果又如何？
./mongo 10.4.4.5:27017
查看rs
>rs.status()
{
"set" : "mongodb_set",
"date" : ISODate("2014-04-04T07:08:08Z"),
"myState" : 1,
"members" : [
{
"_id" : 0,
"name" : "10.4.4.5:27017",
"health" : 1,
"state" : 1,
"stateStr" : "PRIMARY",
"uptime" : 720,
"optime" : Timestamp(1396594084, 4268),
"optimeDate" : ISODate("2014-04-04T06:48:04Z"),
"self" : true
},
{
"_id" : 1,
"name" : "10.4.4.6:27017",
"health" : 1,
"state" : 2,
"stateStr" : "SECONDARY",
"uptime" : 718,
"optime" : Timestamp(1396594084, 4268),
"optimeDate" : ISODate("2014-04-04T06:48:04Z"),
"lastHeartbeat" : ISODate("2014-04-04T07:08:06Z"),
"lastHeartbeatRecv" : ISODate("2014-04-04T07:08:06Z"),
"pingMs" : 0,
"syncingTo" : "10.4.4.5:27017"
}
],
"ok" : 1
}
可见当前连接的是副本集二的primary，此时副本集一的primary去哪里了呢？似乎只能通过10.4.4.4:27017来访问secondary了。
这个难道是mongodb的一个bug？
节点二：10.4.4.5:27017
节点三：10.4.4.6:27017
创建三个节点组成的副本集，竟然产生以下结果：
./mongo 10.4.4.4:27017
查看rs
rs.status()
{
"set" : "mongodb_set",
"date" : ISODate("2014-04-04T06:44:10Z"),
"myState" : 1,
"members" : [
{
"_id" : 0,
"name" : "10.4.4.4:27017",
"health" : 1,
"state" : 1,
"stateStr" : "PRIMARY",
"uptime" : 903,
"optime" : Timestamp(1396592962, 1),
"optimeDate" : ISODate("2014-04-04T06:29:22Z"),
"self" : true
},
{
"_id" : 1,
"name" : "10.4.4.5:27017",
"health" : 1,
"state" : 1,
"stateStr" : "PRIMARY",
"uptime" : 883,
"optime" : Timestamp(1396593341, 1416),
"optimeDate" : ISODate("2014-04-04T06:35:41Z"),
"lastHeartbeat" : ISODate("2014-04-04T06:44:09Z"),
"lastHeartbeatRecv" : ISODate("1970-01-01T00:00:00Z"),
"pingMs" : 0
}
],
"ok" : 1
}
这个状态之一存在，直到关闭当前连接。
./mongo 10.4.4.5:27017
查看rs
rs.status()
{
"set" : "mongodb_set",
"date" : ISODate("2014-04-04T06:43:45Z"),
"myState" : 1,
"members" : [
{
"_id" : 0,
"name" : "10.4.4.5:27017",
"health" : 1,
"state" : 1,
"stateStr" : "PRIMARY",
"uptime" : 878,
"optime" : Timestamp(1396593341, 1416),
"optimeDate" : ISODate("2014-04-04T06:35:41Z"),
"self" : true
},
{
"_id" : 1,
"name" : "10.4.4.6:27017",
"health" : 1,
"state" : 2,
"stateStr" : "SECONDARY",
"uptime" : 858,
"optime" : Timestamp(1396593341, 1416),
"optimeDate" : ISODate("2014-04-04T06:35:41Z"),
"lastHeartbeat" : ISODate("2014-04-04T06:43:45Z"),
"lastHeartbeatRecv" : ISODate("2014-04-04T06:43:45Z"),
"pingMs" : 0,
"syncingTo" : "10.4.4.5:27017"
}
],
"ok" : 1
}
以上现象非常奇怪，副本集id都是一致的，结果产生了两个副本集，都是独立的，而且更奇怪的是，其中有个一副本集具有两个primary。
测试以下两个副本集是否相通：
在第二个副本集上执行数据插入
for(i=1;i<10000;++i){
db.person.insert({
id: i,
name: "mongodb"+i
} )
}
在第二个副本集上查询，结果：
>db.person.find()
{ "_id" : ObjectId("533e51da0fb65d4623fe3607"), "id" : 1, "name" : "mongodb1" }
{ "_id" : ObjectId("533e51da0fb65d4623fe3608"), "id" : 2, "name" : "mongodb2" }
{ "_id" : ObjectId("533e51da0fb65d4623fe3609"), "id" : 3, "name" : "mongodb3" }
{ "_id" : ObjectId("533e51da0fb65d4623fe360a"), "id" : 4, "name" : "mongodb4" }
{ "_id" : ObjectId("533e51da0fb65d4623fe360b"), "id" : 5, "name" : "mongodb5" }
{ "_id" : ObjectId("533e51da0fb65d4623fe360c"), "id" : 6, "name" : "mongodb6" }
{ "_id" : ObjectId("533e51da0fb65d4623fe360d"), "id" : 7, "name" : "mongodb7" }
{ "_id" : ObjectId("533e51da0fb65d4623fe360e"), "id" : 8, "name" : "mongodb8" }
{ "_id" : ObjectId("533e51da0fb65d4623fe360f"), "id" : 9, "name" : "mongodb9" }
{ "_id" : ObjectId("533e51da0fb65d4623fe3610"), "id" : 10, "name" : "mongodb10" }
{ "_id" : ObjectId("533e51da0fb65d4623fe3611"), "id" : 11, "name" : "mongodb11" }
{ "_id" : ObjectId("533e51da0fb65d4623fe3612"), "id" : 12, "name" : "mongodb12" }
{ "_id" : ObjectId("533e51da0fb65d4623fe3613"), "id" : 13, "name" : "mongodb13" }
{ "_id" : ObjectId("533e51da0fb65d4623fe3614"), "id" : 14, "name" : "mongodb14" }
{ "_id" : ObjectId("533e51da0fb65d4623fe3615"), "id" : 15, "name" : "mongodb15" }
{ "_id" : ObjectId("533e51da0fb65d4623fe3616"), "id" : 16, "name" : "mongodb16" }
{ "_id" : ObjectId("533e51da0fb65d4623fe3617"), "id" : 17, "name" : "mongodb17" }
{ "_id" : ObjectId("533e51da0fb65d4623fe3618"), "id" : 18, "name" : "mongodb18" }
{ "_id" : ObjectId("533e51da0fb65d4623fe3619"), "id" : 19, "name" : "mongodb19" }
{ "_id" : ObjectId("533e51da0fb65d4623fe361a"), "id" : 20, "name" : "mongodb20" }
Type "it" for more
而同时在第一个副本集上查询，结果为空，表明它与第二副本集数据不是相通的。
两个副本集同时关闭，再重启，
./mongo 10.4.4.4:27017
查看rs
>rs.status()
{
"set" : "mongodb_set",
"date" : ISODate("2014-04-04T06:58:22Z"),
"myState" : 2,
"syncingTo" : "10.4.4.5:27017",
"members" : [
{
"_id" : 0,
"name" : "10.4.4.4:27017",
"health" : 1,
"state" : 2,
"stateStr" : "SECONDARY",
"uptime" : 135,
"optime" : Timestamp(1396594084, 4268),
"optimeDate" : ISODate("2014-04-04T06:48:04Z"),
"self" : true
},
{
"_id" : 1,
"name" : "10.4.4.5:27017",
"health" : 1,
"state" : 1,
"stateStr" : "PRIMARY",
"uptime" : 133,
"optime" : Timestamp(1396594084, 4268),
"optimeDate" : ISODate("2014-04-04T06:48:04Z"),
"lastHeartbeat" : ISODate("2014-04-04T06:58:21Z"),
"lastHeartbeatRecv" : ISODate("1970-01-01T00:00:00Z"),
"pingMs" : 0
}
],
"ok" : 1
}
注意到，这时候副本集一由当初的两个primary状态转变为正常的primary和secondary。由于现在连接的secondary节点，无法正常查询。
执行
>db.getMongo().setSlaveOk();
>db.person.find()
{ "_id" : ObjectId("533e51da0fb65d4623fe3607"), "id" : 1, "name" : "mongodb1" }
{ "_id" : ObjectId("533e51da0fb65d4623fe3608"), "id" : 2, "name" : "mongodb2" }
{ "_id" : ObjectId("533e51da0fb65d4623fe3609"), "id" : 3, "name" : "mongodb3" }
{ "_id" : ObjectId("533e51da0fb65d4623fe360a"), "id" : 4, "name" : "mongodb4" }
{ "_id" : ObjectId("533e51da0fb65d4623fe360b"), "id" : 5, "name" :
4000
"mongodb5" }
{ "_id" : ObjectId("533e51da0fb65d4623fe360c"), "id" : 6, "name" : "mongodb6" }
{ "_id" : ObjectId("533e51da0fb65d4623fe360d"), "id" : 7, "name" : "mongodb7" }
{ "_id" : ObjectId("533e51da0fb65d4623fe360e"), "id" : 8, "name" : "mongodb8" }
{ "_id" : ObjectId("533e51da0fb65d4623fe360f"), "id" : 9, "name" : "mongodb9" }
{ "_id" : ObjectId("533e51da0fb65d4623fe3610"), "id" : 10, "name" : "mongodb10" }
{ "_id" : ObjectId("533e51da0fb65d4623fe3611"), "id" : 11, "name" : "mongodb11" }
{ "_id" : ObjectId("533e51da0fb65d4623fe3612"), "id" : 12, "name" : "mongodb12" }
{ "_id" : ObjectId("533e51da0fb65d4623fe3613"), "id" : 13, "name" : "mongodb13" }
{ "_id" : ObjectId("533e51da0fb65d4623fe3614"), "id" : 14, "name" : "mongodb14" }
{ "_id" : ObjectId("533e51da0fb65d4623fe3615"), "id" : 15, "name" : "mongodb15" }
{ "_id" : ObjectId("533e51da0fb65d4623fe3616"), "id" : 16, "name" : "mongodb16" }
{ "_id" : ObjectId("533e51da0fb65d4623fe3617"), "id" : 17, "name" : "mongodb17" }
{ "_id" : ObjectId("533e51da0fb65d4623fe3618"), "id" : 18, "name" : "mongodb18" }
{ "_id" : ObjectId("533e51da0fb65d4623fe3619"), "id" : 19, "name" : "mongodb19" }
{ "_id" : ObjectId("533e51da0fb65d4623fe361a"), "id" : 20, "name" : "mongodb20" }
Type "it" for more
这时候可以查找到数据。
./mongo 10.4.4.5:27017
查看rs
rs.status()
{
"set" : "mongodb_set",
"date" : ISODate("2014-04-04T06:56:30Z"),
"myState" : 1,
"members" : [
{
"_id" : 0,
"name" : "10.4.4.5:27017",
"health" : 1,
"state" : 1,
"stateStr" : "PRIMARY",
"uptime" : 22,
"optime" : Timestamp(1396594084, 4268),
"optimeDate" : ISODate("2014-04-04T06:48:04Z"),
"self" : true
},
{
"_id" : 1,
"name" : "10.4.4.6:27017",
"health" : 1,
"state" : 2,
"stateStr" : "SECONDARY",
"uptime" : 20,
"optime" : Timestamp(1396594084, 4268),
"optimeDate" : ISODate("2014-04-04T06:48:04Z"),
"lastHeartbeat" : ISODate("2014-04-04T06:56:30Z"),
"lastHeartbeatRecv" : ISODate("2014-04-04T06:56:30Z"),
"pingMs" : 0,
"lastHeartbeatMessage" : "syncing to: 10.4.4.5:27017",
"syncingTo" : "10.4.4.5:27017"
}
],
"ok" : 1
}
副本集二没有变化。
现在两个副本集的数据相通。但是三个节点共有同一个副本集id，结果实际上分属于两个不同副本集。
现在从副本集一上的10.4.4.4:27017节点退出，改为10.4.4.5:27017登录，结果又如何？
./mongo 10.4.4.5:27017
查看rs
>rs.status()
{
"set" : "mongodb_set",
"date" : ISODate("2014-04-04T07:08:08Z"),
"myState" : 1,
"members" : [
{
"_id" : 0,
"name" : "10.4.4.5:27017",
"health" : 1,
"state" : 1,
"stateStr" : "PRIMARY",
"uptime" : 720,
"optime" : Timestamp(1396594084, 4268),
"optimeDate" : ISODate("2014-04-04T06:48:04Z"),
"self" : true
},
{
"_id" : 1,
"name" : "10.4.4.6:27017",
"health" : 1,
"state" : 2,
"stateStr" : "SECONDARY",
"uptime" : 718,
"optime" : Timestamp(1396594084, 4268),
"optimeDate" : ISODate("2014-04-04T06:48:04Z"),
"lastHeartbeat" : ISODate("2014-04-04T07:08:06Z"),
"lastHeartbeatRecv" : ISODate("2014-04-04T07:08:06Z"),
"pingMs" : 0,
"syncingTo" : "10.4.4.5:27017"
}
],
"ok" : 1
}
可见当前连接的是副本集二的primary，此时副本集一的primary去哪里了呢？似乎只能通过10.4.4.4:27017来访问secondary了。
这个难道是mongodb的一个bug？
内容来自用户分享和网络整理，不保证内容的准确性，如有侵权内容，可联系管理员处理 
相关文章推荐
- 关于MongoDB的group用法
- mysql 测试与mongodb 测试对比
- mongodb 常用接口封装测试
- [MongoDB] Sharding 分片
- [MongoDB] 基本操作
- 图解 MongoDB 地理位置索引的实现原理
- spring-data-mongodb 对象关联 @DBRef学习 之疑问
- MongoDB学习
- [MongoDB] 学习笔记（2）
- java操作mongoDB数据库的简单实例
- java操作mongoDB数据库的简单实例
- mongodb在window下和linux下的部署 和 安装可视化工具
- mongodb的数据备份和恢复
- mongodb拆库分表脚本
- mongodb基础应用
- mongodb实现对对某列求和SUM
- mongoDB第八讲：分片
- mongoDB第八讲：分片
- [MongoDB] Index 索引
- Django操作NOSQL(MongoDB)数据库