startup mongodb with --replSet command option
example :
1. startup one node:
mongod --replSet $setname[/$rmt_ip:$port]
2. startup another node :
mongod --replSet $setname[/$rmt_ip:$port]
3. run ininiate command within one of the nodes only :
db.runCommand({replSetInitiate : <config_object>}) or
rs.initiate(<config_object>)
如果不带参数rs.initiate(),初始化包含$setname[/$rmt_ip:$port]的节点.
<config_object>完整格式:
{
_id : <setname>,
members: [
{
_id : <ordinal>,
host : <hostname[:port]>,
[, priority: <priority>]
[, arbiterOnly : true]
[, votes : <n>]
[, hidden : true]
[, slaveDelay : <n>]
[, buildIndexes : <bool>]
[, initialSync : {
[state : 1|2,]
[_id : <n>,]
[name : <host>,]
[optime : <date>]}]
}
, ...
],
[settings: {
[getLastErrorDefaults: <lasterrdefaults>]
[, heartbeatSleep : <seconds>]
[, heartbeatTimeout : <seconds>]
[, heartbeatConnRetries : <n>]
}]
}
参数说明:
强制参数：
_id - the set name. This must match command line setting. Set names are usually alphanumeric and, in particular, cannot contain the '/' character.
members - an array of servers in the set. For simpler configs, one can often simply set _id and host fields only – all the rest are optional.
_id - each member has an _id ordinal, typically beginning with zero and numbered in increasing order. when a node is retired (removed from the config), its _id should not be reused.
host - host name and optionally the port for the member
Member可选参数:
arbiterOnly - if true, this member will participate in consensus (voting) but receive no data. Defaults to false.
votes - number of votes this member gets in elections. Defaults to 1. Normally you do not need to set this parameter. Sometimes useful when the number of nodes are even or for biasing towards a particular data center.
priority - priority a server has for potential election as primary. The highest priority member which is up will become primary. Default 1.0. Priority zero means server can never be primary (0 and 1 are the only priorities currently supported).
hidden - when true, do not advertise the member's existence to clients in isMaster command responses. (v1.7+)
slaveDelay - how far behind this slave's replication should be (in seconds). Defaults to 0 (as up-to-date as possible). Can be used to recover from human errors (accidentally dropping a database, etc.). This option can only be set on passive members. (v1.6.3+)
buildIndexes - boolean, defaults to true. If the priority is 0, you can set buildIndexes to false to prevent indexes from being created on this member. This could be useful on a machine which is only used for backup as there is less overhead on writes if there are no secondary indexes. Note: the _id index is always created.
initialSync (1.7.4+) - allows you to specify where this server should initially sync from. If not specified, the server will choose the first primary or secondary it finds that is healthy. The default should usually work fine, but you can change this by specifying any of the following options:
state : 1 forces the server to clone from the primary, 2 from a secondary.
_id : the _id of the member to clone from.
name : the host to clone from.
optime : finds a server that is at least this up-to-date to clone from. Can be a Date or Timestamp type.
Set可选参数:
The final optional argument, settings, can be used to set options on the set as a whole. Often one can leave out settings completely from the config as the defaults are reasonable.
getLastErrorDefaults specifies defaults for the getlasterror command. If the client calls getLastError with no parameters, the default object specified here is used. (v1.6.2+)
heartbeatSleep how frequently nodes should send a heartbeat to each other (default: 2 seconds, must be greater than 10 milliseconds).
heartbeatTimeout indicates how long a node needs to fail to send data before we note a problem (default: 10 seconds, must be greater than 10 milliseconds).
heartbeatConnRetries is how many times after heartbeatTimeout to try connecting again and getting a new heartbeat (default: 3 tries).
mongo shell replica set reference commands :
reptest:PRIMARY> rs.help()
rs.status() { replSetGetStatus : 1 } checks repl set status
rs.initiate() { replSetInitiate : null } initiates set with default settings
rs.initiate(cfg) { replSetInitiate : cfg } initiates set with configuration cfg
rs.conf() get the current configuration object from local.system.replset
rs.reconfig(cfg) updates the configuration of a running replica set with cfg (disconnects)
rs.add(hostportstr) add a new member to the set with default attributes (disconnects)
rs.add(membercfgobj) add a new member to the set with extra attributes (disconnects)
rs.addArb(hostportstr) add a new member which is arbiterOnly:true (disconnects)
rs.stepDown([secs]) step down as primary (momentarily) (disconnects)
rs.freeze(secs) make a node ineligible to become primary for the time specified
rs.remove(hostportstr) remove a host from the replica set (disconnects)
rs.slaveOk() shorthand for db.getMongo().setSlaveOk()
db.isMaster() check who is primary
reconfiguration helpers disconnect from the database so the shell will display
an error, even if the command succeeds.
see also http://<mongod_host>:28017/_replSet for additional diagnostic info
或者使用命令:
db.runCommand({})
* { isMaster : 1 }
* { replSetGetStatus : 1 }
* { replSetInitiate : <config> }
* { replSetReconfig: <config> }
* { replSetStepDown : <seconds> }
* { replSetFreeze : <seconds> }
例如:
reptest:PRIMARY> db.runCommand({"replSetGetStatus" : 1})
{
"set" : "reptest",
"date" : ISODate("2011-01-12T02:38:30Z"),
"myState" : 1,
"members" : [
{
"_id" : 0,
"name" : "172.16.3.33:5281",
"health" : 1,
"state" : 2,
"stateStr" : "SECONDARY",
"uptime" : 6595,
"optime" : {
"t" : 1294655286000,
"i" : 1
},
"optimeDate" : ISODate("2011-01-10T10:28:06Z"),
"lastHeartbeat" : ISODate("2011-01-12T02:38:30Z")
},
{
"_id" : 1,
"name" : "db-172-16-3-39.sky-mobi.com.hz.gs:5281",
"health" : 1,
"state" : 1,
"stateStr" : "PRIMARY",
"optime" : {
"t" : 1294655286000,
"i" : 1
},
"optimeDate" : ISODate("2011-01-10T10:28:06Z"),
"self" : true
}
],
"ok" : 1
}
The myState field indicates the state of this server. Valid states are:
0 Starting up, phase 1
1 Primary
2 Secondary
3 Recovering
4 Fatal error
5 Starting up, phase 2
6 Unknown state
7 Arbiter
8 Down
The health field indicates the health of this server. Valid states are:
0 Server is down
1 Server is up
The errmsg field can contain informational messages, as shown above.
force a node to be primary at a given point in time,use below method :
{ replSetStepDown : <seconds> }
Manually tell a member to step down as primary. Node will become eligible to be primary again after the specified number of seconds. (Presumably, another node will take over by then if it were eligible.)
{ replSetFreeze : <seconds> }
'Freeze' state of this member to the extent we can do that. What this really means is that this node will not attempt to become primary until the time period specified expires.
You can call again with {replSetFreeze:0} to unfreeze sooner. A process restart unfreezes the member also.
If the node is already primary, you need to use replSetStepdown instead.
# 修改replica set配置
> // shell v1.7.x:
> // example : give 1st set member 2 votes
> cfg = rs.conf()
> cfg.members[0].votes = 2
> rs.reconfig(cfg)
# 修改配置需要注意
* You must connect to the current primary.
* A majority of members of the set must be up.
# 如两节点的replica环境如果将主节点冻结,在新的主节点修改配置后保存的时候可能会报错,原因是老的主节点未解冻.不能接收到消息.
如下:
reptest:PRIMARY> rs.reconfig(cfg)
Wed Jan 12 11:02:31 query failed : admin.$cmd { replSetReconfig: { _id: "reptest", version: 9, members: [ { _id: 0, host: "172.16.3.33:5281", votes: 2 }, { _id: 1, host: "172.16.3.39:5281", votes: 2.0 } ] } } to: 127.0.0.1:5281
shell got exception during reconfig: Error: error doing query: failed
in some circumstances, the primary steps down and closes connections on a reconfig
null
Wed Jan 12 11:02:31 trying reconnect to 127.0.0.1:5281
Wed Jan 12 11:02:31 reconnect 127.0.0.1:5281 ok
# 因此应该考虑先修改参数，再冻结