mongodb3.2 sharding deploy

  1. 云栖社区>
  2. 博客>
  3. 正文

mongodb3.2 sharding deploy

余二五 2017-11-14 14:21:00 浏览450
展开阅读全文

一、部署软件

pc1pc2pc3分别安装mongodb,操作如下:

[root@pc1 ~]# tail /etc/security/limits.conf

1
2
3
4
5
6
7
8
9
mongod soft nproc 40000
* hard nofile 1000000
* soft nofile 1000000
* soft core unlimited
* soft stack 10240
* - nofile 65535
push - nproc 65535
push - nofile 320000
work - nproc 10000

[root@pc1 ~]# tail /etc/security/limits.d/90-nproc.conf

1
2
*          soft    nproc     1024
root       soft    nproc     unlimited

[root@pc1 ~]#  cat /etc/yum.repos.d/mongodb.repo

1
2
3
4
5
6
[mongodb-org-3.2]
name=MongoDB Repository
baseurl=https://repo.mongodb.org/yum/amazon/2013.03/mongodb-org/3.2/x86_64/
gpgcheck=1
enabled=1
gpgkey=https://www.mongodb.org/static/pgp/server-3.2.asc

[root@pc1 ~]#  yum -y install mongodb-org

[root@pc1 ~]# chkconfig mongodb off

[root@pc1 ~]# mkdir  -p  /data/mongodb/{config,data/{config,logs,shard{1,2,3}_1}}

[root@pc1 ~]# chown -R mongod.mongod  /data/mongodb

[root@pc1 ~]# cp /etc/mongod.conf  /data/mongodb/config/shard1.conf

[root@pc1 ~]# tree /data/

/data/

└── mongodb

    ── config

    │   └── shard1.conf

    └── data

        ── config

        ── logs

        ── shard1_1

        ── shard2_1

        └── shard3_1

二、配置服务

副本集1的配置文件与启动脚本如下:

[root@pc1 ~]# egrep "^[^#$]" /data/mongodb/config/shard1.conf

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
systemLog:
  destination: file
  logAppend: true
  path: /data/mongodb/data/logs/shard1.log
storage:
  dbPath: /data/mongodb/data/shard1_1
  directoryPerDB: true
  journal:
    enabled: true
processManagement:
  fork: true  # fork and run in background
  pidFilePath: /var/run/mongodb/shard1.pid  # location of pidfile
net:
  port: 27027
  #bindIp: 127.0.0.1  # Listen to local interface only, comment to listen on all interfaces.
replication:
  oplogSizeMB: 500
  replSetName: shard1
sharding:
  clusterRole: shardsvr

[root@pc1 ~]#  cp /etc/init.d/mongod  /etc/init.d/shard1

[root@pc1 ~]#  vim /etc/init.d/shard1

[root@pc1 ~]#  egrep "^[^#$]" /etc/init.d/shard1

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
/etc/rc.d/init.d/functions
CONFIGFILE="/data/mongodb/config/shard1.conf"
OPTIONS=" -f $CONFIGFILE"
mongod=${MONGOD-/usr/bin/mongod}
MONGO_USER=mongod
MONGO_GROUP=mongod
SYSCONFIG="/etc/sysconfig/mongod"
if [ -f "$SYSCONFIG" ]; then
    "$SYSCONFIG"
fi
NUMACTL_ARGS="--interleave=all"
if which numactl >/dev/null 2>/dev/null && numactl $NUMACTL_ARGS ls / >/dev/null 2>/dev/null
then
    NUMACTL="numactl $NUMACTL_ARGS"
else
    NUMACTL=""
fi
PIDFILEPATH=`awk -F'[:=]' -v IGNORECASE=1 '/^[[:blank:]]*(processManagement\.)?pidfilepath[[:blank:]]*[:=][[:blank:]]*/{print $2}' "$CONFIGFILE" tr -d "[:blank:]\"'" awk -F'#' '{print $1}'`
PIDDIR=`dirname $PIDFILEPATH`
start()
{
  # Make sure the default pidfile directory exists
  if [ ! -d $PIDDIR ]; then
    install -d -m 0755 -o $MONGO_USER -g $MONGO_GROUP $PIDDIR
  fi
if test -f /sys/kernel/mm/transparent_hugepage/defragthen
echo never > /sys/kernel/mm/transparent_hugepage/defrag
fi
  # Recommended ulimit values for mongod or mongos
  # See http://docs.mongodb.org/manual/reference/ulimit/#recommended-settings
  #
  ulimit -f unlimited
  ulimit -t unlimited
  ulimit -v unlimited
  ulimit -n 64000
  ulimit -m unlimited
  ulimit -u 64000
  echo -n $"Starting mongod: "
  daemon --user "$MONGO_USER" --check $mongod "$NUMACTL $mongod $OPTIONS >/dev/null 2>&1"
  RETVAL=$?
  echo
  [ $RETVAL -eq 0 ] && touch /var/lock/subsys/shard1
}
stop()
{
  echo -n $"Stopping mongod: "
  mongo_killproc "$PIDFILEPATH" $mongod
  RETVAL=$?
  echo
  [ $RETVAL -eq 0 ] && rm -f /var/lock/subsys/shard1
}
restart () {
        stop
        start
}
mongo_killproc()
{
  local pid_file=$1
  local procname=$2
  local -i delay=300
  local -i duration=10
  local pid=`pidofproc -p "${pid_file}" ${procname}`
  kill -TERM $pid >/dev/null 2>&1
  usleep 100000
  local -i x=0
  while [ $x -le $delay ] && checkpid $pid; do
    sleep $duration
    x=$(( $x + $duration))
  done
  kill -KILL $pid >/dev/null 2>&1
  usleep 100000
  checkpid $pid # returns 0 only if the process exists
  local RC=$?
  "$RC" -eq 0 ] && failure "${procname} shutdown" || rm -f "${pid_file}"; success "${procname} shutdown"
  RC=$((! $RC)) # invert return code so we return 0 when process is dead.
  return $RC
}
RETVAL=0
case "$1" in
  start)
    start
    ;;
  stop)
    stop
    ;;
  restart|reload|force-reload)
    restart
    ;;
  condrestart)
    [ -f /var/lock/subsys/shard1 ] && restart || :
    ;;
  status)
    status $mongod
    RETVAL=$?
    ;;
  *)
    echo "Usage: $0 {start|stop|status|restart|reload|force-reload|condrestart}"
    RETVAL=1
esac
exit $RETVAL


副本集2的配置文件与启动脚本如下:

[root@pc1 ~]# egrep "^[^#$]" /data/mongodb/config/shard2.conf

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
systemLog:
  destination: file
  logAppend: true
  path: /data/mongodb/data/logs/shard2.log
storage:
  dbPath: /data/mongodb/data/shard2_1
  directoryPerDB: true
  journal:
    enabled: true
processManagement:
  fork: true  # fork and run in background
  pidFilePath: /var/run/mongodb/shard2.pid  # location of pidfile
net:
  port: 27028
  #bindIp: 127.0.0.1  # Listen to local interface only, comment to listen on all interfaces.
replication:
  oplogSizeMB: 500
  replSetName: shard2
sharding:
  clusterRole: shardsvr

[root@pc1 ~]#  cp /etc/init.d/mongod  /etc/init.d/shard2

[root@pc1 ~]#  vim  /etc/init.d/shard2

[root@pc1 ~]#  egrep "^[^#$]" /etc/init.d/shard2

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
/etc/rc.d/init.d/functions
CONFIGFILE="/data/mongodb/config/shard2.conf"
OPTIONS=" -f $CONFIGFILE"
mongod=${MONGOD-/usr/bin/mongod}
MONGO_USER=mongod
MONGO_GROUP=mongod
SYSCONFIG="/etc/sysconfig/mongod"
if [ -f "$SYSCONFIG" ]; then
    "$SYSCONFIG"
fi
NUMACTL_ARGS="--interleave=all"
if which numactl >/dev/null 2>/dev/null && numactl $NUMACTL_ARGS ls / >/dev/null 2>/dev/null
then
    NUMACTL="numactl $NUMACTL_ARGS"
else
    NUMACTL=""
fi
PIDFILEPATH=`awk -F'[:=]' -v IGNORECASE=1 '/^[[:blank:]]*(processManagement\.)?pidfilepath[[:blank:]]*[:=][[:blank:]]*/{print $2}' "$CONFIGFILE" tr -d "[:blank:]\"'" awk -F'#' '{print $1}'`
PIDDIR=`dirname $PIDFILEPATH`
start()
{
  # Make sure the default pidfile directory exists
  if [ ! -d $PIDDIR ]; then
    install -d -m 0755 -o $MONGO_USER -g $MONGO_GROUP $PIDDIR
  fi
if test -f /sys/kernel/mm/transparent_hugepage/defragthen
echo never > /sys/kernel/mm/transparent_hugepage/defrag
fi
  # Recommended ulimit values for mongod or mongos
  # See http://docs.mongodb.org/manual/reference/ulimit/#recommended-settings
  #
  ulimit -f unlimited
  ulimit -t unlimited
  ulimit -v unlimited
  ulimit -n 64000
  ulimit -m unlimited
  ulimit -u 64000
  echo -n $"Starting mongod: "
  daemon --user "$MONGO_USER" --check $mongod "$NUMACTL $mongod $OPTIONS >/dev/null 2>&1"
  RETVAL=$?
  echo
  [ $RETVAL -eq 0 ] && touch /var/lock/subsys/shard2
}
stop()
{
  echo -n $"Stopping mongod: "
  mongo_killproc "$PIDFILEPATH" $mongod
  RETVAL=$?
  echo
  [ $RETVAL -eq 0 ] && rm -f /var/lock/subsys/shard2
}
restart () {
        stop
        start
}
mongo_killproc()
{
  local pid_file=$1
  local procname=$2
  local -i delay=300
  local -i duration=10
  local pid=`pidofproc -p "${pid_file}" ${procname}`
  kill -TERM $pid >/dev/null 2>&1
  usleep 100000
  local -i x=0
  while [ $x -le $delay ] && checkpid $pid; do
    sleep $duration
    x=$(( $x + $duration))
  done
  kill -KILL $pid >/dev/null 2>&1
  usleep 100000
  checkpid $pid # returns 0 only if the process exists
  local RC=$?
  "$RC" -eq 0 ] && failure "${procname} shutdown" || rm -f "${pid_file}"; success "${procname} shutdown"
  RC=$((! $RC)) # invert return code so we return 0 when process is dead.
  return $RC
}
RETVAL=0
case "$1" in
  start)
    start
    ;;
  stop)
    stop
    ;;
  restart|reload|force-reload)
    restart
    ;;
  condrestart)
    [ -f /var/lock/subsys/shard2 ] && restart || :
    ;;
  status)
    status $mongod
    RETVAL=$?
    ;;
  *)
    echo "Usage: $0 {start|stop|status|restart|reload|force-reload|condrestart}"
    RETVAL=1
esac
exit $RETVAL


副本集3的配置文件与启动脚本如下:

[root@pc1 ~]# egrep "^[^#$]" /data/mongodb/config/shard3.conf

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
systemLog:
  destination: file
  logAppend: true
  path: /data/mongodb/data/logs/shard3.log
storage:
  dbPath: /data/mongodb/data/shard3_1
  directoryPerDB: true
  journal:
    enabled: true
processManagement:
  fork: true  # fork and run in background
  pidFilePath: /var/run/mongodb/shard3.pid  # location of pidfile
net:
  port: 27026
  #bindIp: 127.0.0.1  # Listen to local interface only, comment to listen on all interfaces.
replication:
  oplogSizeMB: 500
  replSetName: shard3
sharding:
  clusterRole: shardsvr

[root@pc1 ~]#  cp /etc/init.d/mongod  /etc/init.d/shard3

[root@pc1 ~]#  vim  /etc/init.d/shard3

[root@pc1 ~]#  egrep "^[^#$]" /etc/init.d/shard3  (略,内容基本与shard2 、shard1相同)

复制配置文件到pc2pc3,并启动服务

[root@pc1 ~]#  scp /data/mongodb/config/*.conf  pc2:/data/mongodb/config/

[root@pc1 ~]#  scp /etc/init.d/shard*  pc2:/etc/init.d/

[root@pc1 ~]#  scp /data/mongodb/config/*.conf  pc3:/data/mongodb/config/

[root@pc1 ~]#  scp /etc/init.d/shard*  pc2:/etc/init.d/

[root@pc1 ~]# 

[root@pc1 ~]#  /etc/init.d/shard1 start

[root@pc1 ~]#  ssh pc2 '/etc/init.d/shard1 start'

[root@pc1 ~]#  ssh pc3 '/etc/init.d/shard1 start'


[root@pc1 ~]#  /etc/init.d/shard2 start

[root@pc1 ~]#  ssh pc2 '/etc/init.d/shard2 start'

[root@pc1 ~]#  ssh pc3 '/etc/init.d/shard2 start'

  

[root@pc1 ~]#  /etc/init.d/shard3 start

[root@pc1 ~]#  ssh pc2 '/etc/init.d/shard3 start'

[root@pc1 ~]#  ssh pc3 '/etc/init.d/shard3 start'


配置副本集1如下操作:

[root@pc1 ~]# mongo localhost:27027

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
MongoDB shell version: 3.2.7
connecting to: localhost:27027/test
> use admin
switched to db admin
> rs.initiate()
{
"info2" "no configuration specified. Using a default configuration for the set",
"me" "pc1:27027",
"ok" : 1
}
shard1:OTHER> rs.add("pc2:27027")
"ok" : 1 }
shard1:PRIMARY> rs.addArb("pc3:27027")
"ok" : 1 }
shard1:PRIMARY> rs.status()
{
"set" "shard1",
"date" : ISODate("2016-06-28T03:59:41.403Z"),
"myState" : 1,
"term" : NumberLong(1),
"heartbeatIntervalMillis" : NumberLong(2000),
"members" : [
{
"_id" : 0,
"name" "pc1:27027",
"health" : 1,
"state" : 1,
"stateStr" "PRIMARY",
"uptime" : 1634,
"optime" : {
"ts" : Timestamp(1467086346, 1),
"t" : NumberLong(1)
},
"optimeDate" : ISODate("2016-06-28T03:59:06Z"),
"infoMessage" "could not find member to sync from",
"electionTime" : Timestamp(1467086301, 2),
"electionDate" : ISODate("2016-06-28T03:58:21Z"),
"configVersion" : 3,
"self" true
},
{
"_id" : 1,
"name" "pc2:27027",
"health" : 1,
"state" : 2,
"stateStr" "SECONDARY",
"uptime" : 54,
"optime" : {
"ts" : Timestamp(1467086346, 1),
"t" : NumberLong(1)
},
"optimeDate" : ISODate("2016-06-28T03:59:06Z"),
"lastHeartbeat" : ISODate("2016-06-28T03:59:40.050Z"),
"lastHeartbeatRecv" : ISODate("2016-06-28T03:59:41.041Z"),
"pingMs" : NumberLong(1),
"syncingTo" "pc1:27027",
"configVersion" : 3
},
{
"_id" : 2,
"name" "pc3:27027",
"health" : 1,
"state" : 7,
"stateStr" "ARBITER",
"uptime" : 35,
"lastHeartbeat" : ISODate("2016-06-28T03:59:40.053Z"),
"lastHeartbeatRecv" : ISODate("2016-06-28T03:59:41.044Z"),
"pingMs" : NumberLong(2),
"configVersion" : 3
}
],
"ok" : 1
}

配置副本集2、配置副本集3的操作与配置副本集基本相同,本文不再罗嗦。注意:配置副本集1时,pc1primarypc2secondary ,而pc3arbiter;配置副本集2时,pc2primarypc3secondary,而pc1arbiter;配置副本集3时,pc3primarypc1secondary,而pc2arbiter


config server 的配置文件与启动文件如下:

[[root@pc1 ~]# egrep "^[^#$]" /data/mongodb/config/config.conf

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
systemLog:
  destination: file
  logAppend: true
  path: /data/mongodb/data/logs/config.log
storage:
  dbPath: /data/mongodb/data/config
  directoryPerDB: true
  journal:
    enabled: true
processManagement:
  fork: true  # fork and run in background
  pidFilePath: /var/run/mongodb/config.pid  # location of pidfile
net:
  port: 27029
  #bindIp: 127.0.0.1  # Listen to local interface only, comment to listen on all interfaces.
replication:
  oplogSizeMB: 500
  replSetName: configrs
sharding:
  clusterRole: configsvr

[root@pc1 ~]#  cp /etc/init.d/mongod  /etc/init.d/mconfig

[root@pc1 ~]#  vim /etc/init.d/mconfig

[root@pc1 ~]#  egrep "^[^#$]" /etc/init.d/mconfig

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
/etc/rc.d/init.d/functions
CONFIGFILE="/data/mongodb/config/config.conf"
OPTIONS=" -f $CONFIGFILE"
mongod=${MONGOD-/usr/bin/mongod}
MONGO_USER=mongod
MONGO_GROUP=mongod
SYSCONFIG="/etc/sysconfig/mongod"
if [ -f "$SYSCONFIG" ]; then
    "$SYSCONFIG"
fi
NUMACTL_ARGS="--interleave=all"
if which numactl >/dev/null 2>/dev/null && numactl $NUMACTL_ARGS ls / >/dev/null 2>/dev/null
then
    NUMACTL="numactl $NUMACTL_ARGS"
else
    NUMACTL=""
fi
PIDFILEPATH=`awk -F'[:=]' -v IGNORECASE=1 '/^[[:blank:]]*(processManagement\.)?pidfilepath[[:blank:]]*[:=][[:blank:]]*/{print $2}' "$CONFIGFILE" tr -d "[:blank:]\"'" awk -F'#' '{print $1}'`
PIDDIR=`dirname $PIDFILEPATH`
start()
{
  # Make sure the default pidfile directory exists
  if [ ! -d $PIDDIR ]; then
    install -d -m 0755 -o $MONGO_USER -g $MONGO_GROUP $PIDDIR
  fi
if test -f /sys/kernel/mm/transparent_hugepage/defragthen
echo never > /sys/kernel/mm/transparent_hugepage/defrag
fi
  # Recommended ulimit values for mongod or mongos
  # See http://docs.mongodb.org/manual/reference/ulimit/#recommended-settings
  ulimit -f unlimited
  ulimit -t unlimited
  ulimit -v unlimited
  ulimit -n 64000
  ulimit -m unlimited
  ulimit -u 64000
  echo -n $"Starting mongod: "
  daemon --user "$MONGO_USER" --check $mongod "$NUMACTL $mongod $OPTIONS >/dev/null 2>&1"
  RETVAL=$?
  echo
  [ $RETVAL -eq 0 ] && touch /var/lock/subsys/mconfig
}
stop()
{
  echo -n $"Stopping mongod: "
  mongo_killproc "$PIDFILEPATH" $mongod
  RETVAL=$?
  echo
  [ $RETVAL -eq 0 ] && rm -f /var/lock/subsys/mconfig
}
restart () {
        stop
        start
}
mongo_killproc()
{
  local pid_file=$1
  local procname=$2
  local -i delay=300
  local -i duration=10
  local pid=`pidofproc -p "${pid_file}" ${procname}`
  kill -TERM $pid >/dev/null 2>&1
  usleep 100000
  local -i x=0
  while [ $x -le $delay ] && checkpid $pid; do
    sleep $duration
    x=$(( $x + $duration))
  done
  kill -KILL $pid >/dev/null 2>&1
  usleep 100000
  checkpid $pid # returns 0 only if the process exists
  local RC=$?
  "$RC" -eq 0 ] && failure "${procname} shutdown" || rm -f "${pid_file}"; success "${procname} shutdown"
  RC=$((! $RC)) # invert return code so we return 0 when process is dead.
  return $RC
}
RETVAL=0
case "$1" in
  start)
    start
    ;;
  stop)
    stop
    ;;
  restart|reload|force-reload)
    restart
    ;;
  condrestart)
    [ -f /var/lock/subsys/mconfig ] && restart || :
    ;;
  status)
    status $mongod
    RETVAL=$?
    ;;
  *)
    echo "Usage: $0 {start|stop|status|restart|reload|force-reload|condrestart}"
    RETVAL=1
esac
exit $RETVAL

[root@pc1 ~]# scp /data/mongodb/config/config.conf  pc2:/data/mongodb/config

[root@pc1 ~]# scp /etc/init.d/mconfig  pc2:/etc/init.d/

[root@pc1 ~]# 

[root@pc1 ~]#  scp /data/mongodb/config/config.conf  pc2:/data/mongodb/config

[root@pc1 ~]#  scp /etc/init.d/mconfig  pc3:/etc/init.d/

[root@pc1 ~]#  

[root@pc1 ~]# /etc/init.d/mconfig  start

Starting mongod:                                           [确定]

[root@pc1 ~]# lsof -i:27029

COMMAND  PID   USER   FD   TYPE DEVICE SIZE/OFF NODE NAME

mongod  2765 mongod    6u  IPv4  11520      0t0  TCP *:27029 (LISTEN)

[root@pc1 ~]# ssh pc2 '/etc/init.d/mconfig  start'

[root@pc1 ~]# ssh pc2 '/etc/init.d/mconfig  start'


为了保持config server的一致性,实现各config server的副本集功能,完成效果如下所示(config server的副本集不支持使用arbiter)

[root@pc1 ~]# mongo localhost:27029

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
MongoDB shell version: 3.2.7
connecting to: localhost:27029/test
configrs:SECONDARY> use admin
switched to db admin
configrs:SECONDARY> rs.slaveOk()
configrs:SECONDARY> rs.status()
{
"set" "configrs",
"date" : ISODate("2016-07-01T03:25:49.471Z"),
"myState" : 2,
"term" : NumberLong(2),
"syncingTo" "pc3:27029",
"configsvr" true,
"heartbeatIntervalMillis" : NumberLong(2000),
"members" : [
{
"_id" : 0,
"name" "pc1:27029",
"health" : 1,
"state" : 2,
"stateStr" "SECONDARY",
"uptime" : 3937,
"optime" : {
"ts" : Timestamp(1467339616, 1),
"t" : NumberLong(2)
},
"optimeDate" : ISODate("2016-07-01T02:20:16Z"),
"syncingTo" "pc3:27029",
"configVersion" : 3,
"self" true
},
{
"_id" : 1,
"name" "pc2:27029",
"health" : 1,
"state" : 2,
"stateStr" "SECONDARY",
"uptime" : 3936,
"optime" : {
"ts" : Timestamp(1467339616, 1),
"t" : NumberLong(2)
},
"optimeDate" : ISODate("2016-07-01T02:20:16Z"),
"lastHeartbeat" : ISODate("2016-07-01T03:25:48.578Z"),
"lastHeartbeatRecv" : ISODate("2016-07-01T03:25:47.481Z"),
"pingMs" : NumberLong(1),
"syncingTo" "pc3:27029",
"configVersion" : 3
},
{
"_id" : 2,
"name" "pc3:27029",
"health" : 1,
"state" : 1,
"stateStr" "PRIMARY",
"uptime" : 3936,
"optime" : {
"ts" : Timestamp(1467339616, 1),
"t" : NumberLong(2)
},
"optimeDate" : ISODate("2016-07-01T02:20:16Z"),
"lastHeartbeat" : ISODate("2016-07-01T03:25:49.146Z"),
"lastHeartbeatRecv" : ISODate("2016-07-01T03:25:47.585Z"),
"pingMs" : NumberLong(1),
"electionTime" : Timestamp(1467339615, 1),
"electionDate" : ISODate("2016-07-01T02:20:15Z"),
"configVersion" : 3
}
],
"ok" : 1
}
configrs:SECONDARY> exit
bye
[root@pc1 ~]#


mongos 路由配置文件与启动脚本内容如下:

[root@pc1 ~]# egrep "^[^#$]" /data/mongodb/config/mongos.conf

1
2
3
4
5
6
7
8
9
10
11
12
13
systemLog:
  destination: file
  logAppend: true
  path: /data/mongodb/data/logs/mongos.log
processManagement:
  fork: true  # fork and run in background
  pidFilePath: /var/run/mongodb/mongos.pid  # location of pidfile
net:
  port: 27017
  #bindIp: 127.0.0.1  # Listen to local interface only, comment to listen on all interfaces.
sharding:
  configDB: pc1:27029,pc2:27029,pc3:27029  
#注: mongodb3.4时格式必须为 configDB: configrs_name/configServer_list

[root@pc1 ~]#  cp /etc/init.d/mongod  /etc/init.d/mongos

[root@pc1 ~]#  vim /etc/init.d/mongos

[root@pc1 ~]#  egrep "^[^#$]" /etc/init.d/mongos

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
/etc/rc.d/init.d/functions
CONFIGFILE="/data/mongodb/config/mongos.conf"
OPTIONS=" -f $CONFIGFILE"
mongod=${MONGOD-/usr/bin/mongos}
MONGO_USER=mongod
MONGO_GROUP=mongod
SYSCONFIG="/etc/sysconfig/mongod"
if [ -f "$SYSCONFIG" ]; then
    "$SYSCONFIG"
fi
NUMACTL_ARGS="--interleave=all"
if which numactl >/dev/null 2>/dev/null && numactl $NUMACTL_ARGS ls / >/dev/null 2>/dev/null
then
    NUMACTL="numactl $NUMACTL_ARGS"
else
    NUMACTL=""
fi
PIDFILEPATH=`awk -F'[:=]' -v IGNORECASE=1 '/^[[:blank:]]*(processManagement\.)?pidfilepath[[:blank:]]*[:=][[:blank:]]*/{print $2}' "$CONFIGFILE" tr -d "[:blank:]\"'" awk -F'#' '{print $1}'`
PIDDIR=`dirname $PIDFILEPATH`
start()
{
  # Make sure the default pidfile directory exists
  if [ ! -d $PIDDIR ]; then
    install -d -m 0755 -o $MONGO_USER -g $MONGO_GROUP $PIDDIR
  fi
if test -f /sys/kernel/mm/transparent_hugepage/defragthen
echo never > /sys/kernel/mm/transparent_hugepage/defrag
fi
  # Recommended ulimit values for mongod or mongos
  # See http://docs.mongodb.org/manual/reference/ulimit/#recommended-settings
  #
  ulimit -f unlimited
  ulimit -t unlimited
  ulimit -v unlimited
  ulimit -n 64000
  ulimit -m unlimited
  ulimit -u 64000
  echo -n $"Starting mongod: "
  daemon --user "$MONGO_USER" --check $mongod "$NUMACTL $mongod $OPTIONS >/dev/null 2>&1"
  RETVAL=$?
  echo
  [ $RETVAL -eq 0 ] && touch /var/lock/subsys/mongos
}
stop()
{
  echo -n $"Stopping mongod: "
  mongo_killproc "$PIDFILEPATH" $mongod
  RETVAL=$?
  echo
  [ $RETVAL -eq 0 ] && rm -f /var/lock/subsys/mongos
}
restart () {
        stop
        start
}
mongo_killproc()
{
  local pid_file=$1
  local procname=$2
  local -i delay=300
  local -i duration=10
  local pid=`pidofproc -p "${pid_file}" ${procname}`
  kill -TERM $pid >/dev/null 2>&1
  usleep 100000
  local -i x=0
  while [ $x -le $delay ] && checkpid $pid; do
    sleep $duration
    x=$(( $x + $duration))
  done
  kill -KILL $pid >/dev/null 2>&1
  usleep 100000
  checkpid $pid # returns 0 only if the process exists
  local RC=$?
  "$RC" -eq 0 ] && failure "${procname} shutdown" || rm -f "${pid_file}"; success "${procname} shutdown"
  RC=$((! $RC)) # invert return code so we return 0 when process is dead.
  return $RC
}
RETVAL=0
case "$1" in
  start)
    start
    ;;
  stop)
    stop
    ;;
  restart|reload|force-reload)
    restart
    ;;
  condrestart)
    [ -f /var/lock/subsys/mongos ] && restart || :
    ;;
  status)
    status $mongod
    RETVAL=$?
    ;;
  *)
    echo "Usage: $0 {start|stop|status|restart|reload|force-reload|condrestart}"
    RETVAL=1
esac
exit $RETVAL

[root@pc1 ~]#  scp /data/mongodb/config/mongos   pc2:/data/mongodb/config/

[root@pc1 ~]#  scp /etc/init.d/mongos  pc2:/etc/init.d/

[root@pc1 ~]#  scp /data/mongodb/config/mongos   pc3:/data/mongodb/config/

[root@pc1 ~]#  scp /etc/init.d/mongos  pc3:/etc/init.d/

[root@pc1 ~]#  /etc/init.d/mongos start

Starting mongod:                                           [确定]

[root@pc1 ~]#  ssh pc2 '/etc/init.d/mongos start'

Starting mongod: [确定]

[root@pc1 ~]#  ssh pc3 '/etc/init.d/mongos start'

Starting mongod: [确定]

[root@pc1 ~]#

有些网朋友说:需要实现mongos路由的副本集(即多台mongos路由之间的复制),但从原理分析,我个人没能理解这种方案的功能,当然本人也尝试过,发现设置副本集之后mongos服务无法启动(并无深入研究,请各位朋友提宝贝意见)


三、实现分片

连接到任何一台mongos路由上,操作如下:

[root@pc1 ~]# mongo

MongoDB shell version: 3.2.7

connecting to: test

mongos> use admin

switched to db admin

mongos> sh.status()

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
--- Sharding Status ---
  sharding version: {
"_id" : 1,
"minCompatibleVersion" : 5,
"currentVersion" : 6,
"clusterId" : ObjectId("577226b7511b1f96da0ddba2")
}
  shards:
  active mongoses:
"3.2.7" : 3
  balancer:
Currently enabled:  yes
Currently running:  no
Failed balancer rounds in last 5 attempts:  0
Migration Results for the last 24 hours:
No recent migrations
  databases:

mongos> sh.addShard("shard1/pc1:27027")

{ "shardAdded" : "shard1", "ok" : 1 }

mongos> sh.addShard("shard3/pc2:27026")

{ "shardAdded" : "shard3", "ok" : 1 }

mongos> sh.addShard("shard2/pc3:27028")

{ "shardAdded" : "shard2", "ok" : 1 }

mongos> sh.status()

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
--- Sharding Status ---
  sharding version: {
"_id" : 1,
"minCompatibleVersion" : 5,
"currentVersion" : 6,
"clusterId" : ObjectId("577226b7511b1f96da0ddba2")
}
  shards:
{  "_id" "shard1",  "host" "shard1/pc1:27027,pc2:27027" }
{  "_id" "shard2",  "host" "shard2/pc1:27028,pc3:27028" }
{  "_id" "shard3",  "host" "shard3/pc2:27026,pc3:27026" }
  active mongoses:
"3.2.7" : 3
  balancer:
Currently enabled:  yes
Currently running:  no
Failed balancer rounds in last 5 attempts:  5
Last reported error:  mongos specified a different config database string : stored : pc1:27029 vs given : pc1:27029,pc2:27029,pc3:27029
Time of Reported error:  Tue Jun 28 2016 07:34:07 GMT+0000 (UTC)
Migration Results for the last 24 hours:
No recent migrations
  databases:

mongos> use config

switched to db config

mongos> db.chunks.find()

mongos> db.settings.find()

{ "_id" : "chunksize", "value" : NumberLong(64) }

mongos> db.settings.save({"_id":"chunksize","value":NumberLong(5)})  #不建议修改chunk的值,本例为了实例效果改为5M。最小的chunksize(1M),默认chunksize为64M

WriteResult({ "nMatched" : 1, "nUpserted" : 0, "nModified" : 1 })

mongos> db.settings.find()

{ "_id" : "chunksize", "value" : NumberLong(5) }

mongos> use admin

switched to db admin

mongos> sh.enableSharding("test")

{ "ok" : 1 }

mongos> sh.shardCollection("test.user",{uid:1})

{ "collectionsharded" : "test.user", "ok" : 1 }

mongos> use test

switched to db test

mongos> db.use.ensureIndex({uid:1})

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
{
"raw" : {
"shard1/pc1:27027,pc2:27027" : {
"createdCollectionAutomatically" true,
"numIndexesBefore" : 1,
"numIndexesAfter" : 2,
"ok" : 1,
"$gleStats" : {
"lastOpTime" : Timestamp(1467102931, 2),
"electionId" : ObjectId("7fffffff0000000000000001")
}
}
},
"ok" : 1
}

mongos>  for (i=1;i<=100000;i++) db.user.insert({uid:'user'+i,age:(i%15),address:'#'+i+',shangdi south road,beijing',preferbooks:['book'+i,'hello world']})

WriteResult({ "nInserted" : 1 })

mongos> sh.status()

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
--- Sharding Status ---
  sharding version: {
"_id" : 1,
"minCompatibleVersion" : 5,
"currentVersion" : 6,
"clusterId" : ObjectId("577234b7c92b86a15a2901d4")
}
  shards:
{  "_id" "shard1",  "host" "shard1/pc1:27027,pc2:27027" }
{  "_id" "shard2",  "host" "shard2/pc2:27028,pc3:27028" }
{  "_id" "shard3",  "host" "shard3/pc1:27026,pc3:27026" }
  active mongoses:
"3.2.7" : 3
  balancer:
Currently enabled:  yes
Currently running:  no
Failed balancer rounds in last 5 attempts:  0
Migration Results for the last 24 hours:
2 : Success
1 : Failed with error 'aborted', from shard1 to shard2
  databases:
{  "_id" "test",  "primary" "shard1",  "partitioned" true }
test.user
shard key: { "uid" : 1 }
unique: false
balancing: true
chunks:
shard11
shard21
shard31
"uid" : { "$minKey" : 1 } } -->> { "uid" "user2" } on : shard2 Timestamp(2, 0)
"uid" "user2" } -->> { "uid" "user8" } on : shard3 Timestamp(3, 0)
"uid"