Watchdog Configuration Example

This tutorial explains the simple way to try "Watchdog". What you need is 2 Linux boxes on which Pgpool-II is installed and a PostgreSQL on the same machine or in the other one. It is enough that 1 node for backend exists. You can use watchdog with Pgpool-II in any mode: replication mode, master/slave mode and raw mode.

This example uses use "osspc16" as an Active node and "osspc20" as a Standby node. "Someserver" means one of them.

8.2.1. Common configurations

Set the following parameters in both of active and standby nodes.

8.2.1.1. Enabling watchdog

First of all, set use_watchdog to on.

use_watchdog = on
                                        # Activates watchdog
     

8.2.1.2. Configure Up stream servers

Specify the up stream servers (e.g. application servers). Leaving it blank is also fine.

trusted_servers = ''
                                        # trusted server list which are used
                                        # to confirm network connection
                                        # (hostA,hostB,hostC,...)
    

8.2.1.3. Watchdog Communication

Specify the TCP port number for watchdog communication.

wd_port = 9000
                                        # port number for watchdog service
    

8.2.1.4. Virtual IP

Specify the IP address to be used as a virtual IP address in the delegate_IP .

delegate_IP = '133.137.177.143'
                                        # delegate IP address
     

Note: Make sure the IP address configured as a Virtual IP should be free and is not used by any other machine.

8.2.2. Individual Server Configurations

Next, set the following parameters for each Pgpool-II . Specify other_pgpool_hostname , other_pgpool_port and other_wd_port with the values of other Pgpool-II server values.

8.2.2.1. Active (osspc16) Server configurations

other_pgpool_hostname0 = 'osspc20'
                                        # Host name or IP address to connect to for other pgpool 0
other_pgpool_port0 = 9999
                                        # Port number for other pgpool 0
other_wd_port0 = 9000
                                        # Port number for other watchdog 0
    

8.2.2.2. Standby (osspc20) Server configurations

other_pgpool_hostname0 = 'osspc16'
                                        # Host name or IP address to connect to for other pgpool 0
other_pgpool_port0 = 9999
                                        # Port number for other pgpool 0
other_wd_port0 = 9000
                                        # Port number for other watchdog 0
    

8.2.3. Starting Pgpool-II

Start Pgpool-II on each servers from root user with "-n" switch and redirect log messages into pgpool.log file.

8.2.3.1. Starting pgpool in Active server (osspc16)

First start the Pgpool-II on Active server.

[user@osspc16]$ su -
[root@osspc16]# {installed_dir}/bin/pgpool -n -f {installed_dir}/etc/pgpool.conf > pgpool.log 2>&1
    

Log messages will show that Pgpool-II has the virtual IP address and starts watchdog process.

LOG:  I am announcing my self as master/coordinator watchdog node
LOG:  I am the cluster leader node
DETAIL:  our declare coordinator message is accepted by all nodes
LOG:  I am the cluster leader node. Starting escalation process
LOG:  escalation process started with PID:59449
LOG:  watchdog process is initialized
LOG:  watchdog: escalation started
LOG:  I am the master watchdog node
DETAIL:  using the local backend node status
    

8.2.3.2. Starting pgpool in Standby server (osspc20)

Now start the Pgpool-II on Standby server.

[user@osspc20]$ su -
[root@osspc20]# {installed_dir}/bin/pgpool -n -f {installed_dir}/etc/pgpool.conf > pgpool.log 2>&1
    

Log messages will show that Pgpool-II has joined the watchdog cluster as standby watchdog.

LOG:  watchdog cluster configured with 1 remote nodes
LOG:  watchdog remote node:0 on Linux_osspc16_9000:9000
LOG:  interface monitoring is disabled in watchdog
LOG:  IPC socket path: "/tmp/.s.PGPOOLWD_CMD.9000"
LOG:  watchdog node state changed from [DEAD] to [LOADING]
LOG:  new outbound connection to Linux_osspc16_9000:9000
LOG:  watchdog node state changed from [LOADING] to [INITIALIZING]
LOG:  watchdog node state changed from [INITIALIZING] to [STANDBY]
     LOG:  successfully joined the watchdog cluster as standby node
DETAIL:  our join coordinator request is accepted by cluster leader node "Linux_osspc16_9000"
LOG:  watchdog process is initialized
     
    

8.2.4. Try it out

Confirm to ping to the virtual IP address.

[user@someserver]$ ping 133.137.177.142
PING 133.137.177.143 (133.137.177.143) 56(84) bytes of data.
64 bytes from 133.137.177.143: icmp_seq=1 ttl=64 time=0.328 ms
64 bytes from 133.137.177.143: icmp_seq=2 ttl=64 time=0.264 ms
64 bytes from 133.137.177.143: icmp_seq=3 ttl=64 time=0.412 ms
   

Confirm if the Active server which started at first has the virtual IP address.

[root@osspc16]# ifconfig
eth0      ...

eth0:0    inet addr:133.137.177.143 ...

lo        ...
   

Confirm if the Standby server which started not at first doesn't have the virtual IP address.

[root@osspc20]# ifconfig
eth0      ...

lo        ...
   

Try to connect PostgreSQL by "psql -h delegate_IP -p port".

[user@someserver]$ psql -h 133.137.177.142 -p 9999 -l
   

8.2.5. Switching virtual IP

Confirm how the Standby server works when the Active server can't provide its service. Stop Pgpool-II on the Active server.

[root@osspc16]# {installed_dir}/bin/pgpool stop
   

Then, the Standby server starts to use the virtual IP address. Log shows:

    LOG:  remote node "Linux_osspc16_9000" is shutting down
LOG:  watchdog cluster has lost the coordinator node
    
LOG:  watchdog node state changed from [STANDBY] to [JOINING]
LOG:  watchdog node state changed from [JOINING] to [INITIALIZING]
LOG:  I am the only alive node in the watchdog cluster
HINT:  skipping stand for coordinator state
LOG:  watchdog node state changed from [INITIALIZING] to [MASTER]
LOG:  I am announcing my self as master/coordinator watchdog node
LOG:  I am the cluster leader node
DETAIL:  our declare coordinator message is accepted by all nodes
    LOG:  I am the cluster leader node. Starting escalation process
LOG:  watchdog: escalation started
    
LOG:  watchdog escalation process with pid: 59551 exit with SUCCESS.
   

Confirm to ping to the virtual IP address.

[user@someserver]$ ping 133.137.177.142
PING 133.137.177.143 (133.137.177.143) 56(84) bytes of data.
64 bytes from 133.137.177.143: icmp_seq=1 ttl=64 time=0.328 ms
64 bytes from 133.137.177.143: icmp_seq=2 ttl=64 time=0.264 ms
64 bytes from 133.137.177.143: icmp_seq=3 ttl=64 time=0.412 ms
   

Confirm that the Active server doesn't use the virtual IP address any more.

[root@osspc16]# ifconfig
eth0      ...

lo        ...
   

Confirm that the Standby server uses the virtual IP address.

[root@osspc20]# ifconfig
eth0      ...

eth0:0    inet addr:133.137.177.143 ...

lo        ...
   

Try to connect PostgreSQL by "psql -h delegate_IP -p port".

[user@someserver]$ psql -h 133.137.177.142 -p 9999 -l
   

8.2.6. More

8.2.6.1. Lifecheck

There are the parameters about watchdog's monitoring. Specify the interval to check wd_interval and the type of lifecheck wd_lifecheck_method . The hearbeat method specify the time to detect a fault wd_heartbeat_deadtime , the port number to receive wd_heartbeat_port , the interval to send wd_heartbeat_keepalive , the IP address or hostname of destination heartbeat_destination<emphasis>0</emphasis> and finally the destination port number heartbeat_destination_port<emphasis>0</emphasis> .

wd_lifecheck_method = 'heartbeat'
                                    # Method of watchdog lifecheck ('heartbeat' or 'query' or 'external')
                                    # (change requires restart)
wd_interval = 10
                                    # lifecheck interval (sec) > 0
wd_heartbeat_port = 9694
                                    # Port number for receiving heartbeat signal
                                    # (change requires restart)
wd_heartbeat_keepalive = 2
                                    # Interval time of sending heartbeat signal (sec)
                                    # (change requires restart)
wd_heartbeat_deadtime = 30
                                    # Deadtime interval for heartbeat signal (sec)
                                    # (change requires restart)
heartbeat_destination0 = 'host0_ip1'
                                    # Host name or IP address of destination 0
                                    # for sending heartbeat signal.
                                    # (change requires restart)
heartbeat_destination_port0 = 9694
                                    # Port number of destination 0 for sending
                                    # heartbeat signal. Usually this is the
                                    # same as wd_heartbeat_port.
                                    # (change requires restart)
    

8.2.6.2. Switching virtual IP address

There are the parameters for switching the virtual IP address. Specify switching commands if_up_cmd , if_down_cmd , the path to them if_cmd_path , the command executed after switching to send ARP request arping_cmd and the path to it arping_path .

if_cmd_path = '/sbin'
                                    # path to the directory where if_up/down_cmd exists
if_up_cmd = 'ip addr add $_IP_$/24 dev eth0 label eth0:0'
                                    # startup delegate IP command
if_down_cmd = 'ip addr del $_IP_$/24 dev eth0'
                                    # shutdown delegate IP command

arping_path = '/usr/sbin'           # arping command path

arping_cmd = 'arping -U $_IP_$ -w 1'
    

You can also bring up and bring down the virtual IP using arbitrary scripts specified wd_escalation_command and wd_de_escalation_command parameters.