KeepAliveD, Rackspace Cloud, Unicast - os dois LB entram no estado MASTER…?

Navegue suas respostas
2

Eu configurei dois balanceadores de carga HAProxy em um par ativo & > Passivo.

KeepAliveD será usado para failover entre os servidores MASTER e BACKUP.

Como a maioria das nuvens, o Multicast não é suportado, por isso não posso usar um IP virtual. Em vez disso, estou tentando usar o Unicast, que eu vi espalhado pela web como uma solução.

Meu problema é que a instância BACKUP KeepAliveD entra no estado MASTER imediatamente. Ele pode fazer ping no servidor MASTER, mas é como se ele não conseguisse perceber que realmente está funcionando.

Eu classificaria a mim mesmo como um administrador de sistema n00b, então, por favor, me perdoe. Por essa razão, espero que haja alguns erros óbvios que estou cometendo e que possam ser corrigidos facilmente ... 

      __[HAProxy Active, KeepAliveD MASTER, 10.179.66.95]
     /
----|
    |
     \__[HAProxy Passive, KeepAliveD BACKUP, 10.179.74.172]

Configs da seguinte forma ...

Versão do KeepAliveD em ambos

1.2.9 (suporte a unicast foi adicionado em 1.2.8 e corrigido em 1.2.9).

link

Nos dois servidores em /etc/sysctl.conf 

# Nonlocal bind for use with KeepAliveD. Allows this instance to take on a non-local IP for failover.
net.ipv4.ip_nonlocal_bind=1

KeepAliveD MASTER em /etc/keepalived/keepalived.conf 

! Configuration File for keepalived

global_defs {
    notification_email {
        [email protected]
    }
    notification_email_from [email protected]
        smtp_server 127.0.0.1
        smtp_connect_timeout 30
        router_id LB_MASTER_ACTIVE
    }

    # Define the script used to check if haproxy is still working
    vrrp_script chk_haproxy {
        script "killall -0 haproxy"   # verify the pid existance
        interval 2                    # check every 2 seconds
        weight 2                      # add 2 points of prio if OK
    }

    # Virtual interface.
   vrrp_instance VI_1 {
    state MASTER
    interface eth1
    virtual_router_id 51
    priority 101
    smtp_alert                  # Activate e-mail notifications.
    #advert_int 1

    authentication {
        auth_type PASS
        auth_pass 1111
    }

    # IP of myself and my peer for unicast based failover.
    vrrp_unicast_bind 10.179.66.95      # My IP.
    vrrp_unicast_peer 10.179.74.172     # The other's IP.

    # Check if HAProxy is running or not.
    track_script {
        chk_haproxy
    }
}

KeepAliveD BACKUP em /etc/keepalived/keepalived.conf 

! Configuration File for keepalived

global_defs {
    notification_email {
        [email protected]
    }
    notification_email_from [email protected]
    smtp_server 127.0.0.1
    smtp_connect_timeout 30
    router_id LB_BACKUP_PASSIVE
}

# Define the script used to check if haproxy is still working
vrrp_script chk_haproxy {
    script "killall -0 haproxy"
    interval 2
    weight 2
}

# Virtual interface.
vrrp_instance VI_1 {
    state BACKUP
    interface eth1
    virtual_router_id 51
    priority 100                # MASTER is priority 101.
    smtp_alert                  # Activate e-mail notifications.
    advert_int 1

    authentication {
        auth_type PASS
        auth_pass 1111
    }

    # IP of myself and my peer for unicast based failover.
    vrrp_unicast_bind 10.179.74.172     # My IP.
    vrrp_unicast_peer 10.179.66.95      # The other's IP.

    # Check if HAProxy is running or not.
    track_script {
        chk_haproxy
    }
}

As mensagens são registradas no KeepAliveD MASTER quando o KeepAliveD é iniciado, tail -f / var / log / messages 

Nov 28 10:54:02 mysql-read-lb-1 Keepalived[30158]: Starting Keepalived v1.2.7 (02/21,2013)
Nov 28 10:54:02 mysql-read-lb-1 Keepalived[30159]: Starting Healthcheck child process, pid=30161
Nov 28 10:54:02 mysql-read-lb-1 Keepalived[30159]: Starting VRRP child process, pid=30162
Nov 28 10:54:02 mysql-read-lb-1 Keepalived_healthcheckers[30161]: Interface queue is empty
Nov 28 10:54:02 mysql-read-lb-1 Keepalived_vrrp[30162]: Interface queue is empty
Nov 28 10:54:02 mysql-read-lb-1 Keepalived_healthcheckers[30161]: No such interface, eth1
Nov 28 10:54:02 mysql-read-lb-1 Keepalived_healthcheckers[30161]: No such interface, eth2
Nov 28 10:54:02 mysql-read-lb-1 Keepalived_healthcheckers[30161]: Netlink reflector reports IP 10.179.66.95 added
Nov 28 10:54:02 mysql-read-lb-1 Keepalived_healthcheckers[30161]: Netlink reflector reports IP 192.168.3.1 added
Nov 28 10:54:02 mysql-read-lb-1 Keepalived_healthcheckers[30161]: Netlink reflector reports IP fe80::be76:4eff:fe08:9227 added
Nov 28 10:54:02 mysql-read-lb-1 Keepalived_healthcheckers[30161]: Netlink reflector reports IP fe80::be76:4eff:fe08:8b4d added
Nov 28 10:54:02 mysql-read-lb-1 Keepalived_healthcheckers[30161]: Registering Kernel netlink reflector
Nov 28 10:54:02 mysql-read-lb-1 Keepalived_healthcheckers[30161]: Registering Kernel netlink command channel
Nov 28 10:54:02 mysql-read-lb-1 Keepalived_healthcheckers[30161]: Opening file '/etc/keepalived/keepalived.conf'.
Nov 28 10:54:02 mysql-read-lb-1 Keepalived_vrrp[30162]: No such interface, eth1
Nov 28 10:54:02 mysql-read-lb-1 Keepalived_vrrp[30162]: No such interface, eth2
Nov 28 10:54:02 mysql-read-lb-1 Keepalived_vrrp[30162]: Netlink reflector reports IP 10.179.66.95 added
Nov 28 10:54:02 mysql-read-lb-1 Keepalived_vrrp[30162]: Netlink reflector reports IP 192.168.3.1 added
Nov 28 10:54:02 mysql-read-lb-1 Keepalived_vrrp[30162]: Netlink reflector reports IP fe80::be76:4eff:fe08:9227 added
Nov 28 10:54:02 mysql-read-lb-1 Keepalived_vrrp[30162]: Netlink reflector reports IP fe80::be76:4eff:fe08:8b4d added
Nov 28 10:54:02 mysql-read-lb-1 Keepalived_vrrp[30162]: Registering Kernel netlink reflector
Nov 28 10:54:02 mysql-read-lb-1 Keepalived_vrrp[30162]: Registering Kernel netlink command channel
Nov 28 10:54:02 mysql-read-lb-1 Keepalived_vrrp[30162]: Registering gratuitous ARP shared channel
Nov 28 10:54:02 mysql-read-lb-1 Keepalived_vrrp[30162]: Opening file '/etc/keepalived/keepalived.conf'.
Nov 28 10:54:02 mysql-read-lb-1 Keepalived_healthcheckers[30161]: Configuration is using : 7559 Bytes
Nov 28 10:54:02 mysql-read-lb-1 Keepalived_vrrp[30162]: Truncating auth_pass to 8 characters
Nov 28 10:54:02 mysql-read-lb-1 Keepalived_vrrp[30162]: Configuration is using : 64400 Bytes
Nov 28 10:54:02 mysql-read-lb-1 Keepalived_vrrp[30162]: Using LinkWatch kernel netlink reflector...
Nov 28 10:54:02 mysql-read-lb-1 Keepalived_vrrp[30162]: VRRP sockpool: [ifindex(3), proto(112), fd(10,11)]
Nov 28 10:54:02 mysql-read-lb-1 Keepalived_healthcheckers[30161]: Using LinkWatch kernel netlink reflector...
Nov 28 10:54:03 mysql-read-lb-1 Keepalived_vrrp[30162]: VRRP_Instance(VI_1) Transition to MASTER STATE
Nov 28 10:54:04 mysql-read-lb-1 Keepalived_vrrp[30162]: VRRP_Instance(VI_1) Entering MASTER STATE
Nov 28 10:54:04 mysql-read-lb-1 Keepalived_vrrp[30162]: Remote SMTP server [127.0.0.1]:25 connected.
Nov 28 10:54:04 mysql-read-lb-1 Keepalived_vrrp[30162]: SMTP alert successfully sent.

As mensagens são registradas no KeepAliveD BACKUP quando o KeepAliveD é iniciado, tail -f / var / log / messages

Você notará que entra no estado MASTER logo de cara, mas deve permanecer em BACKUP ... 

Nov 28 10:57:35 load-balancer-1-passive Keepalived[25048]: Starting Keepalived v1.2.7 (02/21,2013)
Nov 28 10:57:35 load-balancer-1-passive Keepalived[25049]: Starting Healthcheck child process, pid=25050
Nov 28 10:57:35 load-balancer-1-passive Keepalived[25049]: Starting VRRP child process, pid=25052
Nov 28 10:57:35 load-balancer-1-passive Keepalived_healthcheckers[25050]: Interface queue is empty
Nov 28 10:57:35 load-balancer-1-passive Keepalived_healthcheckers[25050]: No such interface, eth1
Nov 28 10:57:35 load-balancer-1-passive Keepalived_healthcheckers[25050]: No such interface, eth2
Nov 28 10:57:35 load-balancer-1-passive Keepalived_healthcheckers[25050]: Netlink reflector reports IP 10.179.74.172 added
Nov 28 10:57:35 load-balancer-1-passive Keepalived_healthcheckers[25050]: Netlink reflector reports IP 192.168.3.2 added
Nov 28 10:57:35 load-balancer-1-passive Keepalived_healthcheckers[25050]: Netlink reflector reports IP fe80::be76:4eff:fe08:93fc added
Nov 28 10:57:35 load-balancer-1-passive Keepalived_healthcheckers[25050]: Netlink reflector reports IP fe80::be76:4eff:fe08:940c added
Nov 28 10:57:35 load-balancer-1-passive Keepalived_healthcheckers[25050]: Registering Kernel netlink reflector
Nov 28 10:57:35 load-balancer-1-passive Keepalived_healthcheckers[25050]: Registering Kernel netlink command channel
Nov 28 10:57:35 load-balancer-1-passive Keepalived_healthcheckers[25050]: Opening file '/etc/keepalived/keepalived.conf'.
Nov 28 10:57:35 load-balancer-1-passive Keepalived_vrrp[25052]: Interface queue is empty
Nov 28 10:57:35 load-balancer-1-passive Keepalived_vrrp[25052]: No such interface, eth1
Nov 28 10:57:35 load-balancer-1-passive Keepalived_vrrp[25052]: No such interface, eth2
Nov 28 10:57:35 load-balancer-1-passive Keepalived_vrrp[25052]: Netlink reflector reports IP 10.179.74.172 added
Nov 28 10:57:35 load-balancer-1-passive Keepalived_healthcheckers[25050]: Configuration is using : 7595 Bytes
Nov 28 10:57:35 load-balancer-1-passive Keepalived_vrrp[25052]: Netlink reflector reports IP 192.168.3.2 added
Nov 28 10:57:35 load-balancer-1-passive Keepalived_vrrp[25052]: Netlink reflector reports IP fe80::be76:4eff:fe08:93fc added
Nov 28 10:57:35 load-balancer-1-passive Keepalived_vrrp[25052]: Netlink reflector reports IP fe80::be76:4eff:fe08:940c added
Nov 28 10:57:35 load-balancer-1-passive Keepalived_vrrp[25052]: Registering Kernel netlink reflector
Nov 28 10:57:35 load-balancer-1-passive Keepalived_vrrp[25052]: Registering Kernel netlink command channel
Nov 28 10:57:35 load-balancer-1-passive Keepalived_vrrp[25052]: Registering gratuitous ARP shared channel
Nov 28 10:57:35 load-balancer-1-passive Keepalived_vrrp[25052]: Opening file '/etc/keepalived/keepalived.conf'.
Nov 28 10:57:35 load-balancer-1-passive Keepalived_vrrp[25052]: Truncating auth_pass to 8 characters
Nov 28 10:57:35 load-balancer-1-passive Keepalived_vrrp[25052]: Configuration is using : 64436 Bytes
Nov 28 10:57:35 load-balancer-1-passive Keepalived_vrrp[25052]: Using LinkWatch kernel netlink reflector...
Nov 28 10:57:35 load-balancer-1-passive Keepalived_vrrp[25052]: VRRP_Instance(VI_1) Entering BACKUP STATE
Nov 28 10:57:35 load-balancer-1-passive Keepalived_vrrp[25052]: VRRP sockpool: [ifindex(3), proto(112), fd(10,11)]
Nov 28 10:57:35 load-balancer-1-passive Keepalived_vrrp[25052]: Remote SMTP server [127.0.0.1]:25 connected.
Nov 28 10:57:35 load-balancer-1-passive Keepalived_healthcheckers[25050]: Using LinkWatch kernel netlink reflector...
Nov 28 10:57:35 load-balancer-1-passive Keepalived_vrrp[25052]: VRRP_Script(chk_haproxy) succeeded
Nov 28 10:57:35 load-balancer-1-passive Keepalived_vrrp[25052]: SMTP alert successfully sent.
Nov 28 10:57:38 load-balancer-1-passive Keepalived_vrrp[25052]: VRRP_Instance(VI_1) Transition to MASTER STATE
Nov 28 10:57:39 load-balancer-1-passive Keepalived_vrrp[25052]: VRRP_Instance(VI_1) Entering MASTER STATE
Nov 28 10:57:39 load-balancer-1-passive Keepalived_vrrp[25052]: Remote SMTP server [127.0.0.1]:25 connected.
Nov 28 10:57:39 load-balancer-1-passive Keepalived_vrrp[25052]: SMTP alert successfully sent.

Interfaces do servidor KeepAliveD MASTER, "ip a": 

[root@load-balancer-1-active keepalived]# ip a
1: lo: <LOOPBACK,UP,LOWER_UP> mtu 16436 qdisc noqueue state UNKNOWN 
    link/loopback 00:00:00:00:00:00 brd 00:00:00:00:00:00
    inet 127.0.0.1/8 scope host lo
    inet6 ::1/128 scope host 
       valid_lft forever preferred_lft forever
2: eth0: <BROADCAST,MULTICAST> mtu 1500 qdisc pfifo_fast state DOWN qlen 1000
    link/ether bc:76:4e:08:92:38 brd ff:ff:ff:ff:ff:ff
3: eth1: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1500 qdisc pfifo_fast state UP qlen 1000
    link/ether bc:76:4e:08:92:27 brd ff:ff:ff:ff:ff:ff
    inet 10.179.66.95/18 brd 10.179.127.255 scope global eth1
    inet6 fe80::be76:4eff:fe08:9227/64 scope link 
       valid_lft forever preferred_lft forever

Interfaces do servidor KeepAliveD BACKUP, "ip a": 

[root@load-balancer-1-passive ~]# ip a
1: lo: <LOOPBACK,UP,LOWER_UP> mtu 16436 qdisc noqueue state UNKNOWN 
    link/loopback 00:00:00:00:00:00 brd 00:00:00:00:00:00
    inet 127.0.0.1/8 scope host lo
    inet6 ::1/128 scope host 
       valid_lft forever preferred_lft forever
2: eth0: <BROADCAST,MULTICAST> mtu 1500 qdisc pfifo_fast state DOWN qlen 1000
    link/ether bc:76:4e:08:4f:b4 brd ff:ff:ff:ff:ff:ff
3: eth1: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1500 qdisc pfifo_fast state UP qlen 1000
    link/ether bc:76:4e:08:93:fc brd ff:ff:ff:ff:ff:ff
    inet 10.179.74.172/18 brd 10.179.127.255 scope global eth1
    inet6 fe80::be76:4eff:fe08:93fc/64 scope link 
       valid_lft forever preferred_lft forever

Sniff em MASTER

Como sugerido por "emy" abaixo, farejando pacotes para ver se a comunicação de verificação de saúde está ocorrendo conforme sugerido aqui: link 

[root@mysql-read-lb-1 ~]# tcpdump -vvv -n -i eth1 host 10.179.74.172
tcpdump: listening on eth1, link-type EN10MB (Ethernet), capture size 65535 bytes

Nada: (

Sniff no BACKUP 

[root@load-balancer-1-passive ~]# tcpdump -vvv -n -i eth1 host 10.179.66.95
tcpdump: listening on eth1, link-type EN10MB (Ethernet), capture size 65535 bytes

Nada: (

    
por Chris Rosillo 28.11.2013 / 12:06

1 resposta

1

Alguns palpites:

Tem certeza de que o tráfego de vrrp passa? Você poderia farejar (por exemplo, ngrep, tcpdump) na porta 112 se os pacotes forem recebidos? (Você deve ver um a cada segundo.) Consulte este link .

Se não, pode ser um problema de firewall.

    
por 28.11.2013 / 14:10

Tags

Implantação contínua do livro de receitas do Chef Como bloquear uma conexão de saída para um local específico