Technology Sharing

한어Русский языкEnglishFrançaisIndonesianSanskrit日本語DeutschPortuguêsΕλληνικάespañolItalianoSuomalainenLatina

• LVS keepalived

  • keepalived background

    • Auxiliary tools for LVS, mainly providing failover and health checks.

    • Working scenario: hot standby of the scheduler, health check of the node server and failover (deletion of entries).

    • The VRRP protocol is used to achieve high availability.

  • keepalived deployment

    • Topology

      • 

      • High availability is achieved between the two schedulers, and the health status of the web servers is checked, and load balancing is performed between the web servers.

    • Configuration of two schedulers

      • Prelude

        • Add or disable the firewall, and disable the kernel security mechanism

        • Turn off NetworkManager

          • If this service is not closed, the drifting IP cannot drift

      • Install keepalived and configure

        • After installation, set it to start automatically

          • systemctl disable keepalived

        • Copy the configuration file: /etc/keepalived/keepalived

        • Configuration file details

          • Global Configuration Section

            • global_defs {

            • notification_email {

              • This module indicates that an email notification needs to be sent when a keep switch operation occurs, and to which email addresses it is sent

            • [email protected]

              • Specific email addresses, one per line

            • [email protected]

            • [email protected]

            • }

            • notification_email_from [email protected]

              • This line defines who sends the alert email (Sender)

            • smtp_server 192.168.200.1

              • It means that you can use the mail service of that server to send mails. Usually you can use 127.0.0.1, which means that the local machine provides the service.

            • smtp_connect_timeout 30

              • Timeout for connecting to stmp service

            • router_id LVS_DEVEL

              • Machine identification: Identification of different servers, which does not affect the actual operation of the service and allows for quick location of mail services

            • vrrp_skip_check_adv_addr

              • No check is done on the master's declared address, and keepalived is run loosely

            • vrrp_strict

              • Enabling this option will not allow unicast communication unless the network is properly configured.

            • vrrp_garp_interval 0

              • Periodically announce IP address ownership. When set to 0, it will only be announced after switching.

            • vrrp_gna_interval 0

            • }

          • VRRP instance section

            • vrrp_instance VI_1 {

              • Instance name. A server can have multiple instances (application scenarios) which will be described later.

            • state MASTER

              • Determine the initial status of the role

                • init position in the vrrp election cycle

            • interface eth0

              • The network card that carries the VIP is based on your own

            • virtual_router_id 51

              • Each instance must use the same VRID to join the same hot standby group.

            • priority 100

              • Because VRRP uses priority to elect the master based on the range (0~255), 0 is reserved, 255 (IP address owner)

            • advert_int 1

              • The interval, in seconds, between master announcement messages (to tell other backups) that the master is still alive.

            • authentication {

              • When other instances want to join the hot standby group, password authentication is required.

            • auth_type PASS

            • auth_pass 1111

            • }

            • virtual_ipaddress {

              • Virtual VIP address: Each virtual VIP represents a cluster

            • 192.168.200.16

            • 192.168.200.17

            • 192.168.200.18

            • }

            • }

          • Virtual Server Section

            • virtual_server 192.168.200.100 443 {

              • Corresponding to the VIP in the instance, it is the beginning of a cluster

            • delay_loop 6

              • Health check interval

            • lb_algo rr

              • Scheduling Algorithm

                • rr|wrr|lc|wlc|lblc|sh|dh

                  Previous Next