Layer 4 server load balancing hot backup configuration example

Network requirements

In Figure 49, physical servers Server A, Server B, and Server C provide FTP services, and are in descending order of hardware configuration.

Configure server load balancing on the LB devices to distribute user requests among the servers based on their hardware performance, and use health monitoring to monitor reachability of the servers.

For high availability purposes, implement hot backup for the LB devices. To ensure traffic continuity, enable synchronization for session extension information and sticky entries on the LB devices.

Figure 49: Network diagram

Configuration procedure

  1. Configure the LB devices:

    • Configure IRF.

      # Set up an IRF fabric for the two LB devices with member IDs 1 and 2. (Details not shown.)

      For more information about IRF, see Virtual Technologies Configuration Guide.

    • Configure Reth interfaces.

      # Create Reth 1, and assign an IP address to Reth 1. Assign member interfaces to Reth 1, and set their priorities.

      [LB] interface reth 1
      [LB-Reth1] ip address 61.159.4.100 24
      [LB-Reth1] member interface gigabitethernet 1/0/1 priority 20
      [LB-Reth1] member interface gigabitethernet 2/0/1 priority 10
      [LB-Reth1] quit
      

      # Create Reth 2, and assign an IP address to Reth 1. Assign member interfaces to Reth 2, and set their priorities.

      [LB] interface reth 2
      [LB-Reth2] ip address 192.168.1.100 24
      [LB-Reth2] member interface gigabitethernet 1/0/2 priority 20
      [LB-Reth2] member interface gigabitethernet 2/0/2 priority 10
      [LB-Reth2] quit
      
    • Configure a redundancy group.

      # Create track entries to monitor the link state of the Ethernet interfaces.

      [LB] track 1 interface gigabitethernet 1/0/1
      [LB] track 2 interface gigabitethernet 1/0/2
      [LB] track 3 interface gigabitethernet 2/0/1
      [LB] track 4 interface gigabitethernet 2/0/2
      

      # Create the redundancy group bkp and assign Reth 1 and Reth 2 to the redundancy group.

      [LB] redundancy group bkp
      [LB-redundancy-group-bkp] member interface reth 1
      [LB-redundancy-group-bkp] member interface reth 2
      

      # Create node 1 (with priority 100) for the redundancy group bkp and bind node 1 to IRF member device 1. Associate track entries 1 and 2 with node 1.

      [LB-redundancy-group-bkp] node 1
      [LB-redundancy-group-bkp-node-1] priority 100
      [LB-redundancy-group-bkp-node-1] bind slot 1
      [LB-redundancy-group-bkp-node-1] track 1 interface gigabitethernet 1/0/1
      [LB-redundancy-group-bkp-node-1] track 2 interface gigabitethernet 1/0/2
      [LB-redundancy-group-bkp-node-1] quit
      

      # Create node 2 (with priority 50) for the redundancy group bkp and bind node 2 to IRF member device 2. Associate track entries 3 and 4 with node 2.

      [LB-redundancy-group-bkp] node 2
      [LB-redundancy-group-bkp-node2] priority 50
      [LB-redundancy-group-bkp-node2] bind slot 2
      [LB-redundancy-group-bkp-node2] track 3 interface gigabitethernet 2/0/1
      [LB-redundancy-group-bkp-node2] track 4 interface gigabitethernet 2/0/2
      [LB-redundancy-group-bkp-node2] quit
      [LB-redundancy-group-bkp] quit
      
    • Enable session synchronization for stateful failover.

      [LB] session synchronization enable
      
    • Configure a server farm.

      # Create the ICMP-type NQA template t1.

      [LB] nqa template icmp t1
      [LB-nqatplt-icmp-t1] quit
      

      # Create the server farm sf, and specify the scheduling algorithm as weighted round robin and health monitoring method as t1.

      [LB] server-farm sf
      [LB-sfarm-sf] predictor round-robin
      [LB-sfarm-sf] probe t1
      [LB-sfarm-sf] quit
      
    • Configure real servers.

      # Create the real server rs1 with IPv4 address 192.168.1.1, port number 21, and weight 150, and add it to the server farm sf.

      [LB] real-server rs1
      [LB-rserver-rs1] ip address 192.168.1.1
      [LB-rserver-rs1] port 21
      [LB-rserver-rs1] weight 150
      [LB-rserver-rs1] server-farm sf
      [LB-rserver-rs1] quit
      

      # Create the real server rs2 with IPv4 address 192.168.1.2, port number 21, and weight 120, and add it to the server farm sf.

      [LB] real-server rs2
      [LB-rserver-rs2] ip address 192.168.1.2
      [LB-rserver-rs2] port 21
      [LB-rserver-rs2] weight 120
      [LB-rserver-rs2] server-farm sf
      [LB-rserver-rs2] quit
      

      # Create the real server rs3 with IPv4 address 192.168.1.3, port number 21, and weight 80, and add it to the server farm sf.

      [LB] real-server rs3
      [LB-rserver-rs3] ip address 192.168.1.3
      [LB-rserver-rs3] port 21
      [LB-rserver-rs3] weight 80
      [LB-rserver-rs3] server-farm sf
      [LB-rserver-rs3] quit
      
    • Configure a virtual server.

      # Create the TCP virtual server vs with VSIP 61.159.4.100, and specify its default master server farm sf. Enable synchronization for session extension information and sticky entries, and enable the virtual server.

      [LB] virtual-server vs type tcp
      [LB-vs-tcp-vs] virtual ip address 61.159.4.100
      [LB-vs-tcp-vs] default server-farm sf
      [LB-vs-tcp-vs] connection-sync enable
      [LB-vs-tcp-vs] sticky-sync enable
      [LB-vs-tcp-vs] service enable
      [LB-vs-tcp-vs] quit
      
  2. Configure the physical servers:

    # Specify the default gateway 192.168.1.100 for physical servers Server A, Server B, and Server C. (Details not shown.)

Verifying the configuration

# Display information about the redundancy group bkp.

[LB] display redundancy group bkp
Redundancy group bkp (ID 1):
  Node ID      Slot          Priority   Status        Track weight
  1            Slot1         100        Primary       255
  2            Slot2         50         Secondary     255

Preempt delay time remained     : 0    min
Preempt delay timer setting     : 1    min
Remaining hold-down time        : 0    sec
Hold-down timer setting         : 1    sec
Manual switchover request       : No

Member interfaces:
    Reth1                 Reth2
Member failover groups:

Node 1:
  Track info:
    Track    Status       Reduced weight     Interface
    1        Positive     255                GE1/0/1
    2        Positive     255                GE1/0/2
Node 2:
  Track info:
    Track    Status       Reduced weight     Interface
    3        Negative     255                GE2/0/1
    4        Negative     255                GE2/0/2

# Display brief information about all real servers.

[LB] display real-server brief
Real server      Address              Port  State     VPN instance     Server farm
rs1              192.168.1.1          21    Active                     sf
rs2              192.168.1.2          21    Active                     sf
rs3              192.168.1.3          21    Active                     sf

# Display detailed information about all server farms.

[LB] display server-farm
Server farm: sf
  Description:
  Predictor: Round robin
  Proximity: Disabled
  NAT: Enabled
  SNAT pool:
  Failed action: Keep
  Active threshold: Disabled
  Slow-online: Disabled
  Selected server: Disabled
  Probe information:                                                            
    Probe success criteria: All                                                 
    Probe method:                                                               
    t1
  Total real server: 3
  Active real server: 3
  Real server list:
  Name             State    VPN instance     Address              Port  Weight Priority
  rs1              Active                   192.168.1.1          21    150    4
  rs2              Active                   192.168.1.2          21    120    4
  rs3              Active                   192.168.1.3          21    80     4

# Display detailed information about all virtual servers.

[LB] display virtual-server
Virtual server: vs
  Description:
  Type: TCP
  State: Active
  VPN instance:
  Virtual IPv4 address: 61.159.4.100/32
  Virtual IPv6 address: --
  Port: 0
  Primary server farm: sf (in use)
  Backup server farm:
  Sticky:
  LB policy:
  Connection limit: --
  Rate limit:
    Connections: --
    Bandwidth: --
    Inbound bandwidth: --
    Outbound bandwidth: --
  Connection synchronization: Enabled
  Sticky synchronization: Enabled
  Bandwidth busy protection: Disabled
  Interface bandwidth statistics: Disabled
  Route advertisement: Disabled