170 likes | 313 Views
Redundant Routers. Jordan Curzon Charles Smutz. Purpose. Redundancy/Reliability Remove Single points of failure Manual Load Balancing Use both internet Links Other Services DNS DHCP Network Management. Redundancy through HA. Availability Monitoring Resource Management IP Takeover
E N D
Redundant Routers Jordan Curzon Charles Smutz
Purpose • Redundancy/Reliability • Remove Single points of failure • Manual Load Balancing • Use both internet Links • Other Services • DNS • DHCP • Network Management
Redundancy through HA • Availability Monitoring • Resource Management • IP Takeover • Gratuitous ARP • http://www.ultramonkey.org/3/ip_address_takeover.html • Permanent/Service Addresses
Switch Redundancy • Distribution switches as single point of failure
Other Redundancy Issues • Power • Cable Runs
Routing with Iptables • Standard linux software router • Firewall with Iptables rules • NAT (MASQ) • Port Forwarding (DNAT) • Same Rule Set on each router • Separate chain for packets destined for each public address (P1, P2) • Same port can be use on each address • Shorter chains for better performance
DNS • Master/Slave • Slave loads zones from Master via zone transfer • Resolve requests instead of forwarding to BYU servers
DHCP • Failover • Primary and Secondary Server split pool • Same config file—both have static leases • Included sym link to primary/secondary config • Proprietary communication over arbitrary ports (519,520) • Compile from Source!
Manual Load Balancing • Static leases point to R1, dynamic to R2 for primary default gateway
Config Versioning/Syncing • Murcurial
Network Management • Ping permanent Address • Optionally send Email with HA • Optionally use SNMP
Live Demo • Off to 385