dhcp failover with 'shared network'

suprstar · 10-04-2011, 08:47 AM

I have 2 subnets in a shared network, I'm deciding which ip's to give out by the mac address. When 1 server is working by itself, no problems. When I add the failover piece, I get 'load balancing to peer' in syslog, but syslog on server 2 shows nothng. server 1 & 2 are identical except the failover piece has 'secondary' on server 2, ips/ports switched, 'split' commented out on server 2. With a tcpdump, I can see traffic going back and forth, dhcpd is running, just the failover is acting like it can't hear server 1....

The relevant chunks of dhcpd.conf are:

Code:

failover peer "dhcp-failover" {
  primary;
  address 10.250.0.8;
  port 519;
  peer address 10.250.0.9;
  peer port 520;
  max-response-delay 60;
  max-unacked-updates 10;
  mclt 600;
  split 128;
  load balance max seconds 3;
}

log(info, concat( "DEBUG-", binary-to-ascii(16,8,":",substring(hardware,1,3))));
class "RD2-MGMT" {
        match if binary-to-ascii(16,8,":",substring(hardware, 1, 3)) = "0:10:e7";
}

subnet 10.250.0.0 netmask 255.255.255.192 {}

shared-network customers {
        # This is one subnet 10.0.0.0/22
        subnet 10.0.0.0 netmask 255.255.252.0  {
                default-lease-time 86400;
                max-lease-time 86400;
                option subnet-mask 255.255.252.0;
                option broadcast-address 10.0.3.255;
                option routers 10.0.0.1;
 
                pool {
                  failover peer "dhcp-failover";
                  deny members of "RD2-MGMT";
                  range 10.0.0.10 10.0.3.199;
                }

        }

        # This is the another subnet 192.168.0.0/21
        subnet 192.168.0.0 netmask 255.255.248.0 {
                default-lease-time 86400;
                max-lease-time 86400;
                option subnet-mask 255.255.248.0;
                option broadcast-address 192.168.7.255;
                option routers 192.168.0.1;

                pool {
                    failover peer "dhcp-failover";
                    range 192.168.0.1 192.168.7.252;
                    allow members of "RD2-MGMT";
                }
        }
}

Thanks in advance,

suprstar · 10-04-2011, 03:23 PM

An update:

Is there something wrong with my failover declarations? I'm defining ports 519 and 520 as per the numerous examples on the net. But when I tcpdump it later on, I see that the primary server is using some arbitrary port, while the second server is using the port I tell it to. A tcpdump on both servers shows the same thing..

primary server:

Code:

failover peer "dhcp-failover" {
  primary;
  address 10.250.0.8;
  port 519;
  peer address 10.250.0.9;
  peer port 520;
  max-response-delay 60;
  max-unacked-updates 10;
  mclt 600;
  split 128;
  load balance max seconds 3;
}

secondary server:

Code:

failover peer "dhcp-failover" {
  secondary;
  address 10.250.0.9;
  port 520;
  peer address 10.250.0.8;
  peer port 519;
  max-response-delay 60;
  max-unacked-updates 10;
  mclt 600;
  #split 128;
  load balance max seconds 3;
}

tcpdump on primary server:
------------------------------------------
16:13:34.960694 IP svc2.local.520 > svc1.local.33077: Flags [P.], seq 265:277, ack 420, win 124, options [nop,nop,TS val 1038201325 ecr 1037458126], length 12
16:13:34.960715 IP svc1.local.33077 > svc2.local.520: Flags [.], ack 277, win 142, options [nop,nop,TS val 1037458614 ecr 1038201325], length 0
16:13:53.010044 IP svc1.local.33077 > svc2.local.520: Flags [P.], seq 420:432, ack 277, win 142, options [nop,nop,TS val 1037463126 ecr 1038201325], length 12
16:13:53.010282 IP svc2.local.520 > svc1.local.33077: Flags [.], ack 432, win 124, options [nop,nop,TS val 1038205837 ecr 1037463126], length 0
16:13:54.960715 IP svc2.local.520 > svc1.local.33077: Flags [P.], seq 277:289, ack 432, win 124, options [nop,nop,TS val 1038206325 ecr 1037463126], length 12
16:13:54.960736 IP svc1.local.33077 > svc2.local.520: Flags [.], ack 289, win 142, options [nop,nop,TS val 1037463614 ecr 1038206325], length 0
16:14:13.018180 IP svc1.local.33077 > svc2.local.520: Flags [P.], seq 432:444, ack 289, win 142, options [nop,nop,TS val 1037468128 ecr 1038206325], length 12
16:14:13.018350 IP svc2.local.520 > svc1.local.33077: Flags [.], ack 444, win 124, options [nop,nop,TS val 1038210839 ecr 1037468128], length 0
16:14:14.960707 IP svc2.local.520 > svc1.local.33077: Flags [P.], seq 289:301, ack 444, win 124, options [nop,nop,TS val 1038211325 ecr 1037468128], length 12
16:14:14.960728 IP svc1.local.33077 > svc2.local.520: Flags [.], ack 301, win 142, options [nop,nop,TS val 1037468614 ecr 1038211325], length 0

tcpdump on secondary server:
------------------------------------------
16:13:35.002085 IP svc2.local.520 > svc1.local.33077: Flags [P.], seq 1:13, ack 12, win 124, options [nop,nop,TS val 1038201325 ecr 1037458126], length 12
16:13:35.002372 IP svc1.local.33077 > svc2.local.520: Flags [.], ack 13, win 142, options [nop,nop,TS val 1037458614 ecr 1038201325], length 0
16:13:53.051706 IP svc1.local.33077 > svc2.local.520: Flags [P.], seq 12:24, ack 13, win 142, options [nop,nop,TS val 1037463126 ecr 1038201325], length 12
16:13:53.051727 IP svc2.local.520 > svc1.local.33077: Flags [.], ack 24, win 124, options [nop,nop,TS val 1038205837 ecr 1037463126], length 0
16:13:55.002107 IP svc2.local.520 > svc1.local.33077: Flags [P.], seq 13:25, ack 24, win 124, options [nop,nop,TS val 1038206325 ecr 1037463126], length 12
16:13:55.002352 IP svc1.local.33077 > svc2.local.520: Flags [.], ack 25, win 142, options [nop,nop,TS val 1037463614 ecr 1038206325], length 0
16:14:13.059817 IP svc1.local.33077 > svc2.local.520: Flags [P.], seq 24:36, ack 25, win 142, options [nop,nop,TS val 1037468128 ecr 1038206325], length 12
16:14:13.059837 IP svc2.local.520 > svc1.local.33077: Flags [.], ack 36, win 124, options [nop,nop,TS val 1038210839 ecr 1037468128], length 0
16:14:15.002101 IP svc2.local.520 > svc1.local.33077: Flags [P.], seq 25:37, ack 36, win 124, options [nop,nop,TS val 1038211325 ecr 1037468128], length 12
16:14:15.002326 IP svc1.local.33077 > svc2.local.520: Flags [.], ack 37, win 142, options [nop,nop,TS val 1037468614 ecr 1038211325], length 0