Failover again (sigh)

[Tina Siegenthaler]

Hi list

Still trying to get failover working as I suppose it should (well,  
maybe it's me who's wrong here)... and ran into a new problem.

I decided to try and put my secondary into partner-down state (after  
shutting down the primary), to test if failover is really working. No  
sense in finding out that it  actually doesn't when one server REALLY  
has failed, is it...

So I stopped dhcpd on the secondary (DHCP-0352) and edited the  
dhcpd.leases file to read:

failover peer "DHCP-0352" state {
   my state partner-down;
   partner state normal at 5 2006/12/15 13:01:03;
   mclt 3600;

as it says in the manpage. However, as soon as I restart dhcpd, the  
state will switch back to communications-interrupted:

failover peer "DHCP-0352" state {
   my state communications-interrupted at 3 2007/01/03 14:16:07;
   partner state normal at 5 2006/12/15 13:01:03;
   mclt 3600;

Is it supposed to do that? Shouldn't it rather stay in partner-down?


Second question: It still seems that the primary (when both servers  
are running normally) handles nearly all leases. For example, I get  
the following log entries on my primary (DHCP-0465):

Jan  3 11:16:44 DHCP-0465 dhcpd: DHCPDISCOVER from 00:14:51:a9:6e:10  
(Erika-0494) via 130.60.225.1: load balance to peer DHCP-0465
Jan  3 11:16:45 DHCP-0465 dhcpd: DHCPDISCOVER from 00:14:51:a9:6e:10  
(Erika-0494) via 130.60.225.1: load balance to peer DHCP-0465
Jan  3 11:16:47 DHCP-0465 dhcpd: DHCPDISCOVER from 00:14:51:a9:6e:10  
(Erika-0494) via 130.60.225.1: load balance to peer DHCP-0465
Jan  3 11:16:52 DHCP-0465 dhcpd: DHCPDISCOVER from 00:14:51:a9:6e:10  
(Erika-0494) via 130.60.225.1
Jan  3 11:16:53 DHCP-04654 dhcpd: DHCPOFFER on 130.60.225.34 to  
00:14:51:a9:6e:10 (Erika-0494) via 130.60.225.1
Jan  3 11:16:54 DHCP-0465 dhcpd: DHCPREQUEST for 130.60.225.34  
(130.60.23.4) from 00:14:51:a9:6e:10 (Erika-0494) via 130.60.225.1
Jan  3 11:16:54 DHCP-0465 dhcpd: DHCPACK on 130.60.225.34 to  
00:14:51:a9:6e:10 (Erika-0494) via 130.60.225.1

It seems to me that the primary is trying to pass the "DISCOVER" of  
"Erika-0494" to the secondary. Right? But the secondary obviously  
isn't responding, so the primary finally handles the request on its  
own. I'm not seeing any log entries at 11:16 on the secondary.

On the other hand, if the situation is the other way round, the  
primary will take over a request from the secondary:

Log on secondary (DHCP-0352):

Jan  3 12:26:34 DHCP-0352 dhcpd: DHCPDISCOVER from 00:30:65:81:c9:e4  
via en0: load balance to peer DHCP-0352
Jan  3 12:26:35 DHCP-0352 dhcpd: DHCPDISCOVER from 00:30:65:81:c9:e4  
via en0: load balance to peer DHCP-0352

Log on primary (DHCP-0465):

Jan  3 12:26:35 zool-admin04 dhcpd: DHCPDISCOVER from  
00:30:65:81:c9:e4 via en0
Jan  3 12:26:36 zool-admin04 dhcpd: DHCPOFFER on 130.60.23.56 to  
00:30:65:81:c9:e4 via en0
Jan  3 12:26:37 zool-admin04 dhcpd: DHCPREQUEST for 130.60.23.56  
(130.60.23.4) from 00:30:65:81:c9:e4 via en0
Jan  3 12:26:37 zool-admin04 dhcpd: DHCPACK on 130.60.23.56 to  
00:30:65:81:c9:e4 via en0

All in all, dhcpd.log on the primary is filled with entries  
(DISCOVER, OFFER, REQUEST, ACK every other minute or so), whereas I  
get nearly none on the secondary (one entry every hour or so).
The primary was my first DHCP server and was running for quite some  
time all alone, without any failover. The secondary was set up later.  
Would it be a good idea to delete the dhpcd.leases file on both  
servers and restart dhcpd? I'd do that on a weekend of course...


TIA, Tina