2

On our DHCP server I see this quite a bit on some of our leases and I am not sure why its being reported like this. 

lease <IP Address> {
starts 5 2015/08/21 13:22:20;
ends 5 2015/08/21 13:22:18;
tstp 5 2015/08/21 13:22:20;
tsfp 5 2015/08/21 13:22:20;
atsfp 5 2015/08/21 13:22:20;
binding state backup;
hardware ethernet <MAC address>;
uid "\001\000\017\224A\271`";

As you can see here, it ended before it began by 2 seconds. This is unlike the lease below where it appears to be more accurate as far as times go and is in line with the settings configured by the conf file, which is 48 hours default lease time. 

lease <IP Address> {
starts 4 2015/10/22 01:59:48;
ends 6 2015/10/24 01:59:48;
tstp 6 2015/09/19 01:57:16;
tsfp 0 2015/10/25 01:59:48;
atsfp 0 2015/10/25 01:59:48;
cltt 3 2015/09/16 01:57:16;
binding state active;
next binding state expired;
hardware ethernet <MAC Address>
uid "\001\000\017\224A\3010";

Can anyone explain why this can happen and what does it mean as far as lease validity?

user53029
  • 629
  • 3
  • 14
  • 36

1 Answers1

2

Looking at the man-page for the leases file, your first block is a lease that can be handed out by the secondary server in a failover configuration; the second block is a lease that has been handed out to a client. If this isn't a failover configuration, did you copy the lease file from somewhere?

Andreas Turriff
  • 74
  • 5
  • Yes, I found that out shortly after I posted by reading about the "binding state". Thanks for clarifying. It is a failover config. So why does the lease in the first block now have to be handed out by the secondary server? Why not the primary server? Is that just the way a failover pair in linux works? – user53029 Oct 22 '15 at 18:35
  • Having never actually played with dhcp failover, I'm not certain. I suspect the answer is that the two servers split up the address pool between them. – Andreas Turriff Oct 23 '15 at 19:18