Question

[Glenn Satchell]

Hi Leslie,

You've got "split 0" in the failover section - this causes dhcpd to 
prefer one of the servers for all renewals, so this might be a reason 
only one server is logging any renewals. "split 128" does a 50-50 
balance between the servers and is the more common setting.

"not responding (recovering)" sounds like the failover protocol hasn't 
synced between the servers, so it won't respond to any requests while 
it's in that state to prevent both dhcpd servers handing out conflicting 
responses, eg both hand out the same IP to different clients. dhcpd will 
log messages to syslog when failover states change.

You could use tcpdump or wireshark to look and see if there is any 
traffic on tcp port 647 between the servers. Also the dhcpd.conf man 
page in the section titled FAILOVER STARTUP goes through the steps and 
may help debug what's going on.

I think you might have already done this, bu check there is no local 
firewall rules on the server blocking failover traffic between those 
ports.

dhcpd.leases is the current lease file, you can run tail -f to see 
updates being appended. Every hour it renames this to dhcpd.leases~ and 
writes out a new dhcpd.leases from the in-memory copy of the data. 
dhcpd.leases is only ever appended to, so you can have multiple entries 
for the same IP address - the last one is the active state. When the 
file gets re-written it removes all the duplicate records.

regards,
Glenn

On 2022-06-03 13:47, Leslie Rhorer wrote:

> I don't think so.  The secondary server seems to have gone completely 
> silent, now, but I am getting a ton of them on the primary server, now.
> 
> On 6/2/2022 8:06 PM, Richard L. Hamilton wrote: 
> https://serverfault.com/questions/313008/isc-dhcp-fails-to-sync-leases-between-peers
> 
> Probably the two DHCP servers weren't able to talk to each other, not 
> surprising when one was having problems.
> 
> On Jun 2, 2022, at 7:42 PM, Leslie Rhorer 
> <lesrhorer at siliconventures.net> wrote:
> 
> During troubleshooting of my recent issue, I got tons of duplicates of 
> errors like the following.  They seem to have stopped, now, but I am 
> curious what they meant.
> 
> Jun  1 00:31:56 Backup dhcpd[15785]: DHCPDISCOVER from 
> 60:01:94:f0:41:48 via enp11s0: not responding (recovering)
> 
> --
> ISC funds the development of this software with paid support 
> subscriptions. Contact us at https://www.isc.org/contact/ for more 
> information.
> 
> dhcp-users mailing list
> dhcp-users at lists.isc.org
> https://lists.isc.org/mailman/listinfo/dhcp-users