[ednog] Anyone using the ISC DHCPD failover protocol?

[Todd Caine]

Yeah, you will need to keep your configuration in sync for the  
failover pair and restart all servers that have a modified config  
file.  We are storing our DHCP configuration in a relational database  
and autogenerating the dhcpd.conf files using a Perl script.  The  
idea is to keep the master configuration data in a central location  
so that we can update it in one place and push it out to the DHCP  
servers in an automated fashion.

You might also want to think about breaking up the dhcpd.conf file  
into seperate files and use the include statement in dhcpd.conf to  
pull in all the pieces.  You might move all of your shared-network/ 
subnet/pool declarations to a single file that will be copied to the  
relevant DHCP servers.

At this point we are still restarting the DHCP servers manually to  
get them to re-read the updated configuration files, which is a bit  
tedious.

Todd

On Sep 21, 2005, at 11:40 AM, Chris Fabri wrote:
>
> I'm also looking into this.   My understanding is that beyond the
> dhcpd.config and leases issues, you also need to keep your configs in
> sync, and when changes are made, still restart both/all dhcpd  
> processes.
>
> How have those who've deployed this managed this portion?    Or is  
> this
> simpler than what I'm thinking?     chris
-------------- next part --------------
An HTML attachment was scrubbed...
URL: https://puck.nether.net/pipermail/ednog/attachments/20050921/3ed0e888/attachment.html