Primary server stuck in "recovering" on restarts

[Patrick Trapp]

We make our changes on a DHCP server (we call it our staging server) that is not in service but shares the configuration of the servers that are. That allows us to test the config with a matching DHCP server and avoid edits on the production servers that could inadvertently break things. No one logs into the production servers to make their changes.

On the DHCP staging server, we 
1. make our edits
2. test the changes using the -t parameter
3. commit the changes to a local git repository
4. push the git changes to a central repository
5. call a script on each of the two production servers, staggered with the primary being first.

The script on the production servers that is called
1. pulls the git changes from the central repository to the production server
2. restarts the DHCP service

---------------------------------------------------------------

These servers provide boot parameters in addition to addresses to the client devices. At any given time, we may have up to eight valid combinations of boot parameters in use. For a long time, we used individual host entries to associate the desired boot parameters with the MAC address of the device. In the spirit of "if it's not broken, don't fix it", I really didn't mess with it for a long time, but in the last year we have started using groups and host entries to accomplish the same thing with a lot less confusion.

The specific configuration changes are not relevant to the issue, I think. If I restart the servers using my "dhcp-sync" script without making any changes, the primary server will be stuck in "recovering" when the script completes. And it will apparently stay that way until someone does something to fix it. It no longer goes back into service on its own.

Patrick

________________________________________
From: dhcp-users-bounces at lists.isc.org [dhcp-users-bounces at lists.isc.org] on behalf of Steven Carr [sjcarr at gmail.com]
Sent: Wednesday, July 30, 2014 11:41 AM
To: Users of ISC DHCP
catSubject: Re: Primary server stuck in "recovering" on restarts

On 30 July 2014 17:13, Patrick Trapp <ptrapp at nex-tech.com> wrote:
> We have a lot of minor tweaks to host entries and group entries in our
> configuration - maybe up to a dozen a day - so we have scripted out some of
> the steps for synchronizing our configurations between the servers. The last
> script stops the primary server, restarts it, stops the secondary server,
> and restarts it. This has worked great for years and was working on 4.2.2
> (the most recent version prior to this patch/upgrade).

Can you post exactly what your script is doing? commands etc. and some
examples of what the configuration changes are?

Usually all you need to do is restart the dhcpd process (don't mess
with OMAPI), staggered is usually best, they will briefly go into
communications interrupted but should then come back active.

Steve
_______________________________________________
dhcp-users mailing list
dhcp-users at lists.isc.org
https://lists.isc.org/mailman/listinfo/dhcp-users