Thank you very much Simon. I will probably try the nsupdate route, the first option you presented does seem to be the least attractive, or at least, ungentle, way to proceed.
-----Original Message-----
From:
[hidden email] [mailto:
[hidden email]] On Behalf Of Simon Hobson
Sent: Thursday, April 02, 2015 11:27 AM
To: Users of ISC DHCP;
[hidden email]
Subject: Re: problem with static range in dynamic table
"Cuttler, Brian (HEALTH)" <
[hidden email]> wrote:
> Except-I set my available address range to 10.57.36.10 - 10.57.39.150, as I have some machines that needed static entries.
>
> I'm finding that when I update the forward and reverse table with a static entry, in the 10.57.39.151-250 range, that the tables NEVER propagate. Updating SOA serial, restarting the server is not the solution, the dns slaves never pull the tables, even after the table expiration date.
>
> The work-around, which is really not supportable, has been to remove the tables from the slave servers and restart named on them.
It's a BIND question rather than DHCP, but ...
Oh, notices it's copied to the BIND list I'm not subscribed to.
From what you write, it sounds like you are updating the zone files, then restarting the server. This does *not* work (properly) for dynamically updated zones.
There are 3 (that I can think of) ways to do it :
1) Stop the server (politely), remove the journal (.jnl) files, edit the zone file, start the server.
2) Use "rndc freeze your_zone_goes_here", edit the zone file, then use "rndc unfreeze your_zone_goes_here".
3) Use nsupdate.
1) is a bit brute force. When you stop the server, it flushes it's caches to the zone file. If you remove the journal files, then it'll use only the zone file when it starts up.
2) This is the "polite" way to edit the zone files. When you freeze a zone, BIND flushes changes to the zonefile, removes the journal files, and you can then edit the zone file - the zone will continue being served with the in-memory copy. When you unfreeze the zone, the zone file is read into memory.
3) This is the best way on a busy zone as it allows all the other processing to carry on normally - including updates from DCHP.
Options 1 and 2 also remove the ability to use incremental zone transfers - when the journal files are removed, the zone history is lost, so the server can't create an incremental update from older versions.
_______________________________________________
dhcp-users mailing list
[hidden email]
https://lists.isc.org/mailman/listinfo/dhcp-users_______________________________________________
dhcp-users mailing list
[hidden email]
https://lists.isc.org/mailman/listinfo/dhcp-users
Locked
