Am 2005-07-14 09:55:10, schrieb /dev/rob0: > A couple other tips, in case you've not managed a dynamic zone before: Right. > 1. ALL zone changes will have to be done using nsupdate(8), unless you > stop named and delete the zone's journal file (/var/named/*.jnl or the > *.jnl in whatever directory you are using.) Do not attempt manual edits > of your dynamic zone's file without doing this. Also, understand that > deletion of the journal inhibits IXFR; slaves will have to transfer the > whole zone, not just the changed data. > > 2. Comments in your nicely-edited zone file are gone. I had made mine I was already thinking on this issue. My curent zonefiles are autogenerated by a script from a postgresql. There will be no problem to change the script to use 'nsupdate', ok, the script will be much bigger and slower, but it schould work too. > such a pretty one, with human-friendly sorting of records, and comments; Me too. :-) > named and nsupdate ate them. :) $INCLUDE's and $GENERATE's will be > included and generated into the zone file. named will do its own simple > dictionary sort of records by name. So 'nano', 'mc' and #brain' are obsolete! :-/ > 3. Notify is your friend, and so is BIND 9. I use Zoneedit service for > the parent domain, and I tried setting them up as slave of the dynamic > subzone. But they don't work with notifies, and they only do AXFR. I > deleted that out of zoneedit and switched to a network of BIND 9 slaves. > The slaves get their notifies and start their IXFR immediately. Changes > are propagated in 2 seconds (when network lag is bad) or less. OK. > Wouldn't you also use a variable in place of the RR name? I do. Will Yes, I have around 30 Workstations, Laptops and servers to update > there only be one user? I've offered mine to friends, so I've got a > handful of users. I started this because a paying customer was on a > dynamic IP (PPPoE), and we needed to be able to find that host. Me too. > I use TXT records in mine. I generate a UTC string of update and store > as a TXT record. I think that's a useful feature, because it tells me > when a user last updated his/her record. I do logging too. And I recommend it to all peoples which do such scripting. Recording the source IP with timestamp is essential. Maybe you/me will get hack attempts > 86400 is a bit long for a TTL, if this is truly a dynamic IP. My own is Yes its 100% dynamic IP. > dynamic in form but not in practice, so I can use longer TTL's, but even > so I choose to keep it no more than 3600. I use 900 for a friend on > dialup, 1800 for IP's which change occasionally, and 3600 for mostly I think, I will use this values too. > static IP's. (TTL value is set in the user record which the CGI script > uses for authentication, and I use a variable for that, too.) OK > > This is easier as I was thinking... > > Indeed. Have fun with it. :-) Now I am working since 6 years and 4 month with Linux (since 2.0.36) and each day I have more fun. :-) Greetings Michelle -- Linux-User #280138 with the Linux Counter, http://counter.li.org/ Michelle Konzack Apt. 917 ICQ #328449886 50, rue de Soultz MSM LinuxMichi 0033/3/88452356 67100 Strasbourg/France IRC #Debian (irc.icq.com) -- Attached file included as plaintext by Ecartis -- -- File: signature.pgp -- Desc: Digital signature -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.2.5 (GNU/Linux) iD8DBQFC1pyLC0FPBMSS+BIRAn98AKCTK8oJyCDbL2FzEaJ6jhOCLHxBaQCgpb1n oGz014em6MKezQlv+pSAktI= =jzNx -----END PGP SIGNATURE-----