Serial Number Issues

[Joseph S D Yao]

You give as your return address:

To: WndrGrl <wndrgrl.nospam at nospam.hotmail.nospam>

That should be enough to earn you no answer.  I notice nobody else has.
This is severely deprecated on this mailing list/newsgroup.

On Sun, Dec 17, 2000 at 12:53:52AM -0800, WndrGrl wrote:
> 
> In an upgrade from Bind 8.2.2p5 to Bind 9.0.1, I ran into a number of
> problems, the most pressing of which is serial numbers.
> 
> Bind 9.0.1 forced me to drop a digit from my serial numbers.
> 
> Previous format:  YYYYMMDDXXX
> New format: YYYYMMDDXX

My guess is that BIND 9.0.1 checks, because your previous format was
also illegal under BIND 8.*.

2^32 = 4294967296
       YYYYMMDDXXX

You were always wrong.  There has been no change.

> My slave name servers are not under my direct control; two are at
> granitecanyon and two are at Southwestern Bell.
> 
> Previous messages to the list/group suggest that the SN 0 trick no
> longer works properly in Bind 8.2x and higher, which of course isn't
> documented in the Grasshopper Book (I have both 2nd and 3rd editions
> on hand).  I can't honestly say I understand the wrap-around trick,
> and now that Bind 9 treats serial numbers as unsigned 32-bit integers,
> I also do not know that it'll work.

BIND 8 always did.

What do your slaves tell you is the current serial number?

> Short of having to kill the delegation on the remote name server and
> then set it back up, what is the solution to this issue?
> 
> Oh, one other wrinkle:  The remote name servers use Bind, but I can't
> figure out what version.
> 
> So, without control or even decent communication with those in
> control, how do I fix this problem?

I would suspect - no difference.  Use the wraparound technique.

-- 
Joe Yao				jsdy at cospo.osis.gov - Joseph S. D. Yao
COSPO/OSIS Computer Support					EMT-B
-----------------------------------------------------------------------
This message is not an official statement of COSPO policies.