Sequence number obsession

[Gerald (Jerry) Carter]

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

On Sun, 27 Apr 2003, Ken Cross wrote:

> Samba-folk:
> 
> We seem to be a bit obsessive-compulsive when it comes to getting a
> domain's sequence number.  I have network traces where we obtain the
> sequence number >20 times within a second.
> 
> The "force" argument to refresh_sequence_number in winbindd_cache.c
> appears to be true quite often.  Do we really need sub-second updates of
> the sequence number?
> 
> The patch below to SAMBA_3_0 changes it so that it doesn't update the
> sequence number if it is valid and has been updated within the last
> second.  Note that this overrides the "force" argument.

Ken,

I think the original idea behind forcing an update of the sequence number 
was to keep the cache straight.  The sequence number opdate is forced to 
True whenever our local cache fails and we have to query the PDC.  If we 
don't update the sequence number then we can get a false positive that 
says a cache entry is out of date.  This is a rough guess since I have 
just brielfy looked at the code recently.




cheers, jerry
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.2.0 (GNU/Linux)
Comment: For info see http://quantumlab.net/pine_privacy_guard/

iD8DBQE+y4QXIR7qMdg1EfYRAlIbAJ0XA4dtrpvJLysNRGiAwk5kOcIcDwCgrjlQ
LzXKc1W0KOVAZHT3PxmbUak=
=gwNx
-----END PGP SIGNATURE-----