-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1
Brian Millett wrote:
Ok, man pages..... Gee it was working as it should. There was not a
negative request as the old addr had been reassigned and was alive. Only
the name had been reassigned to a different ip.
That's as it should be. The entry in the database will time out after a
while and after reloading the new IP address is available. If one needs
the change to take effect immediately, remove the hosts cache with
nscd --invalidate=hosts
If this is a frequent thing, disable nscd caching completely. But there
is no bug, this is how nscd is supported to work.
So, should nscd be updated when a 'host','dig', or
'nslookup' request is
done?
All these programs don't use the libc resolver at all, they have nothing
whatsoever to do with nscd. By design, I might add.
I am playing with the /etc/nscd.conf to tune it for my particular
setup, but I was wondering what triggers a refresh, or update of the
cached database?
Just reduce the positive-time-to-live value for hosts if the current one
is too long for you. It is not in general.
Also, there is not an entry to the bugzilla for nscd.
nscd is part of glibc. No need to file a bug since there is none.
- --
➧ Ulrich Drepper ➧ Red Hat, Inc. ➧ 444 Castro St ➧ Mountain View, CA ❖
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.2.6 (GNU/Linux)
iD8DBQFBkmrL2ijCOnn/RHQRAh1TAJ9rRZjDC+Kjgzgc4wCQNTCZyndRKQCgoVAB
EpTAl+uah5DNIHmdAM80jWE=
=RGvM
-----END PGP SIGNATURE-----