Exceptional handling of glue credibility - why?
Ladislav Vobr
lvobr at ies.etisalat.ae
Sun Aug 1 03:59:54 UTC 2004
I have posted several posts about the above, but still I feel there is
no consistency in it. Of course most possibly I am wrong, and I am
bothering you still with the same questions and I am sorry if it is so,
but I am really trying my best to understand this behavior to be able to
troubleshoot daily customer cases.
Why recent binds doesn't provide the glue credibility cache records out
of it's cache to the recursive clients? Why does it put such an
incredible headache to get an authoritative answer for the recursive
client at the moment the client asks, while before when it was fetching
the data through some other requests it really doesn't care which
credibility it is, just simply cache it.
This puts a big load on the server, and creates unexplainable
situations, when although the data are in the cache bind gets very busy
doing something nobody really needs and wants and expect.
saying that as long as the data are cached there is really no need to
fetch them again and bind should be able to answer from the cache is
really untrue and the real life is quite different.
saying that if there is a local failure, but the data is cached, it
should not affect the bind performance is untrue as well, since bind
retries incredibly even for the cached names be it glue credibility.
can any body shed some light on this and the logic behind it?
One more question, why some binds treats exactly the same data different
way -(delegation ns and a records are sometimes in the answer section,
sometimes in the additional section, sometimes cached as glue, sometimes
as an answer, and some times the glue credibility is not provided to the
clients but answer credibility always is.... simple isn't it?)
Ladislav
More information about the bind-users
mailing list