selecttest tool

[JINMEI Tatuya / 神明達哉]

At Fri, 08 Aug 2008 17:45:00 -0500,
Walter Gould <gouldwp at auburn.edu> wrote:

> > 1. you should specify a larger max-cache-size in named.conf.  the
> >    default size of 9.5.0 (32MB) is normally too conservative for a
> >    busy server with many clients, reducing cache hit rate and making
> >    the server busier, and possibly triggering subsequent performance
> >    troubles as a result.
> > 2. #1 may be sufficient for you, but you may also want to raise
> >    recursive-clients (whose default is 1000) to some reasonable large
> >    value (e.g., 10000).  I suspect the primary reason for the SERVFAIL
> >    in your case is that the server hits the recursive-clients quota.
> >    I suspect the server left warning messages like ""no more recursive
> >    clients: ...".  Raising recursive-clients will at least solve this
> >    issue.

> So, why would we have not had these same issues arise six months or a 
> year ago? Do you think it is related to upgrading to these newer BIND 
> versions or possibly just a timing coincidence?

I expected this question:-)  There are several possible explanations:

First, if you upgrade from 9.4 or 9.3 or even older version to
9.5.0-P1 or 9.5.0-P2, this may be due to the new default of
max-cache-size in 9.5.

Second, since the port randomization patch can generally make the
server busier (even with a reasonably large max-cache-size), it may
take longer time to handle each recursive client, which may then led
to hitting the recursive-client quota.

But these are just a guess at the moment, and I'm not confident about
those theories.  Recursive server's behavior is so complicated and
varies among users, and something beyond my imagination could easily
happen.

---
JINMEI, Tatuya
Internet Systems Consortium, Inc.