I have been seeing the same thing here with Solaris 2.6 and internal
DNS.
Chris Tilbury wrote:
>
> On Thu, Jan 13, 2000 at 01:54:07AM +0100, Henrik Nordstrom wrote:
>
> > Duane Wessels wrote:
> >
> > > The number of dnsserver processes affects how many
> > > outstanding queries you can have at one time.
> > > If you didn't specify enough dnsserver child process
> > > you got warnings, or your users had to wait longer
> > > than necessary for one to become free.
> >
> > Not to mention that Squid aborts with a fatal error if the DNS backlog
> > queue to the external dnsserver processes grows a little large. This is
> > commonly seen when there are DNS server performance problems.
>
> Someone else has already mentioned this here, but the downside is that the
> internal dnsserver code doesn't seem to be 100% reliable yet (at least on
> Solaris, which I think is the platform the other report came from).
>
> Certainly here, on our fairly loaded site cache (50 requests per second,
> Solaris 7, with 60Gbyte Cache, 1Gbyte memory and two CPUs), we get fairly
> spurious (and regular) "server failures" reported by squid for resolving
> hostnames.
>
> We're still using the external dnsservers at the moment because of that.
>
> Chris
>
> --
> SQUID Frequently Asked Questions -> http://squid.nlanr.net/Squid/FAQ/FAQ.html
> -----------------------------------------------------------------------------
> Chris Tilbury, UNIX Systems Administrator, IT Services, University of Warwick
> PHONE: 024 7652 3365 / FAX: 024 7652 3267 / MAIL: Chris.Tilbury@warwick.ac.uk
This archive was generated by hypermail pre-2.1.9 : Tue Dec 09 2003 - 16:50:22 MST