I've received two responses both suggesting I investigate using a custom
version of the resolver+ software. I've done some work playing with the
code but have not been able to affect the timeout going from the
primary nameserver to a secondary server. I hope to put some more time
on it when I get a chance.
>We run DNS at our site and have a primary and secondary name server defined
>in our /etc/resolv.conf files. This sounds good until one of our suns is
>unable to get to the primary because the machine is down or more likely
>there is a network problem. It seems to take about a minute and a half to
>timeout on the primary server before going to the next server in the
>resolv.conf file. Is this timeout value something that is configurable in
>the kernel or elsewhere???
>We have sparc2s and sparc10s running 4.1.x and no NIS.
>Any help would be appreciated. I will summarize.
One way to change this, is to compile a modified version of the
resolver routines, into your shared C library.
The package "resolv+" purportedly handles this. There are a couple of
known problems with resolv+, noteably it can confuse sendmail... and I
believe there was one other thing, too. They were both documented in
the resolv+ .tar.Z file.
A summary could be helpful...
Dan Stromberg - OAC/DCS email@example.com
Thanks to those who responded
Greg Barry Boeing Computer Services Richland
firstname.lastname@example.org (509) 376-1652
This archive was generated by hypermail 2.1.2 : Fri Sep 28 2001 - 23:08:01 CDT