Hi, Let me put 2 cents here, as this fix hit some huge lab enviroment too.
Company have multiple sites, each of them have own domain name like "siteX.sub.domain". Each site has park of test servers, on whose you can find hell mix of all OS kinds, and they constantly provisioned by automatic installation system. All this zoo coexists happily with help of DHCP's option 15 "abuse", where provisioned domain set by DHCP to "siteX.sub.domain sub.domain domain other.sub.domain other.domain". Although it is deprecated, everyone is happy, the "search" line in /etc/resolv.conf is correctly populated, and most important - this works for *BSD, Solaris, old OS (RH/CentOS/OEL < 6) and newer ones, out from the box!!! So, with help of option 15, all servers can be reached by short names inside site and semi-short names crossite, eg: site-filer01 wil point to local filer01 and site-filer01.siteX wil point to siteX filer. Thus, it allowing users to use short names and admins to have unified kickstarts, post-install and repo files in "one place", without spreading thousands of them across sites and later on manage this nightmare. Until mentioned fix has been integrated, all Debian installations was provisioned without any problems, dozens a day. One day it stopped. Suddenly. No reasons. Now we are looking for a fix of this fix, instead of testing product compliance on this OS :( I've already reopened original udchpc bug at busybox bugtracker ( https://bugs.busybox.net/show_bug.cgi?id=3979), but it looks like I'm asking for a miracle... What can be done on this matter? Best regards, Daniel
