<blockquote style="margin: 0pt 0pt 0pt 0.8ex; border-left: 1px solid rgb(204, 204, 204); padding-left: 1ex;" class="gmail_quote">Basically, there are cases where kickstarts go interactive when ksdevice<br></blockquote><div>
<blockquote class="gmail_quote" style="margin: 0pt 0pt 0pt 0.8ex; border-left: 1px solid rgb(204, 204, 204); padding-left: 1ex;"> is set at the installer kernel command line. I've seen it both with<br> ksdevice=link and ksdevice=eth1, and it winds up throwing an error about<br>
how it can't retrieve an RPM. Manually retrying works, and the<br> installation continues normally.<br> <br> Is that similar to what you were seeing?</blockquote><div><br>not quite. allow me to backstep.<br><br>fetching a kickstart file from the local disk as in ks=hd:sda1:/foo.ks works in the rhel5 installer but not in rhel6b2. i'd sure like that ability back. probably ought to file a bug on that.<br>
<br>next i tried nfs. i set ksdevice=eth1, that didn't work, but if i set ksdevice=<mac address> it works, and becomes clear that rhel6 anaconda is naming that device eth3, instead of eth1 as it was named by the rhel5 installer. so then going forward with ksdevice=<mac address>, anaconda barfed out twice on me, <a name="b622558" href="https://bugzilla.redhat.com/show_bug.cgi?id=622558">the first</a> could well have been that kickstart is also having some problem with the ethernet device names, so i went on without kickstart.<br>
<br><a href="https://bugzilla.redhat.com/attachment.cgi?id=437684">the second</a> also probably ought to be filed as a new bug, it seems anaconda tripped over a bad raid superblock (that came from trying to shrink a raid lvm physical volume). after i deleted the awry partitions it got past that.<br>
<br>having leaped all those hurdles, i have now installed rhel6b2.<br><br>thank you kindly,<br>greg<br></div></div>