On Tue, 13 Jan 2004, Martinez, Carlos R wrote:
>I'm having an issue with a Kickstart installation that gets stuck after it
>installs the kernel package. It just hangs indefinitely.
Sounds like the post-install script is hanging. Careful though; I might
be wrong about where the postinstalls run.
For example:
$ rpm -q kernel
kernel-2.4.20-20.9
$ rpm -q --scripts kernel
[snip]
postinstall scriptlet (using /bin/sh):
cd /boot
ln -sf vmlinuz-2.4.20-20.9 vmlinuz
ln -sf System.map-2.4.20-20.9 System.map
ln -sf module-info-2.4.20-20.9 module-info
[ -x /usr/sbin/module_upgrade ] && /usr/sbin/module_upgrade
[ -x /sbin/mkkerneldoth ] && /sbin/mkkerneldoth
if [ -x /sbin/new-kernel-pkg ] ; then
/sbin/new-kernel-pkg --mkinitrd --depmod --install 2.4.20-20.9
fi
[snip]
$ rpm -qf /bin/sh /boot /usr/sbin/module_upgrade /sbin/mkkerneldoth /sbin/new-kernel-pkg
bash-2.05b-20.1
filesystem-2.2.1-3
kudzu-0.99.99-1
initscripts-7.14-1
mkinitrd-3.4.42-1
If you didn't have bash, filesystem, kudzu, initscripts and mkinitrd
installed *before* kernel's postinstall starts, something might break.
Because most commands are guarded by "[ -x foo ] && foo", I wouldn't
expect a hang. What does ps/pstree show?
>To me it seems as if the package order gets out of whack. I know my
>Kickstart server is good as to I can successfully install a server that
>is on my local LAN. The server that I'm installing is over the WAN
>across country. I use HTTP as my installation method. The noticeable
>difference between the two servers is when the kernel package gets
>installed. On the server that has problems it is on package 164 and on
>the server that it installs successfully it is package 132. Both
>servers are identical and the only difference is location. Are there
>problems installing via HTTP over the WAN where the package list can
>get out of whack?
I'd be very surprised. Are you sure there are no HTTP proxies
(transparent or otherwise) getting in the way? Are you willing to tell
us the WAN URL, if it's publically accessible?
Cheers,
Phil
