OK, there is something somewhat unrelated to all that other stuff that I
would like to run by the developers.
Basically, I'm writing a kickstart install script for redhat. I put
all of the failsafe rpms in the file it that I'm interested in and
they get installed just fine (after running genhdlist). My dilemma is
this; you get a section at the end to run a bunch of shell commands
after the installation completes, but you don't have access to the
installation media anymore, because it's running in a 'chroot'
environment.
I'd like to run the commands:
$FAILSAFEINSTALLDIR/fsinstall
$FAILSAFEINSTALLDIR/guiinstall server
$FAILSAFEINSTALLDIR/guiinstall client (remember -I'm exporting my display)
Here are the problems:
1) I've installed the rpms, but AFAIK, I don't have the fsinstall or
guiinstall scripts anywhere on the filesystem. I could make a
little RPM that contained the install scripts, but I was kind of
hoping that you guys could just add the scripts to the RPMs so
they would be there
2) The fsinstall script prompts the user. This isn't good for my
kickstart script. I'd like to pass it a command line flag to
tell it on install to create the config file. Similarly, on
upgrade installation, I'd like to pass it a flag to just leave
the database alone.
Since I don't have the source at my
disposal yet, Is there some kind of flag we can pass to
/usr/sbin/cdbreinit so it won't ask the question? I see that if
I pass it an argument --help and it munched away for a little
while and came back empty handed (YIKES! what did I do?)
FYI, I thought of a way to work around this, but I like the method I
proposed above better. The workaround would be:
pass a --chroot flag to the %post section,
copy the scripts from the distribution media
Do a manual 'chroot' command and
run the rest of my customized stuff
run the failsafe scripts:
echo y | $FAILSAFEINSTALLDIR/fsinstall
In the end, I'd like to also define the cluster, resources, and
resource groups from a script using cluster_mgr. (Does anyone have
an example?)
-Eric.
|