Sharpe, Sam J wrote: > > Our kickstart include files are available over http, and in the %pre > section, we grab them with wget. The advantage is that we can have a > specific overall kickstart file for a machine or group of machines, but > we can maintain the sections that are common to all machines separately. > Typically this might involve the root password or installation key being > set in the overall kickstart but the %packages section simply contains a > common set of packages which are set with an %include. > > Our kickstarts are a bit complicated because we first use wget to grab > some details of the machine from a central database and process that > information to make decisions about what to include. For example, a > kickstart might %include /tmp/packagelist.ks but which particular common > package list we wget and from what location is a decision made during > the %pre section based on what the type of the machine is. > > Your kickstart might simply contain: > > <snip> > %pre > wget -q -O- > http://installserver.you.domain/path/to/kickstarts/packagelist.ks > > /tmp/packagelist.ks > <snip> > %packages > %include /tmp/packagelist.ks > <snip> > > Now that doesn't seem like jumping through hoops to me...
Ok thanks for the responses, it makes more sense now, I was thinking that Anaconda would scan the kickstart file for %include and retrieve them from the same location (ie. applying the path relatively), maybe that would be a nice feature to suggest to Redhat. I think I'll look into the http/wget way, not a big fan of NFS, especially for things that are frequently mounted/unmounted. Cheers Tim Edwards _______________________________________________ rhelv5-list mailing list [email protected] https://www.redhat.com/mailman/listinfo/rhelv5-list
