Once upon a time, Miroslav Lichvar <mlichvar(a)redhat.com> said:
The trouble is that a fragment having a different name cannot
disable
servers specified in a different fragment. If anaconda wanted to
override the default servers, it would need to know the name of the
fragment.
So, that brings up something I've been meaning to file a bug about:
anaconda has behavior around NTP that nothing else does. If anaconda
gets NTP servers from DHCP, it deletes the pool entries from the default
config and replaces that with the DHCP-provided servers. This is wrong;
it assumes that the installation network is the same as the production
network, and overrides default config without any prompt or kickstart
config.
When an operational system gets NTP servers from DHCP (e.g. via
NetworkManager), it does not touch the config in /etc; it instead
notifies the NTP service to add the server.
IMHO anaconda is just wrong to do this, and it breaks using an
installation network. Right now, I have to hack around it in a
kickstart by telling anaconda to not configure NTP at all, but the
enable the service and add the package.
--
Chris Adams <linux(a)cmadams.net>