Hey Chris,
That’s really very strange. It sounds like it should have just worked…fresh OS install should never have any problems during install.
Reinstalling Postfix and Dovecot won’t configure them for Virtualmin (we do that stuff during the Virtualmin installation).
You’ve got two easy (easier than fixing everything manually, because even I can’t keep up with everything that goes on during installation, in both virtualmin-base and in the install.sh script itself–the install process is maintained by both me and Jamie, and neither of us knows everything it does without looking closely at the different pieces of the code) options:
You can attempt a full re-install of Virtualmin, by using the virtualmin-uninstall.sh script first (you don’t need to re-image the whole system to do this–the script removes everything that it needs to start fresh, or at least does as far as I know). It’s ugly and not recommended if you already have any domains (because it will kill them without backing them up), but it will probably work. Then again, the same thing that happened durind your first install might happen again and we’ll still have to figure out what failed–we might at least note an error hiding in the output somewhere (error logging in the install script isn’t yet ideal…I’ve recently added a nice shell logging library and it’s getting better with each release, but I’m having to capture errors from so many different sources, and some things don’t throw errors even if something doesn’t go right, that it’s difficult to catch everything). The virtualmin-uninstall.sh script can be found here:
http://software.virtualmin.com/lib/virtualmin-uninstall.sh
Be sure to also download the latest install.sh before attempting this…maybe something has been fixed since the first install. When you go to http://www.virtualmin.com/serial the Download link is always the latest version of install.sh available.
Or, and this one might be simpler, you can just reinstall Postfix and Dovecot (to get fresh configuration files) and then run the standalone perl script that does the configuration steps (this is pretty much pulled verbatim out of the virtualmin-base package). This script is named virtualmin-base-standalone.pl and can be found in this directory (you’ll need to use your serial number and license key as the username and password to login–these can be found in /etc/virtualmin-license, or in the serial numbers page on this website):
http://software.virtualmin.com/wbm
Or, you can email me the details for logging in, and I’ll fix it for you personally. I’d kind like to see where it blew up, as I suspect there’s more than just this trouble–mail is only halfway through the configuration process, and if it didn’t get done, nothing after it will have gotten done. And that’ll lead to surprises later. Either of the methods of resolving the issue given here will probably fix it–but I can’t be sure of that. The same thing might happen again (but maybe we’ll see the error this time and can address it in the install script for another go).
Sorry you’ve run into this trouble. I think Fedora Core 5 installs are still a bit iffy, though I didn’t have trouble during my test installs. I think we just need a bit more real world work with it to get it solid. But, for all systems the logging needs to get a lot better–it shouldn’t be nearly invisible when something goes so wrong with the install that major features aren’t working. 