Good afternoon, Sean...

On Sunday 18 September 2005 1:39 pm, Sean O'Connell wrote:

If you are willing to do kickstarts, then, yes. you can do this. Add a reinstall entry to your /etc/grub.conf using vnc (you connect to myserver:1)

title Reinstall System kernel /vmlinuz-reinstall ks=<URL>/server-reinstall.cfg vnc initrd /initrd-reinstall.img

where vmlinuz-reinstall and intrd-reinstall.img are the vmlinuz and initrd.img from the pxe subdirectory of os/<arch>/images/pxeboot that you need to download and drop in /boot (the above assumes /boot is its own partition otherwise you'll need /boot/vm..). There are a bunch of issues with doing this:

1. are you comfortable with kickstart?
2. are the machines using static addressing or dhcp? if static, then

you'll have to add a few command line parameters:

ip=w.x.y.z netmask=255.255.255.0 gateway=w.x.y.1 dns=w.x.y.2

3. you'll probably want to add a vncpassword=xxxxx flag as well

You'll probably want to play with a local machine first (and make sure someone is there to push a power button :).

Hope that helps,

It did work incredibly well to a point, despite my fumble-fingered approach to typing in the commands, as given. I *should* have listened to your advice of trying it locally before taking a shot at it remotely, though, because things froze because of a bad vmlinuz-reinstall location. However, I had someone at the remote site who read me the errors, and I fixed the problem.

The situation that originated this query took place recently when I tried to do a CD-based upgrade from 3.3 to 4.1, and although the install medium did search for previous instances of CentOS, it never found one (it was there, really!) and continued to try reinstalling the entire OS. I finally resorted to telling the installation to only format the /boot partition and *not* format the / partition, which seems to have worked, as it overwrote only those applications that had been upgraded/changed and saved all the configuration files.

However, I think once I get the knack of performing remote upgrades, I'll be eventually doing more of them.

Thanks for everyone's help with this. If anyone else has encountered the problem I met going from 3.3 -> 4.1, I'd like to hear about it, as well. The only reasons I am eager to upgrade to 4.1, as I alluded to earlier, is the more recent 4.1 kernel, expanded management utility applications, and x.org rather than XFree86.

I have not had *any* stability or security issues in 3.3 at all. Much to the contrary, it has been more stable than anything else I have tried in the recent past, including various other distributions too numerous to mention. I've even begun using a highly-modified workstation configuration at both my office and home, and am reasonably content with that, as well, with only a few minor exceptions. ;-)

Dave

Why, good evening, Bob!!!

On Sunday 18 September 2005 4:51 pm, Robert Hanson wrote:

Dave,

Although I have had decent to good success with 3.x to 4.x upgrades myself, I do not recommend it in most circumstances.... and recommend backup of configs and data and then fresh install and reconfig and restore data etc

Kind regards,

I just wanted to make certain it would even *work*, and it did. This winter I would really *like* to avoid long trips to Republic's mines, and if I can make this work remotely, it will save me some time spent sweating on those icy lonesome roads. FWIW, I was able to do two different kinds of upgrades this afternoon, one remotely (across the room on a different network) and one using 4.1 CD's and an existing 3.3 CentOS box.

However, it was good hearing from you. How is everything?

Dave