On 5/18/2013 9:24 PM, Kyle Brantley wrote:
On 5/16/2013 9:23 PM, Luke S. Crawford wrote:
On 05/16/2013 12:43 PM, Ruben Rodriguez <cha0s> wrote:
Thanks guys, but neither of these seems to be where the addresses
currently in /etc/resolv.conf are coming from. I don't believe I set
this up myself and have no idea how to change it. If everything breaks
whenever this change happens I'm probably not going to bother trying to
fix it, more like tar scp, rm :P
Yeah, most of your /etc/resolv.conf files were setup with the image
we use to install you (I want to get away from using images and make
you all use the install kernel/initrd for your preferred distro...
but that hasn't happened yet.)
This may be $DISTRO specific, but if you gave me a way of loading up
the installer kernel + initrd and then allowed me to specify a
kickstart URL to actually run the install, that would satisfy all of
my needs (and desires when it comes to installation, actually!).
Perhaps just maintain all of the common network installers and provide
the option to boot them up from the rescue console, but also provide
the option to edit the command line for those kernels as well? That
way you don't have to worry about an image for debian or fedora or
etc, all you need is a few hundred megs per distro and those who want
to perform custom / automated installs still can. Maybe for the people
who "just want a working distro" provide an option to use a kickstart
(or similar) file that would install the system using defaults without
end user intervention.
--Kyle
I'm working on setting up my second node now and trying to manually
install the OS (dropped vmlinuz and initrd.img into /boot, added an
entry for grub), and I've found a few problems with doing it this way.
1) CentOS 6 requires at least 512MB of RAM to install. This is a problem
for my 256MB VPS.
I've managed to mostly work around this by commenting out lines 1244
through 1257 as well as 2131 in loader/loader.c from the anaconda SRPM
and rebuilding the initrd to contain this new version of /sbin/loader.
It looks like there is another check for RAM once you actually get into
anaconda (starts on line 376 of the 'anaconda' script) and I'm not
positive on how I'll be able to work around that yet, but I also have
yet to get there. I'm sure that just adding 'return True' to line 377
would fix it, but this script isn't contained within the initrd, but
rather install.img (which is hosted in the same location as the rest of
the installer tree... I can't easily push a new version of this without
hosting the entire install mirror).
2) I can't actually seem to manually boot the installer. :(
No matter what I choose for the kernel boot options, even just as simple
as 'kernel /boot/vmlinuz' and nothing else, I can't seem to actually get
to the installer screen. The entire node seems to die during the kernel
init, and I think that this is truncating the actual error message. The
last few things on my screen:
NET: Registered protocol family 17
registered taskstats version 1
XENBUS: Device with no driver: device/vbd/51712
XENBUS: Device with no driver: device/vbd/51776
XENBUS: Device with no driver: device/vbd/51792
XENBUS: Device with no driver: device/vif/0
drivers/rtc/hctosys.c: unable to open rtc device (rtc0)
Initalizing network drop monitor service
Has anyone had success with manually installing CentOS here, or does
everyone wind up installing it elsewhere and transferring the image over?
--Kyle
_______________________________________________
discuss mailing list
[email protected]
http://lists.prgmr.com/mailman/listinfo/discuss
