I now have debian installed on my other partition and have managed to get X working under Gnome but not KDE. Unfortunately the installer ignored my rh7.3 on the other partition and overwrote the MBR - good job I had made a boot disk. Anyway, tried to mod the redhat lilo.conf to boot debian too but I can't get it to work. boot is /dev/hda in both cases. rh root is /dev/hda1 and debian is /dev/hda2. I can only have one root statement in the preamble block and lilo does not like path names starting /dev in the image line. So how to do this? Ian
On Sun, 27 Apr 2003 11:27:56 -0000 MJ Ray <markj@cloaked.freeserve.co.uk> wrote:
Alternatively, try grub. Beware, as the path specs are a little different.
Grub will definitely enable you to do what you want, though I still think it is possible with LILO too. With LILO most of the inteligence is in the map installer, the 'lilo' command you run to set things up, and the code that actually does the booting is therefore very simplistic, but small and light and loads very fast. By comparison, grub has a second stage loader that is able to read proper file systems rather than just follow a block list so it is very flexible and can cope with things getting changed, but does take noticably longer to load because the second stage loader is bigger. Though I still have LILO for my main boot loader for the extra speed, I have a copy of grub on a floppy as a completely general purpose boot disk - just the job for booting a configuration that is otherwise broken. Steve.
On Sunday 27 Apr 2003 12:27 pm, MJ Ray wrote:
Ian Bell <ian@redtommo.com> wrote:
how to do this?
Not sure. Can you put root statements in the option blocks?
Yes, in fact you have to because root is on a different partitionfor each distro. The solution I have come up with is to temporarily mount the debian file system under the RH one (at /mnt/deb) and copy debian's vmlinuz into RH /boot. lilo then accepts this and boots debian OK. Now that one is solved it is time to find out why KDE does not work. X windows gives a fatal server error 4. What does that indicate? By the way I am using debian 3.0 from the DVD I got last year at Linux Expo. I know 3.01 is out now - did this make any changes to KDE? Ian
On Sun, 27 Apr 2003 18:30:03 +0100 Ian Bell <ian@redtommo.com> wrote:
Now that one is solved it is time to find out why KDE does not work. X windows gives a fatal server error 4. What does that indicate?
When does this error occur? Is it before login, i.e. is it when you are expecting a graphical login screen like kdm or gdm to appear, or is it after you have logged in? X writes a log file which may help diagnose this - where this gets written may depend on which login manager you are using. For gdm2 I have the log file /var/log/XFree86.0.log. I have also seen a log called gdm:0.log or some such similar thing. Steve.
On Monday 28 Apr 2003 10:36 pm, Steve Fosdick wrote:
On Sun, 27 Apr 2003 18:30:03 +0100
Ian Bell <ian@redtommo.com> wrote:
Now that one is solved it is time to find out why KDE does not work. X windows gives a fatal server error 4. What does that indicate?
When does this error occur? Is it before login, i.e. is it when you are expecting a graphical login screen like kdm or gdm to appear, or is it after you have logged in?
X writes a log file which may help diagnose this - where this gets written may depend on which login manager you are using. For gdm2 I have the log file /var/log/XFree86.0.log. I have also seen a log called gdm:0.log or some such similar thing.
Steve.
Strangely XFree86.0.log and kdm.log do not show the error. k login manager comes up OK and I can login as any user for a gnome session. If I login for a kde session it gets to the intialsing peripheralspart, briefly flashes a console with the fatal server error 4 then restarts klogin. Ian
On Wednesday 30 Apr 2003 9:29 am, MJ Ray wrote:
Ian Bell <ian@redtommo.com> wrote:
a kde session it gets to the intialsing peripheralspart, briefly flashes = a=20 console with the fatal server error 4 then restarts klogin.
There might be a $HOME/.xsession-errors file for that user.
Good clue. When I run kde, errors apear in here, principaly from kdeinit whcih sees an IO error and kiils its child (presumable X). It is intersting that when I run gnome, which seems to work, a whole bunch of erros appears in here (which I can look at from within gnome) many of which are gtk warnings but there is a gtk critical, though these do not seem to stop gnome from functioning. Ian
On Tuesday 29 Apr 2003 11:24 pm, Ian Bell wrote:
On Monday 28 Apr 2003 10:36 pm, Steve Fosdick wrote:
On Sun, 27 Apr 2003 18:30:03 +0100
Ian Bell <ian@redtommo.com> wrote:
Now that one is solved it is time to find out why KDE does not work. X windows gives a fatal server error 4. What does that indicate?
When does this error occur? Is it before login, i.e. is it when you are expecting a graphical login screen like kdm or gdm to appear, or is it after you have logged in?
X writes a log file which may help diagnose this - where this gets written may depend on which login manager you are using. For gdm2 I have the log file /var/log/XFree86.0.log. I have also seen a log called gdm:0.log or some such similar thing.
Steve.
Strangely XFree86.0.log and kdm.log do not show the error. k login manager comes up OK and I can login as any user for a gnome session. If I login for a kde session it gets to the intialsing peripheralspart, briefly flashes a console with the fatal server error 4 then restarts klogin.
Ian
I tell a fib, it is signal 4 not server error 4. Ian
On Sat, 26 Apr 2003 12:26:28 +0100 Ian Bell <ian@redtommo.com> wrote:
I now have debian installed on my other partition and have managed to get X working under Gnome but not KDE. Unfortunately the installer ignored my rh7.3 on the other partition and overwrote the MBR - good job I had made a boot disk. Anyway, tried to mod the redhat lilo.conf to boot debian too but I can't get it to work. boot is /dev/hda in both cases. rh root is /dev/hda1 and debian is /dev/hda2. I can only have one root statement in the preamble block and lilo does not like path names starting /dev in the image line. So how to do this?
Have you tried mounting the other disk partition? So for example, from your Redhat installation: mkdir /mnt/debian mount /dev/hda2 /mnt/debian then from within lilo.conf you could try naming you images like: image=/vllinuz label=Redhat image=/mnt/debian/vmlinuz label=Debian I have never tried this, but from what I understand, the 'lilo' command (map installer) asks the kernel for the block maps of the named kernels when you run it, so the actual boot loader doesn't care about files or partitions it just load the blocks it has been told to, so device and file names refer to the state of play when the map installer is run, not when the system boots. As far as the roots for the different linuxes are concerned, the kernels can store their own record of what their root filesystem should be, using the rdev command. Steve.
On Sun, 27 Apr 2003 17:43:34 +0100 I wrote:
As far as the roots for the different linuxes are concerned, the kernels can store their own record of what their root filesystem should be, using the rdev command.
I just had another thought about this and checked the man entry for lilo.conf. The root= setting can be global, or you can specify a different one in each image block. Steve.
On 2003.04.26 12:26 Ian Bell wrote:
I now have debian installed on my other partition and have managed to get X working under Gnome but not KDE. Unfortunately the installer ignored my rh7.3 on the other partition and overwrote the MBR - good job I had made a boot disk. Anyway, tried to mod the redhat lilo.conf to boot debian too but I can't get it to work. boot is /dev/hda in both cases. rh root is /dev/hda1 and debian is /dev/hda2. I can only have one root statement in the preamble block and lilo does not like path names starting /dev in the image line. So how to do this?
Ian
Mine (Debian Testing) is set up so: boot=/dev/sda delay=40 lba32 compact vga=normal map=/boot/map image=/boot/vmlinuz-2.4.20 root=/dev/sdb1 label=linux-pwroff append="apm=power-off" read-only image=/boot/vmlinuz-2.4.19 root=/dev/sdb1 label=linux-2419 append="apm=power-off" read-only As I understand it the 'boot' statement specifies the MBA and the two 'root' statements specify the partition to boot. Is that what you wanted to know or have I missed something obvious here? Barry Samuels http://www.beenthere-donethat.org.uk The Unofficial Guide to Great Britain
get it to work. boot is /dev/hda in both cases. rh root is /dev/hda1 and debian is /dev/hda2. I can only have one root statement in the preamble block and lilo does not like path names starting /dev in the image line. So how to do this?
AFAIR this is how I did it. Hope the explanation is clear. Make sure you have a boot disk for both distros if they are already installed. Basically, rather than trying to do the whole thing from within one instance of lilo on the MBR, each distro is installed (or lilo run) to place the boot loader on the appropriate boot partition for that distro. Only the last distro to be installed (or to be more exact, the final instance of lilo that is run) has lilo installed onto the mbr and it has lilo.conf set up to redirect to each of the other secondary boot loaders. This is essentially what happens in a dual-boot with windows for example and the lilo stanza is very similar. This what worked for me. In distroA, edit lilo.conf so that it is installed onto the boot partition of distroA, not onto the mbr of hda e.g. I have boot=/dev/hda2 as first line . Also the stanza for distroA contains a line pointing to the root partition e.g. root=/dev/hda5 Run lilo of course. Now boot into distroB and edit lilo.conf so that there is an entry that points towards the 'secondary boot loader' which was installed in distroA using 'other' e.g. other=/dev/hda2 label=distroA The stanza for distroB - in distroB lilo.conf - contains the line root=/dev/hdc3 since that is where root partition for distroB resides. First line of lilo.conf is boot=/dev/hda i.e installed onto MBR. (And run lilo). Finally make sure that /dev/hda is marked as the bootable partition with fdisk if this has been changed - if not then you should be able to ignore that. OK, so the boot process starts with lilo on the MBR. Selecting DistroB takes it to the boot and root partitions as usual but selecting distroA chains to the secondary boot loader which then sets up the boot and root partitions. This method is elegant and easy to manage since each distro has its own lilo.conf . It also avoids complicated lilo entries, copying kernels from other distros into one boot partition etc. HTH - let me know if you try it and have any problems. I'm working from notes and shaky memory but I'm pretty sure I haven't missed out anything essential. If you don't have a /boot partition then you'll need to experiment - my guess is to point at the root partition in that case. Regards Syd
participants (6)
-
Barry Samuels -
Ian Bell -
MJ Ray -
Steve Fosdick -
Steve Fosdick -
Syd Hancock