Trying to upgrade to Linux kernel 2.4.13

Trying to upgrade to Linux kernel 2.4.13

Post by Scott Danzi » Tue, 30 Oct 2001 15:58:03



Hi,

I haven't compiled a kernel manually before on my own.  I got the
standard Red Hat 7.2 distribution, and I'm trying to upgrade it to
2.4.13 so my Promise Fasttrak100 RAID controller will work.  I end up
with a kernel panic error message from my latest attempt, saying that it
couldn't detect /sysroot, /sysroot/initrd.  I'm thinking that was
because maybe /dev/hda2 isn't the correct boot partition once the RAID
drives get recognized, but I'm not even totally sure I set things up
right.  Can someone see if what I'm doing is wrong?  Here are the steps
I took:

I first extracted the 2.4.13 kernel source, and followed the README
instructions closely, running man xconfig, making sure that RAM disk
functionality was turned on.  I then installed the modules.  I noticed
in /boot there wasn't an initrd-2.4.13.img in there, so I first tried
following the initrd.txt that came with the kernel doc.  It said to copy
"all the files that are needed to properly use the initrd environment"
.. no clue what those were...so I looked online.  One post said to just
go into /boot and type   mkinitrd initrd-2.4.13.img 2.4.13  and that'd
work.  It looked nice and simple and it produced a reasonably-sized 265k
file.  Since I'm running grub, I went into /boot/grub/grub.conf and
followed the format of the 2.4.7-10 kernel that Red Hat set up.  Well, I
tried selecting the 2.4.13 kernel after that was done and it's not
working.  I notice there's no /boot/modules-info-2.4.13 nor a
/boot/kernel.h-2.4.13 and I'm hoping that's okay.  Should I just start
guessing /dev/hd*'s or is there something I didn't do right?

Thanks for any help you can offer.

Scott Danzig

 
 
 

Trying to upgrade to Linux kernel 2.4.13

Post by Scott Danzi » Tue, 30 Oct 2001 17:11:00


Just as a latest update on this, I installed the ext3 kernel patch and
left /dev/hda2 as the root partition.  Now it brings me to Kudzu,
offering for me to detect new hardware..the keyboard doesn't work..
eventually it brings me to a login prompt with a non-functional
keyboard.  Here's a log of it:

Oct 29 02:48:46 localhost sysctl: net.ipv4.ip_forward = 0
Oct 29 02:48:46 localhost sysctl: net.ipv4.conf.default.rp_filter = 1
Oct 29 02:48:46 localhost rc.sysinit: Configuring kernel parameters:
succeeded
Oct 29 02:48:46 localhost date: Mon Oct 29 02:48:40 EST 2001
Oct 29 02:48:46 localhost rc.sysinit: Setting clock  (localtime): Mon
Oct 29 02:48:40 EST 2001 succeeded
Oct 29 02:48:46 localhost rc.sysinit: Loading default keymap succeeded
Oct 29 02:48:46 localhost rc.sysinit: Setting default font (lat0-sun16):
  succeeded
Oct 29 02:48:46 localhost rc.sysinit: Activating swap partitions:
succeeded
Oct 29 02:48:46 localhost rc.sysinit: Setting hostname
localhost.localdomain:  succeeded
Oct 29 02:48:46 localhost fsck: /: clean, 117274/2375680 files,
674458/4745199 blocks
Oct 29 02:48:46 localhost rc.sysinit: Checking root filesystem succeeded
Oct 29 02:48:46 localhost rc.sysinit: Remounting root filesystem in
read-write mode:  succeeded
Oct 29 02:48:46 localhost rc.sysinit: Finding module dependencies:
succeeded
Oct 29 02:48:46 localhost fsck: /boot: clean, 47/12048 files,
12458/48163 blocks
Oct 29 02:48:46 localhost rc.sysinit: Checking filesystems succeeded
Oct 29 02:48:46 localhost mount: mount: fs type devpts not supported by
kernel
Oct 29 02:48:46 localhost rc.sysinit: Mounting local filesystems:  failed
Oct 29 02:48:46 localhost rc.sysinit: Enabling local filesystem quotas:
  succeeded
Oct 29 02:48:47 localhost rc.sysinit: Enabling swap space:  succeeded
Oct 29 02:48:49 localhost kudzu: Updating /etc/fstab succeeded
Oct 29 02:49:30 localhost kudzu:  failed
Oct 29 02:49:30 localhost kudzu: Hardware configuration timed out.
Oct 29 02:49:30 localhost kudzu: Run '/usr/sbin/kudzu' from the command
line to re-detect.
Oct 29 02:49:31 localhost sysctl: net.ipv4.ip_forward = 0
Oct 29 02:49:31 localhost sysctl: net.ipv4.conf.default.rp_filter = 1
Oct 29 02:49:31 localhost network: Setting network parameters:  succeeded
Oct 29 02:49:31 localhost ifup: Cannot send dump request: Connection
refused
Oct 29 02:49:31 localhost network: Bringing up interface lo:  succeeded
Oct 29 02:49:36 localhost xinetd: xinetd startup succeeded
Oct 29 02:49:37 localhost lpd: lpd startup succeeded
Oct 29 02:49:38 localhost sendmail: sendmail startup succeeded
Oct 29 02:49:38 localhost gpm: gpm startup succeeded
Oct 29 02:49:39 localhost crond: crond startup succeeded
Oct 29 02:49:40 localhost xfs: xfs startup succeeded
Oct 29 02:49:40 localhost anacron: anacron startup succeeded
Oct 29 02:49:40 localhost atd: atd startup succeeded
Oct 29 02:49:40 localhost rhnsd: rhnsd startup succeeded
Oct 29 02:49:40 localhost wine: /etc/rc5.d/S99wine:
/proc/sys/fs/binfmt_misc/register: No such file or directory
Oct 29 02:49:40 localhost wine: /etc/rc5.d/S99wine:
/proc/sys/fs/binfmt_misc/register: No such file or directory
Oct 29 02:49:40 localhost rc: Starting wine:  failed

Thanks again :)

Scott Danzig


> Hi,

> I haven't compiled a kernel manually before on my own.  I got the
> standard Red Hat 7.2 distribution, and I'm trying to upgrade it to
> 2.4.13 so my Promise Fasttrak100 RAID controller will work.  I end up
> with a kernel panic error message from my latest attempt, saying that it
> couldn't detect /sysroot, /sysroot/initrd.  I'm thinking that was
> because maybe /dev/hda2 isn't the correct boot partition once the RAID
> drives get recognized, but I'm not even totally sure I set things up
> right.  Can someone see if what I'm doing is wrong?  Here are the steps
> I took:

> I first extracted the 2.4.13 kernel source, and followed the README
> instructions closely, running man xconfig, making sure that RAM disk
> functionality was turned on.  I then installed the modules.  I noticed
> in /boot there wasn't an initrd-2.4.13.img in there, so I first tried
> following the initrd.txt that came with the kernel doc.  It said to copy
> "all the files that are needed to properly use the initrd environment"
> .. no clue what those were...so I looked online.  One post said to just
> go into /boot and type   mkinitrd initrd-2.4.13.img 2.4.13  and that'd
> work.  It looked nice and simple and it produced a reasonably-sized 265k
> file.  Since I'm running grub, I went into /boot/grub/grub.conf and
> followed the format of the 2.4.7-10 kernel that Red Hat set up.  Well, I
> tried selecting the 2.4.13 kernel after that was done and it's not
> working.  I notice there's no /boot/modules-info-2.4.13 nor a
> /boot/kernel.h-2.4.13 and I'm hoping that's okay.  Should I just start
> guessing /dev/hd*'s or is there something I didn't do right?

> Thanks for any help you can offer.

> Scott Danzig



 
 
 

Trying to upgrade to Linux kernel 2.4.13

Post by Steven Sanfratell » Thu, 01 Nov 2001 20:18:16


it sounds like a mess.  you might have more than one problem.
i'm only guessing, be careful when doing any of my suggestions.
make sure you have a way of undoing anything that you do.

Quote:> > file.  Since I'm running grub, I went into /boot/grub/grub.conf and
> > followed the format of the 2.4.7-10 kernel that Red Hat set up.  Well, I

if you edited grub's menu (grub.conf is the menu right?  my
grub uses a different name ... menu.lst) then you MUST have
entered something for the filename of the kernel image.  did
you just enter the previous kernel?  maybe you entered a kernel
which doesn't exist and grub is defaulting to the previous
kernel.  whatever name you entered into grub's menu, then THAT
is the kernel which you are booting.

after compiling a kernel it tends to be called zImage or
one of a few other names.  it's located under the source
tree of the kernel.  check the documentation that packaged
with the kernel to find where the file is.  this can be
copied to the /boot dir and renamed to whatever.

Quote:> > .. no clue what those were...so I looked online.  One post said to just
> > go into /boot and type   mkinitrd initrd-2.4.13.img 2.4.13  and that'd
> > work.  It looked nice and simple and it produced a reasonably-sized 265k
> > file.  

i don't know what that command does.  maybe it made some init
scripts that are meant to be used with your new kernel.  but if
grub is actually booting the previous kernel, then you might have
an incompatability between the new scripts that this command created
and the previous kernel.  using the original selection off of grub's
menu, can you still boot the old kernel?  if so, then there's probably
no incomaptability issues.

Quote:> Oct 29 02:48:46 localhost mount: mount: fs type devpts not supported by
> kernel
> Oct 29 02:48:46 localhost rc.sysinit: Mounting local filesystems:  failed
> Oct 29 02:49:40 localhost wine: /etc/rc5.d/S99wine:
> /proc/sys/fs/binfmt_misc/register: No such file or directory
> Oct 29 02:49:40 localhost wine: /etc/rc5.d/S99wine:
> /proc/sys/fs/binfmt_misc/register: No such file or directory
> Oct 29 02:49:40 localhost rc: Starting wine:  failed

i'm not sure what devpts is, but i think it's fundamentally
important.  it might allow access to everything under the
/dev directory (harddrives, floppies, serial ports, most hardware).
you might want to make sure that the new kernel is compiled to
support devpts mount type.

well if wine is causing the failure, you might try (temporarilly)
moving the /etc/rc5.d/S99wine link out of that directory in order
to remove it from start up.  if you have webmin or some higher
level configuration program, use that program to disable wine.

  -steve s.