Thanks Daniel,
I'm sorry it's taken so long to reply but I had to get away from
from it all for a while and fight fires on other fronts.
Previously, trying to boot into runlevel 3 I observed a timeout at
"INIT: Entering runlevel : 3
Starting system log daemon: syslog"
.. after what seemed a long time it reverted to runlevel 1. The same
with 2 and 5.
Daniel Bush wrote:
> Adam,
> There's not much help I can give. But I've added some comments below.
>
> On 11/29/05, *Adam Felix Bogacki* <[EMAIL PROTECTED]
> <mailto:[EMAIL PROTECTED]>> wrote:
>
> I found I had a kernel problem after trying to install
> dependencies for
> 'scribes' in succession when on reboot
> I could not get back into my 2.6.5 linux drive.An early error
> message was ..
>
> Error inserting genrtc
> (lib/modules.2.6.5-1-686/kernel/drivers/char/genrtc.ko): Device or
> resource busy.
> /etc/init.d/syslog start
>
> .. and later
>
> Kernel panic: VFS: Unable to mount root fs on unknown-block(0,0)
>
>
> Are you sure there is no way you could get 2.6.5 back up and running
> rather than trying to compile and boot a new kernel on an
> already-broken system based on 2.6.14 as you tried below?
Good question - and that was was the first area I explored.
Tux:~# mkinitrd /boot/initrd-2.6.5 -o 2.6.5custom
/usr/sbin/mkinitrd: /boot/initrd-2.6.5 is not a valid kernel version
I could not see 2.6.5 or 2.4.14 vmlinuz or linuz images on the system.
Symlinks remained in / or /boot but I could not find the actual files ..
unless I am missing something.
The only working kernels seem to be 2.6.14 and WinXP, on which I write this.
>
> ...
>
> After a bit of fiddling with /etc/lilo.conf I finally manged to boot
> into my own drive with the following messages
> which I reproduce manually in some detail in the hope that they mean
> something to someone ..
>
>
> ---------------------------------------------------------------------------------------------------------------------------------------
>
> Not running depmod because /lib/modules/2.6.14 is not writable
> Loading modules ..
> grep: error while loading shared libraries libpcre.so.3: cannot open
> shared object file: No such file or directory
> FATAL: Could not load /lib/modules/2.6.14/modules.dep: No such
> file or
> directory
> Creating device-mapper files
> [screen moves .........]
> /etc/rc5.d/S19autofs: line33: basename: command not found
> .......
>
>
> ...
>
>
> Debian GNU/Linux testing/unstable Tux tty1
>
> Tux login: adam
> Password:
>
> -----------------------------------------------------------------------------------------------------------------
>
> .. so I was in on tty1 at run-level 1
>
> I quickly found that only my Linux root and WinXP partitions were
> auto-mounted - both specified in /etc/lilo.conf
> - and I had to manually mount all other partitions except swap
> (which I
> could not quite work out how to do).
>
>
> swapon / swapoff
> You can use it on files too. (Also see mkswap - not that I think you
> need it here, though).
'swapon -a' works, but does not seem to make any difference.
>
> Do you have an /etc/rc.d/rc.sysinit file or similar?
Tux:/# find -name rc.sysinit -print
./opt/ltsp/i386/etc/rc.sysinit
I could not find an /etc/rc.d/rc.sysinit file .. but found many
./etc/rc*.d files, eg /etc/rc0.d, /etc/rcS.d. /etc/rc2.d .. etc.
Hmm .. maybe I could copy ./opt/ltsp/i386/etc/rc.sysinit to /etc/rc.d/
...
> You can see how the bootup shell script uses swapon and the 'mount'
> command to do stuff (like mounting root dir read-only before going
> read-write).
'mount -a' .. worked
>
> Finally,
>
> Tux: /home/adam/linux-2.6.14# make modules_install
> if [ -r System.map -a -x /sbin/depmod ] ; then /sbin/depmod -ae -F
> System.map 2.6.14; fi
>
> Tux: /lib/modules# ls
> 0.0.0 2.4.18-bf2.4 2.4.25-1-686 2.6.14 2.6.5-1-686
>
> Tux: /lib/modules/2.6.14# ls
> build modules.ccwmap modules.inputmap modules.seriomap source
> modules.dep modules.isapnpmap modules.symbols
> kernel modules.alias modules.ieee1394map modules.pcimap modules.usbmap
>
> .............
>
> As it now stands, I can boot into WinXP on which this is written and
> 2.6.14 at will
> but the latter is text mode (init 1), I have to mount other
> partitions
> manually,
> have no access to the internet preventing a dist-upgrade, and no
> X-server.
>
>
> If you mean runlevel 1, then is your networking script file
> ('network') in /etc/rc1.d (or similar) prefixed with a K or an S? On
> my system it is a 'K' - which means any networking will be killed (an
> 'S' means it will be started if not already on). Runlevel 1 on my
> system has a lot of K's and not a lot of S's. (The shell script file
> that manages changes in runlevels on my system is /etc/rc/rc.d.)
Looking at /etc/rc1.d all I seemed to see were K prefixes.
> You can start your rc* scripts manually using something like
> ./XXXnetwork start ( or 'restart')
Initially I had no success ..
Tux::~# /etc/init.d/network start
bash: /etc/init.d/network: Permission denied
. than I realised I was trying to as root to start the network in level 1 ..
Tux:/etc/init.d# ls
..included nethack (green)
network (white)
networking (green)
nfs-common (green)
nfs-kernel-server (green)
/etc/inittab conf file told me
#The default runlevel
id: 5 : initdefault:
#Boot-time system configuration/initialisation script
#This is first run except when booting in emergency (-b) mode
si::sysinit:/etc/init.d/rcS
#What to do in single-user mode
~~:S:wait:/sbin/sulogin
' mount -a ' worked like a charm, mounting the whole fstab !
' ifconfig -a ' told me that both RX and TX packets were 0
... upon further slugging away ..
Tux:~# /etc/init.d/sysklogd start
starting system log daemon: syslogd [5649.824461] process 'syslogd' is
using obsolete setsockopt SO_BSDCOMPAT
> (in the directory the file is in or just go to init.d). There's also
> often /sbin/service (in redhat at least)
>
> Also, telinit can be used to change runlevels. (see man page) Might
> be better to run that than starting up stuff in isolation. However,
> it looks like all your /etc/rc* scripts are failing to run which
> probably relates to your mount problem -- really, really not sure..
>
>
At that point I found I could change runlevels successfully via
/sbin/telinit 5
.. X and GDM came up normally but after logging in I found that GNOME
terminal would not give me a cursor with the message "There was an error
creating the child process for this terminal" .. I rebooted in sawfish
and other sessions and found no cursor
on other terminals (eg. Xterm).
There was no connectivity ..
"Failed to run network-admin as root.
Failed to fork new process: No such file or directory."
On reboot, it hung (timed out) again at
"INIT: Entering runlevel 5
Starting system log daemon: syslogd"
> /etc/fstab exists and I don't understand why it does not mount all
> partitions at start-up.
>
>
> This may relate to the sysinit file I was mentioning about above (or
> it may not).
> There is a cryptic note in my sysinit file saying that contrary to
> standard usage, file systems are NOT unmounted in single user mode.
> Anyway, this hints to the fact that changing runlevels may not only
> change the services but may possibly also change what is mounted.
>
> Only other things I can think of are:
> mount -a
> will try to mount everything in fstab.
Worked like magic !
> mount -a -t no<comma-list of file system types>
> will mount all available systems except those types that come after
> 'no'. You might see stuff like that in your sysinit file.
>
>
> Check /etc/mtab to see what was mounted at startup and any other
> invocations of 'mount' that you ran after that.
> Check /proc/mounts to see mounts...
>
> I have correctly AFAIK set up /etc/init.d/network but there is no
> internet connection.
>
>
> Dumb question: you did start it? (as per a previous comment above).
Unsuccessful - see above.
> Are your interfaces up? (ifconfig)
Negative. (See above).
>
> I feel I've achieved something but have not yet arrived at a
> working system.
>
> You really sure you can't resurrect your old system?
I could not find the vmlinuz and linuz images of the other kernels.
'nano /etc/rcS.d/README' gave me
The scripts in this directory whose names begin with 'S' are executed
once when booting the
system, even when booting directly into single user mode. The scripts
are all symbolic links,
the real files being located in /etc/init.d/.
The following sequence points are defined at this time.
* After the S40 scripts have executed, all local file systems are
mounted and networking is available.
All device drivers have been initialised.
.........................
Also,
Tux:/etc/init.d# /etc/init.d/networking start
* /etc/network/options is deprecated
Setting up IP spoofing protection ... done (rp_filter)
Configuring network interfaces ... ifup:interface lo already configured
SIOCSIFADDR: No such device
eth0: ERROR while getting interface flags: No such device
SIOCSIFNETMASK: No such device
eth0: ERROR while getting interface flags: No such device
Failed to bring up eth0
..... and the same for eth1 ....
Also,
Tux:/# find -name S40* -print
./etc/rc0.d/S40umountfs
./etc/rcS.d/S40shorewall
./etc/rcS.d/S40hostname.sh
./etc/rcS.d/S40networking
./etc/rcS.d/S40hotplug
./etc/rcS.d/S40ifrename
./etc/rc2.d/S40dhcp3-server
./etc/rc3.d/S40dhcp3-server
./etc/rc4.d/S40dhcp3-server
./etc/rc5.d/S40dhcp3-server
./etc/rc6.d/S40umountfs
In summary, there is a problem with rc scripts, pre-existing
installation of the 2.6.14 kernel.
After the OOPS I received heaps of 'rc' error messages - I reproduce (by
hand) some I noted
at the time .. apologies for any human error .. I still get them ..
---------------------------------
grep: error while loading shared libraries:libpcre.so.3:cannot open
shared object file: No such file or directory
/etc/rc5.d/S20linuxconf: line 18: /var/run/runlevel.dir:No such file or
directory
/etc/rc5.d/S20mon: line 32: basename: command not found
/etc/rc5.d/S20postgresql-7.4:line 3:
/usr/share/postgresql-common/init.d-functions: No such file or directory
/etc/rc5.d/S20postgresql-7.4: tr: command not found
/etc/rc5.d/S20Xprint: ## FATAL ERROR: Unknown XProjectRoot
Starting NFS common utilities: statdstatd: Could not chdir: No such file
or directory
grep: error while loading shared libraries:libpcre.so.3:cannot open
shared object file: No such file or directory
/usr/sbin/apache is not executable, not starting/reloading ... failed.
Debian GNU/Linux testing/unstable Tux tty 1
Tux login:
----------------------------
Postgresql and printing are not critical for booting or networking.
Having got as far as the GUI gives me grounds for optimism and I'll try
copying 'rc.sysinit ' from /opt/ltsp .. but the broader
question remains - is this fixable ? I don't know what else the OOPS did
to my system.
I have enough invested in this system to keep trying .. I'm hopeful
that connectivity will enable me
to dist-update/upgrade and fix these problems.
Apt and Mutt both work and enable me to use init 3, but at the moment do
not connect.
I'd be grateful for any constructive advice.
Adam Bogacki.
[EMAIL PROTECTED]
--
SLUG - Sydney Linux User's Group Mailing List - http://slug.org.au/
Subscription info and FAQs: http://slug.org.au/faq/mailinglists.html
