LinuxQuestions.org

LinuxQuestions.org (/questions/)
-   Slackware - Installation (https://www.linuxquestions.org/questions/slackware-installation-40/)
-   -   Won't boot after Slackware 11.0 installation (https://www.linuxquestions.org/questions/slackware-installation-40/wont-boot-after-slackware-11-0-installation-528790/)

vasco-t 02-13-2007 04:56 PM

Won't boot after Slackware 11.0 installation
 
Hi,
I am new to Linux so apologies if I don't describe my problem correctly.

I have spent a day installing SLACKWARE 11.0 on an old DEC Celebris PC that has 128Mb RAM, one 4Gb SCSI disk and two CD drives.

The BIOS will not boot from the CD drive so I created a boot floppy using the SCSI2.S kernel and then created two install floppies. These I used to start and run the install process. I partitioned the HDD as 1Gb Linux Swap and 3Gb Linux date (ext2).

I ran the "newbie" installation option as I only wanted to install a fairly minimal system with XWindows and KDE bearing in mind that my HDD was rather small.

The installation went well. At the end I confirmed that I wanted to use the SCSI2.S kernel and I created an emergency floppy as well as installing LILO.

When I booted the new system the boot process seemed to go well up until after it had scanned the SCSI controller for devices. Then it gave me the error:

"Unable to handle kernel paging request at virtual address 0832d06"

It then continued the process and ended with:

<0> Kernel panic: Aiee, killing interupt handler!
ln interrupt handler - not syncing


I would like to print all of the boot message but I don't know how to trap the message in a file and copy it to this forum as I have seen many other members do, (sorry, I'm new to Linux).

Can anybody help me please?

Thanks.

onebuck 02-14-2007 08:15 AM

Quote:

Originally Posted by vasco-t
Hi,
I am new to Linux so apologies if I don't describe my problem correctly.

I have spent a day installing SLACKWARE 11.0 on an old DEC Celebris PC that has 128Mb RAM, one 4Gb SCSI disk and two CD drives.

The BIOS will not boot from the CD drive so I created a boot floppy using the SCSI2.S kernel and then created two install floppies. These I used to start and run the install process. I partitioned the HDD as 1Gb Linux Swap and 3Gb Linux date (ext2).

I ran the "newbie" installation option as I only wanted to install a fairly minimal system with XWindows and KDE bearing in mind that my HDD was rather small.

The installation went well. At the end I confirmed that I wanted to use the SCSI2.S kernel and I created an emergency floppy as well as installing LILO.

When I booted the new system the boot process seemed to go well up until after it had scanned the SCSI controller for devices. Then it gave me the error:

"Unable to handle kernel paging request at virtual address 0832d06"

It then continued the process and ended with:

<0> Kernel panic: Aiee, killing interupt handler!
ln interrupt handler - not syncing


I would like to print all of the boot message but I don't know how to trap the message in a file and copy it to this forum as I have seen many other members do, (sorry, I'm new to Linux).

Can anybody help me please?

Thanks.


Hi,

First, you don't need that big of a swap! You could set the swap to 256MB and you would have enough space. The 1GB of swap is a waste.

What type of scsi controller?

You could boot using the install cd and get information;

Code:

#lspci -vv            #find out what the kernel finds
#dmesg |grep hd      #what hd devices
#dmesg |grep sd      #what scsi devices

By using this information you can decide what kernel you should use.

You should read the Release notes. I would select the huge26 kernel at this point if you are not sure. If you read the release notes, you will see why.

The reason I'm suggesting the huge26 is that the kernel supports about everything but the kitchen sink. From the Release notes;

Quote:

These kernels have compiled-in support
for essentially all of the hard drive controllers and filesystems that
are supported by the 2.6.x kernel series, and should be helpful for
people with new SATA controllers or other hardware that 2.4.33.3
doesn't support. This kernel doesn't have any built in network support
and is primarily intended for installing from local media (although
huge26.s will support an NFS installation if you run the 'network'
script to load your network module, and you're installing from CD or
DVD). After the packages have installed and you've reached the menu
where you are prompted to select a kernel to install, BE CERTAIN to
install huge26.s or test26.s (whichever one you used to install) or
there's a chance that your machine will be unable to boot with the
stock sata.i kernel that you'll end up with.

Also, the kernel module packages for 2.6.x are not part of the
standard installation, so if you install using huge26.s or test26.s
kernel you'll have to install the corresponding kernel-modules package
when you're done. They can be found under /extra/linux-2.6.17.13/ (or
ISO number two under /extra/linux-2.6.17.13), or
/testing/packages/linux-2.6.18/ (or ISO number four under
/testing-2.6.18/packages/linux-2.6.18). You might also consider
installing the other packages there and building an initrd according to
the instructions in README.initrd if you'd like to be running a slimmer
kernel without a lot of unneeded drivers taking up RAM, or better yet,
visit ftp.kernel.org in /pub/linux/kernel/v2.6/ and grab the latest
version of the 2.6 kernel to build your own. It'll probably have some
important fixes.
BTW, Welcome to Slackware!

You should check my sig for some good online reference.

vasco-t 02-14-2007 09:51 AM

Thanks Gary.
I can't boot directly from the CD so I am using floppies.

My SCSI controller is an on-board NCR SDMS V3.0 PCI SCSI V2.0
Tried the lspci -vv but got "No such file or directory"
The two dmesg commands tells me that:

hda and hdb are both CD ROM drives attached to the ide-cdrom driver

sda is a 4513Mb SCSI disk chan 0, id 0, lun 0
partitions sda1 = Linux data (bootable)
sda2 = Linux swap

Should I try to reduce the size of the swap partition by deleting it and re-creating it with cfdisk ?

My kernel is scsi2.s as that is the one on the floppy.

The HDD (sda1) boots and starts the boot sequence but it is within this process that it fails.

How can I change the kernel?

Cheers

bioe007 02-15-2007 06:55 PM

for getting info to files:

Code:

$dmesg > myPath/myFilename
you can send the output of any (?) command to a file with the > operator. To catch only errors I think you do:

Code:

$dmesg &> myPath/myFilename
for resizing stuff i like gPartEd - but that makes a live CD and sounds like your machine doesnt like it.

to copy the kernel:

1- insert Slack11 D1 # I think this its disk 1
Code:

mount /mnt/cdrom 
# move to directory containing Huge26.s
cp Huge26.s /boot
vi /etc/lilo.conf # to include an entry for Huge26.s
/sbin/lilo

7- reboot

Some quick editor commands to help with lilo.conf: copy = yy , paste = p, deletechar = x, inserttext = a

I find its easiet to just use yy and p to duplicate the lilo.conf entry thats already there, then edit the image and label entries to reflet the new kernel.

Alternatively you can make a new boot floppy that contains the Huge26.s kernel much the same way you made your scsi2.s floppy. (makebootdisk)

Quote:

Originally Posted by vasco-t
lspci -vv but got "No such file or directory"

It might be easier to re-install Slack and select Huge26.s as your initial kernel. Then once up and running you can start learning more.

While this seems like a real pain, sometimes when first starting out its easier to do a full install and then remove stuff later to free up disk space. <-IMHO

I'd also recommend against installing KDE if processor cycles and disk space are at a premium. I think fluxbox might run better on your system. Its harder to make look fancy, but you'll be staring at a bouncing gear forever in KDE :)

Sorry these are fast & loose directions, I don't have my Slack11 CDs with me.

welcome to Slackware - its fun in here :)

onebuck 02-16-2007 07:33 AM

Quote:

Originally Posted by vasco-t
Thanks Gary.
I can't boot directly from the CD so I am using floppies.

My SCSI controller is an on-board NCR SDMS V3.0 PCI SCSI V2.0
Tried the lspci -vv but got "No such file or directory"
The two dmesg commands tells me that:

hda and hdb are both CD ROM drives attached to the ide-cdrom driver

sda is a 4513Mb SCSI disk chan 0, id 0, lun 0
partitions sda1 = Linux data (bootable)
sda2 = Linux swap

Should I try to reduce the size of the swap partition by deleting it and re-creating it with cfdisk ?

My kernel is scsi2.s as that is the one on the floppy.

The HDD (sda1) boots and starts the boot sequence but it is within this process that it fails.

How can I change the kernel?

Cheers

Hi,

To boot from any recognized media, try the sbootmgr. Just create the sbootmgr from /rootdisk with the rawrite command. Once you create this you can then boot that floppy and you will see any recognized bootable devices. Just cursor down to the desired device, of course you should have the install cd1 in the cdrom.

As for the swap size, I would just start my installation over. Just create the new swap and / with the desired size then do the installation. If you do a full install then you will be limited with hd space available. You can always use pkgtool to trim the install.

Since the kernel boots with the floppies you created, just select that kernel from the install cd since the cd is recognized. This way you will get a good copy. I just don't trust floppy use that much anymore.
I've had a lot of problems over the years with floppy media. Not my choice today.

bioe007 gave you some advice on how to capture files. If you have the space with the current system on the hard drive, just create a small partition to use as a transfer medium. Especially if you will use M$ windows on the machine. You could redirect the capture to the fd0 (floppy) that is mounted with a usable filesystem or a flash drive if you can mount one.

You should really read some of the online reference in my sig to get some background. Several good guides.

I concur that you should use something other than kde on your system. I prefer xfce4, I use it on a lot of the older equipment. My laptop, the one I use to access my machines, LQ and whatever is a old Gateway 2500 solo PII 333 with 192 MB ram (this helps here). The laptop runs smooth with the xfce4 and has a small footprint with a great display. I couldn't beat the cost.

vasco-t 02-17-2007 09:33 AM

Thanks BIOE007 and ONEBUCK foe your advice and help.

In fact I rebuilt SlakWare on the HDD and installed LILO on the MBR. Then I somehow managed to get the huge26 kernel onto the HDD, (don't ask me how!). I then edited LILO.CONF adding a second pointer to it. I renamed the labels on both pointers do that I would recognise the choices offered. On rebooting I expected LILO to come up with the two possible kernels for me to choose but it only came up with one and it wasn't either of my modified labels, wierd. The boot continued successfully and did not get my error message; I got right to the end !!!
However, I ma not sure what kernel I am using. Is there a way to check?

I logged in as root and tried the LSCPI and DMESG commands as advised and this time they worked. The results are as follows:

Quote:

00:00.0 Host bridge: Intel Corporation 82434LX/NX [Mercury/Neptune] Processor to PCI bridge (rev 11)
Control: I/O- Mem+ BusMaster+ SpecCycle- MemWINV- VGASnoop- ParErr+ Stepping- SERR+ FastB2B-
Status: Cap- 66MHz- UDF- FastB2B- ParErr- DEVSEL=slow >TAbort- <TAbort- <MAbort+ >SERR- <PERR-
Latency: 32

00:01.0 SCSI storage controller: LSI Logic / Symbios Logic 53c810 (rev 02)
Control: I/O+ Mem+ BusMaster+ SpecCycle- MemWINV- VGASnoop- ParErr- Stepping- SERR+ FastB2B-
Status: Cap- 66MHz- UDF- FastB2B- ParErr- DEVSEL=medium >TAbort- <TAbort- <MAbort- >SERR- <PERR-
Latency: 64
Interrupt: pin A routed to IRQ 11
Region 0: I/O ports at ec00 [size=256]
Region 1: Memory at fedfbf00 (32-bit, non-prefetchable) [size=256]

00:02.0 Non-VGA unclassified device: Intel Corporation 82378ZB/IB, 82379AB (SIO, SIO.A) PCI to ISA Bridge (rev 88)
Control: I/O+ Mem+ BusMaster+ SpecCycle+ MemWINV- VGASnoop- ParErr- Stepping- SERR- FastB2B-
Status: Cap- 66MHz- UDF- FastB2B- ParErr- DEVSEL=medium >TAbort- <TAbort- <MAbort- >SERR- <PERR-
Latency: 0

00:06.0 VGA compatible controller: Matrox Graphics, Inc. MGA 2064W [Millennium] (rev 01) (prog-if 00 [VGA])
Control: I/O+ Mem+ BusMaster- SpecCycle- MemWINV- VGASnoop- ParErr- Stepping+ SERR- FastB2B-
Status: Cap- 66MHz- UDF- FastB2B+ ParErr- DEVSEL=medium >TAbort- <TAbort- <MAbort- >SERR- <PERR-
Interrupt: pin A routed to IRQ 9
Region 0: Memory at fedfc000 (32-bit, non-prefetchable) [size=16K]
Region 1: Memory at fe000000 (32-bit, prefetchable) [size=8M]
Expansion ROM at <unassigned> [disabled] [size=64K]

Quote:

hda: TOSHIBA CD-ROM XM-5302TA, ATAPI CD/DVD-ROM drive
hdb: TOSHIBA CD-ROM XM-5602B, ATAPI CD/DVD-ROM drive
hda: attached ide-cdrom driver.
hda: ATAPI 4X CD-ROM drive, 256kB Cache
hdb: attached ide-cdrom driver.
hdb: ATAPI 8X CD-ROM drive, 256kB Cache
SCSI device sda: 8813870 512-byte hdwr sectors (4513 MB)
Quote:

Attached scsi disk sda at scsi0, channel 0, id 0, lun 0
SCSI device sda: 8813870 512-byte hdwr sectors (4513 MB)
sda: sda1 sda2
UMSDOS 0.86k (compatibility level 0.4, fast msdos)
Next I tried STARTX but it came up with the message:
Quote:

Could not start kstartupconfig. Check your installation.
I guess there is some configuration to be done there.

I will have a look at all the places you both suggested and try to boost my knowledge a bit.

Oh, by the way, the Celebris is a dual processor PC. Don't know if that makes a difference to SlackWare.

Cheers and many thanks.

ryanoa 02-17-2007 05:16 PM

Anytime you edit lilo.conf you need to run
Code:

/sbin/lilo
to make the changes stick.

To take advantage of your dual processors you need to compile your own kernel with support for smp or you could install the linux-smp-2.6.17.13 kernel from the /extra directory on the cd (If you do so, don't forget to install the modules also)

For the x errors: I would run xorgsetup to get started. Hopefully that will make an acceptable xorg.conf for you. If not, run xorgconfig and manually enter your setup info.

Hope this helps,
Ryan

bioe007 02-17-2007 07:10 PM

Quote:

Originally Posted by vasco-t
Thanks BIOE007 and ONEBUCK foe your advice and help.

welcome to Slackware. A newbie myself, I have never found anything less than exemplary help here at LQ. I'm just a newbie trying to help out others as I have been helped so much.

I'm still strongly suggesting an alternative to KDE, it really is a processor hog (ok, not really fair to KDE but by comparison, it is).

I should have told you this before, but to change from the default (KDE) windows manager to another you need to modify a file called '.xinitrc'

try this command:

$ echo startfluxbox > /home/<username>/.xinitrc # where <username> is your username

that should start you off in fluxbox (i'm not being paid for advertising or anything, but thats the only small WM I have experience with, fxce4 may be a better alternative, but I dont know anything about it.

another fopah (sp?) I committed early on was trying to run X as root, if you're doing that add another user (use adduser, it will prompt you for everything needed). Then log in as that user and you can try startx from there.

Once you get X up and running you can start tinkering more.

vasco-t 02-20-2007 03:30 AM

Thanls Ryan and Bioe007, much appreciate your help.

Ryan: Tried the LILO stuff and it worked a treat
Will try the kernel compile later as th thought of it scares the socks off of me!
Neither xorgsetup not xorgconfig fixed the problem; xwindows starts (x cursor on screen then a window telling me "Could not start kstartupconfig. Check your installation."

Bioe007: I will try your alternative but thought I would create a new user and try the startx from there: result - same error as noted to Ryan above.
So it's on the FLUXBOX...

(By the way, it's spelt "faux pas")

vasco-t 02-20-2007 04:00 AM

Tried the "echo startfluxbox > /home/<username>/.xinitrc" in my new user account, logged out, logged in again and nothing happened - just the normal Linux prompt.

Is FLUXBOX installed by default?

Checked the .initrc file and it just had startfluxbox in it.

All these different window systems are a bit confusing.
What are their relationships?

Cheers

erklaerbaer 02-20-2007 09:21 AM

as ryanoa said, you don't need to compile a kernel for yourself; just install the 2.6.17.13-smp one (+modules).

kstartupconfig is a kde program. if you want to use kde you need at least kdebase kdelibs and qt.

your X server seems to be running( although probably not properly configured )

how do you start up X? startx from a shell, XDM or KDM? you know KDM ignores the .xinitrc :)


for the future: please start one thread per problem ;)

vasco-t 02-20-2007 11:54 AM

Sorry, I'll start a new thread on SMP

vasco-t 02-20-2007 03:09 PM

Well, now I set up the run level to be 4, rebooted and this kernel now hangs after "Starting up X11 session manager..."

So now it's reformat HDD and start from scratch again.
Wonder if it will ever work...

bioe007 02-20-2007 03:15 PM

You don't have to despair and reinstall everything...

When it hangs try either/all of these: Ctrl+Alt+Backspace, Ctrl+F6 then Ctrl+C

that should get you back to the shell prompt. Then immediately go and change your inittab back to runlevel 3.. ;)

<later>
There are a couple of things that could be wrong here:

1) your X server is not configured, did you run (as root) xorgsetup?

1.b) If not, then run it (#xorgsetup) hopefully that works, then try startx, as a normal user again.

1.b.1) if xorgsetup crumps on you try xorgconfig (also as root), if you get through that try startx as a normal user again.

1.c) try the vesa xorg.conf (as root):
# mv /etc/X11/xorg.conf /etc/X11/MYxorg.conf
# cp /etc/X11/xorg.conf-vesa /etc/X11/xorg.conf
# exit
log in and try startx (it wont be pretty but will/should isolate the problem)

I have seen this problem where Slack11 puts fluxbox somewhere the fluxbox startup script does not find it.

Let me switch to linux and I'll check for specifics on my computer..

<later>
can you:
Code:

$ slocate fluxbox | grep bin
$ vi ~/.fluxbox/startup

check the last (or nearly last line) ?

exec /usr/local/bin/fluxbox -log "~/.fluxbox/log"

That is mine, but I think the sometimes the path looks in X11/bin or something. In which case you should fix it too look like whatever the path you found to fluxbox is from your slocating..

At any rate, don't set your machine to runlevel 4 until you get your x server set up. In fact, I never boot into runlevel 4 (not that theres anything wrong with it), i just dont mind typing startx.

-WM- Window managers are just what they sound like, fluxbox KDE and GNOME are variants of the same layer that rides atop the 'X' server, which is what allows stuff to be graphical on a linux machine. You need both to have a desktop with windows/icons etc...

The ability to use different window managers is -one- of the many niceties about linux, on fast machines we can choose a heavy duty, fully featured lots of bells and whistles WM (KDE) or on an old machine (like my P3) you can choose something bare-bones that maybe requires a bit more learning to fully configure the way you like it.

(Disclaimer-> ^That^ is a quick summary by a fellow N00B)

vasco-t 02-20-2007 03:35 PM

Tried both of your suggestions individually and together:
Quote:

Ctrl+Alt+Backspace, Ctrl+F6 then Ctrl+C
Neither worked. The Ctrl+F6 just echoed what looks like an ascii keycode.
Only way I seem to be able to move forward is ctl+alt+delete (reboot)


All times are GMT -5. The time now is 05:20 PM.