Main Page | Report Page

 

  Linux Forum Index » Linux - Slackware Forum » slackware 13.1 upgrade blowout...

Author Message
Mike Jones...
Posted: Thu Oct 21, 2010 3:30 am
 
Responding to Thomas Overgaard:

Quote:
Res wrote:

any suggestions?

It might be because the kernel in 13.1 is build to automatically detect
and build a RAID if possible and then sda and sdb doesn't exist anymore.

Try this at the boot-screen: Hit the <TAB> key and use this boot option:
Linux raid=noautodetect

If successful you could either build a new kernel without autodetection
or you could add this option to your /etc/lilo.conf:
append="raid=noautodetect"
and then run the lilo command.

Last solution might have a drawback because now your drives must be
plugged into the correct slots at the mobo. I had my second drive
plugged into the easiest accessible slot and it worked fine as sdb
during setup and install, but after I added this option it showed up as
sdf so I had to rearrange cabling to get it going.


All these new auto-everythingies are turning Linux into Widnos! >:(

"I'm sorry Dave. I can't allow you to do that."

Bah!

--
*=( http://www.thedailymash.co.uk/
*=( For all your UK news needs.
 
Thomas Overgaard...
Posted: Thu Oct 21, 2010 4:56 am
 
Res wrote:

Quote:
any suggestions?

It might be because the kernel in 13.1 is build to automatically detect

and build a RAID if possible and then sda and sdb doesn't exist anymore.

Try this at the boot-screen: Hit the <TAB> key and use this boot option:
Linux raid=noautodetect

If successful you could either build a new kernel without autodetection
or you could add this option to your /etc/lilo.conf:
append="raid=noautodetect"
and then run the lilo command.

Last solution might have a drawback because now your drives must be
plugged into the correct slots at the mobo. I had my second drive
plugged into the easiest accessible slot and it worked fine as sdb
during setup and install, but after I added this option it showed up as
sdf so I had to rearrange cabling to get it going.
--
Thomas O.

This area is designed to become quite warm during normal operation.
 
Martha Adams...
Posted: Thu Oct 21, 2010 9:39 am
 
On Thu, 21 Oct 2010 13:30:30 -0000, Mike Jones <luck at (no spam) dasteem.invalid>
wrote:

Quote:
Responding to Thomas Overgaard:

Res wrote:

any suggestions?

It might be because the kernel in 13.1 is build to automatically detect
and build a RAID if possible and then sda and sdb doesn't exist anymore.

Try this at the boot-screen: Hit the <TAB> key and use this boot option:
Linux raid=noautodetect

If successful you could either build a new kernel without autodetection
or you could add this option to your /etc/lilo.conf:
append="raid=noautodetect"
and then run the lilo command.

Last solution might have a drawback because now your drives must be
plugged into the correct slots at the mobo. I had my second drive
plugged into the easiest accessible slot and it worked fine as sdb
during setup and install, but after I added this option it showed up as
sdf so I had to rearrange cabling to get it going.


All these new auto-everythingies are turning Linux into Widnos! >:(

"I'm sorry Dave. I can't allow you to do that."

Bah!
=========================

I thought Slackware didn't do that egregious jazz. If they've started
that, then Slackware has lost its principal goodness.

Titeotwawki -- Martha Adams [2010 Oct 21]
 
Chris Ahlstrom...
Posted: Thu Oct 21, 2010 3:14 pm
 
Martha Adams pulled this Usenet face plant:

Quote:
I thought Slackware didn't do that egregious jazz. If they've started
that, then Slackware has lost its principal goodness.

Titeotwawki -- Martha Adams [2010 Oct 21]

Hi Martha! Miss you in the COLA cesspool! Glad to see you're still
Linuxing.

--
Angels we have heard on High
Tell us to go out and Buy.
-- Tom Lehrer
 
Res...
Posted: Thu Oct 21, 2010 6:11 pm
 
On Thu, 21 Oct 2010 12:56:34 +0200, Thomas Overgaard wrote:

Quote:
Res wrote:

any suggestions?

It might be because the kernel in 13.1 is build to automatically detect
and build a RAID if possible and then sda and sdb doesn't exist anymore.

Try this at the boot-screen: Hit the <TAB> key and use this boot option:
Linux raid=noautodetect

If successful you could either build a new kernel without autodetection
or you could add this option to your /etc/lilo.conf:
append="raid=noautodetect"
and then run the lilo command.


nope that not it sadly Sad
its almost like the custom kernel has decided it doesnt know about the
sata controller, not using modules, using builtin, cant figure out why it
worked in 13.0 but borks 13.1

I tried all sd* nothing is found :(

very strange, this is a (thankfully) non critical server, so no GUI.

Quote:

Last solution might have a drawback because now your drives must be
plugged into the correct slots at the mobo. I had my second drive
plugged into the easiest accessible slot and it worked fine as sdb
during setup and install, but after I added this option it showed up as
sdf so I had to rearrange cabling to get it going.

yeah, but lilo sees them as sda* and sdb* so should the kernel, I've just
booted back into opensuse live (PITA using a mouse in it LOL) chrooted to
sda1 and rebuilding the kernel, but it a;ll looks good
 
Mike Spencer...
Posted: Thu Oct 21, 2010 10:19 pm
 
"Martha Adams" <mhada at (no spam) verizon.net> wrote:

Quote:
I thought Slackware didn't do that egregious jazz. If they've
started that, then Slackware has lost its principal goodness.

Titeotwawki -- Martha Adams [2010 Oct 21]

I've already reverted my main box from Slack 12 to 11.0 because the
new libraries don't support a mission critical, non-OSS app.

Now it looks like I'll be sticking with 12.{0,1} (or 13.0? Is it worth
the trouble?) on the *other* boxen, more or less forever. (I'm old
enough that I might just get away with that.)

Just tonight, my wife, a not very gruntled Windows user, set up a
spare box for dialup net access and Windoes announced that it was
dialing 1-800-red-mond (or something) to get her a list of local ISPs
and immediately commenced to dial out. Gak! Why I've been
Microsoft-free for over a decade. Don't need all kinds of auto stuff,
don't want it, won't have it.

Titeotwawki indeed, Martha.

--
Mike Spencer Nova Scotia, Canada
 
Mike Jones...
Posted: Thu Oct 21, 2010 11:24 pm
 
Responding to Mike Spencer:

[...]
Quote:

Now it looks like I'll be sticking with 12.{0,1} (or 13.0? Is it worth
the trouble?) on the *other* boxen, more or less forever. (I'm old
enough that I might just get away with that.)



10.2 or 12.2, IME\IMO

--
*=( http://www.thedailymash.co.uk/
*=( For all your UK news needs.
 
Eef Hartman...
Posted: Fri Oct 22, 2010 5:47 am
 
Res <res at (no spam) ausics.net> wrote:
Quote:
00:1f.2 IDE interface: Intel Corporation 82801FB/FW (ICH6/ICH6W) SATA
Controller (rev 03)
Subsystem: Hewlett-Packard Company Device 3005
Kernel driver in use: ata_piix

That's for an Intel motherboard S-ATA controller (in your case the
ICH6 chipset) without "ahci" enabled in the bios (then the driver
would have been "ahci", of course).
As far as I know "ata_piix" also needs "libata", but that
one is needed for ahci too, at least, MY (newer) HP just uses
ahci and libata (HP xw4600 workstation - ICH9 chipset).
--
******************************************************************
** Eef Hartman, Delft University of Technology, dept. SSC/ICT **
** e-mail: E.J.M.Hartman at (no spam) tudelft.nl - phone: +31-15-27 82525 **
******************************************************************
 
Martin...
Posted: Fri Oct 22, 2010 12:53 pm
 
Res wrote:
Quote:
initial boot screen knows its there, tells me sda and sdb, reports all
the right number of sda1-sda8 ans sdb1-sdb13, gets to where it runs fsck
and aborts as device not found, enter rescue shell, no devices as found,
blkid also shows empty, cfdisk and fdisk report fatal error no devices sda
or sdb, even got adventourous and tried sd(a .. h) Smile

just rephrasing: are you saying, the kernel boot messages recognize the
disks and their partitions, but when you go into the rescue shell and ls
/dev/sd[ab]* nothing is found?

In that case, can you create the device files manually just for testing
purposes (man mknod)? If that works, most likely something in your
udev/dbus/hal configuration is the problem.

Martin
 
Res...
Posted: Fri Oct 22, 2010 7:37 pm
 
On Fri, 22 Oct 2010 20:53:09 +0200, Martin wrote:

Quote:
Res wrote:
initial boot screen knows its there, tells me sda and sdb, reports all
the right number of sda1-sda8 ans sdb1-sdb13, gets to where it runs
fsck and aborts as device not found, enter rescue shell, no devices as
found, blkid also shows empty, cfdisk and fdisk report fatal error no
devices sda or sdb, even got adventourous and tried sd(a .. h) :)

just rephrasing: are you saying, the kernel boot messages recognize the
disks and their partitions, but when you go into the rescue shell and ls
/dev/sd[ab]* nothing is found?


*nods*

Quote:
In that case, can you create the device files manually just for testing
purposes (man mknod)? If that works, most likely something in your
udev/dbus/hal configuration is the problem.



negative, tried all that, ro, wont re-mount rw.
it has to be kernel since using huge-kernel it works, so it's back online
temp to do its thing today, then its back offline, not needed again till
next weekend so i can play more monday.
 
Res...
Posted: Fri Oct 22, 2010 7:40 pm
 
On Fri, 22 Oct 2010 13:47:50 +0200, Eef Hartman wrote:

Quote:
Res <res at (no spam) ausics.net> wrote:
00:1f.2 IDE interface: Intel Corporation 82801FB/FW (ICH6/ICH6W) SATA
Controller (rev 03)
Subsystem: Hewlett-Packard Company Device 3005 Kernel driver in
use: ata_piix

That's for an Intel motherboard S-ATA controller (in your case the ICH6
chipset) without "ahci" enabled in the bios (then the driver would have
been "ahci", of course).
As far as I know "ata_piix" also needs "libata", but that one is needed
for ahci too, at least, MY (newer) HP just uses ahci and libata (HP
xw4600 workstation - ICH9 chipset).

yep, they are all included, so I'm guessing its an odd option in a
totally different config section, when i play on monday i'll try 2.6.36.,
its untar'd ready to go, not that i expect it to work either Smile
 
Res...
Posted: Fri Oct 22, 2010 7:45 pm
 
On Fri, 22 Oct 2010 11:36:56 +0000, Douglas Mayne wrote:

Quote:


Custom kernel is used (prolly tomora i'll resort to slackware huge
kern to see if it resolves, i also need sleep, had 7 hours in 3 days)
i am using the libsata stuff, even removed the deprecated ide stuff
incase something was used wrongly.



Yep, boots fine with PVs huge kernel
but the whole idea of custom kernels is to remove bloat and slackware's
kernel version is a tad old

The modular kernel is not the same thing exactly as the huge kernel. I

Yeah, but for all intents and purposes, my custom kernel is essentially
the modular.

Quote:


Yeah, it's a puzzle. I have mostly stopped tweaking the kernel because
there are too many places to introduce breakage. YMMV.

*nods*


Quote:

If you haven't ever built an initrd, then it is worth it just for the
exercise. It's not that hard, and at least on my hardware, it's much
quicker than recompiling the kernel. man mkinitrd

Built plenty, but I prefer in-kernel on critical options, in case someone
forgets it and thinks a system is hosed when it wont boot.
 
Res...
Posted: Fri Oct 22, 2010 11:22 pm
 
On Sat, 23 Oct 2010 11:37:28 +1000, Res wrote:

Quote:
On Fri, 22 Oct 2010 20:53:09 +0200, Martin wrote:

Res wrote:
initial boot screen knows its there, tells me sda and sdb, reports
all the right number of sda1-sda8 ans sdb1-sdb13, gets to where it
runs fsck and aborts as device not found, enter rescue shell, no
devices as found, blkid also shows empty, cfdisk and fdisk report
fatal error no devices sda or sdb, even got adventourous and tried
sd(a .. h) :)

just rephrasing: are you saying, the kernel boot messages recognize the
disks and their partitions, but when you go into the rescue shell and
ls /dev/sd[ab]* nothing is found?


*nods*

In that case, can you create the device files manually just for testing
purposes (man mknod)? If that works, most likely something in your
udev/dbus/hal configuration is the problem.



negative, tried all that, ro, wont re-mount rw. it has to be kernel
since using huge-kernel it works, so it's back online temp to do its
thing today, then its back offline, not needed again till next weekend
so i can play more monday.

I hate replying to myself, but Razz anyway the config for anyone bored
enough to bother is at http://members.ausics.net/res/fox.config
ive had enough of kernels for this weekend lol, I'll try again after the
weekend Wink
 
andrew...
Posted: Sat Oct 23, 2010 5:19 am
 
On 2010-10-23, Michael Talbot-Wilson <mtw at (no spam) calypso.view.net.au> wrote:

Quote:
I also had a bad experience upgrading from 13.0 to 13.1 and eventually
had to re-install 13.0. And I'd labelled all my partitions, hda and
sda, and was mounting on LABEL in /etc/fstab so the libata thing
shouldn't have made much difference.

Agree that Xorg is to blame for a lot of it. It could only stay up
for seconds in 13.1. A great pity that legacy XFree86 can't load Xorg
drivers.

I retreated to 13.0 after wrestling for some time with intel display
problems under 13.1, Slackware 13.0 being rock solid in this regard for
which I am grateful Smile. I am hoping for fewer problems with 13.2 when
it arrives....

Andrew
--
Do you think that's air you're breathing?
 
Michael Talbot-Wilson...
Posted: Sat Oct 23, 2010 5:14 pm
 
On 2010-10-23, Grant <omg at (no spam) grrr.id.au> wrote:

Quote:
I tend to edit 'vi /mnt/etc/lilo.conf' before the reboot to get rid of
first boot surprises Wink Follow with 'lilo -r /mnt'

I rarely use startx, so there's a range of issues I don't see.

I just noticed that I didn't follow the upgrade directions, didn't
really read them.

So the fault is all mine, presumably.

Can't blame Slackware.
 
 
Page 1 of 2    Goto page 1, 2  Next
All times are GMT - 5 Hours
The time now is Sun Apr 20, 2014 9:25 pm