System Drops to manual mount root prompt after HDD duplication

[David Cramblett]

My FreeBSD 5.2.1 server had a 4.5 GB HDD. I decided to upgrade it with
a larger drive. I installed a new drive on the second IDE channel which
made it ad2, of course, my original drive was ad0. I created a
partition, boot loader and matching slices on the new drive. Then I
copied the old drive to the new drive using tar. Once finished, I
removed the original drive and installed the new one on the primary
channel. When I booted up everything appeared normal, but when the
system starts to mount "/" it gives no error or warning and just drops
to a "Manual mount root specification" prompt. If I type "ufs:ad0s1a"
it boots up and everything is perfect. This is the same slice "/" was
on the old drive as well.


I have tried the following with no success:

Checked /etc/fstab

boot0cfg -v -B ad0

bsdlabel -B ad0s1

tried booting from a cd, going into post install config, fdisk, and set
the partition as bootable, it already was.

Since upgrading the hard disk, I have upgraded the system to 5.5 and
then to 6.2. This system has been working great for over a week now,
just have this boot problem.


--------------

Here is my fstab:

/dev/ad0s1b none swap sw 0 0
/dev/ad0s1a / ufs rw 1 1

--------------

Output from bsdlabel
# bsdlabel ad0s1

# /dev/ad0s1:
8 partitions:
# size offset fstype [fsize bsize bps/cpg]
a: 585018626 1048576 4.2BSD 2048 16384 28552
b: 1048576 0 swap
c: 586067202 0 unused 0 0 # "raw" part,
don't edit


--------------

Output from boot0cfg
# boot0cfg -v ad0
# flag start chs type end chs offset size
1 0x80 0: 1: 1 0xa5 1023:254:63 63 586067202

version=1.0 drive=0x80 mask=0xf ticks=182
options=nopacket,update,nosetdrv
default_selection=F1 (Slice 1)



Thanks,

David

[Dag-Erling Smørgrav]

Post by David Cramblett
My FreeBSD 5.2.1 server had a 4.5 GB HDD. I decided to upgrade it
with a larger drive. I installed a new drive on the second IDE
channel which made it ad2, of course, my original drive was ad0. I
created a partition, boot loader and matching slices on the new drive.
Then I copied the old drive to the new drive using tar. Once
finished, I removed the original drive and installed the new one on
the primary channel. When I booted up everything appeared normal, but
when the system starts to mount "/" it gives no error or warning and
just drops to a "Manual mount root specification" prompt. If I type
"ufs:ad0s1a" it boots up and everything is perfect. This is the same
slice "/" was on the old drive as well.

What's in your /boot.config and /boot/loader.conf?

DES

[Eric Anderson]

Post by David Cramblett
My FreeBSD 5.2.1 server had a 4.5 GB HDD. I decided to upgrade it with
a larger drive. I installed a new drive on the second IDE channel which
made it ad2, of course, my original drive was ad0. I created a
partition, boot loader and matching slices on the new drive. Then I
copied the old drive to the new drive using tar. Once finished, I
removed the original drive and installed the new one on the primary
channel. When I booted up everything appeared normal, but when the
system starts to mount "/" it gives no error or warning and just drops
to a "Manual mount root specification" prompt. If I type "ufs:ad0s1a"
it boots up and everything is perfect. This is the same slice "/" was
on the old drive as well.
Checked /etc/fstab
boot0cfg -v -B ad0
bsdlabel -B ad0s1
tried booting from a cd, going into post install config, fdisk, and set
the partition as bootable, it already was.
Since upgrading the hard disk, I have upgraded the system to 5.5 and
then to 6.2. This system has been working great for over a week now,
just have this boot problem.
--------------
/dev/ad0s1b none swap sw 0 0
/dev/ad0s1a / ufs rw 1 1
--------------
Output from bsdlabel
# bsdlabel ad0s1
# size offset fstype [fsize bsize bps/cpg]
a: 585018626 1048576 4.2BSD 2048 16384 28552
b: 1048576 0 swap
c: 586067202 0 unused 0 0 # "raw" part,
don't edit

Could it be because your root partition is not at offset 0?

Eric

[David Cramblett]

Post by Dag-Erling Smørgrav

Post by David Cramblett
My FreeBSD 5.2.1 server had a 4.5 GB HDD. I decided to upgrade it
with a larger drive. I installed a new drive on the second IDE
channel which made it ad2, of course, my original drive was ad0. I
created a partition, boot loader and matching slices on the new drive.
Then I copied the old drive to the new drive using tar. Once
finished, I removed the original drive and installed the new one on
the primary channel. When I booted up everything appeared normal, but
when the system starts to mount "/" it gives no error or warning and
just drops to a "Manual mount root specification" prompt. If I type
"ufs:ad0s1a" it boots up and everything is perfect. This is the same
slice "/" was on the old drive as well.

What's in your /boot.config and /boot/loader.conf?
DES

I have no boot.config.


[***@www /]# cat /boot/loader.conf
# -- sysinstall generated deltas -- #
userconfig_script_load="YES"


I also tried an option I found while searching for a solution

root_disk_unit="0"

although that did not help.

David

[David Cramblett]

Post by Eric Anderson

Post by David Cramblett
My FreeBSD 5.2.1 server had a 4.5 GB HDD. I decided to upgrade it
with a larger drive. I installed a new drive on the second IDE
channel which made it ad2, of course, my original drive was ad0. I
created a partition, boot loader and matching slices on the new
drive. Then I copied the old drive to the new drive using tar. Once
finished, I removed the original drive and installed the new one on
the primary channel. When I booted up everything appeared normal, but
when the system starts to mount "/" it gives no error or warning and
just drops to a "Manual mount root specification" prompt. If I type
"ufs:ad0s1a" it boots up and everything is perfect. This is the same
slice "/" was on the old drive as well.
Checked /etc/fstab
boot0cfg -v -B ad0
bsdlabel -B ad0s1
tried booting from a cd, going into post install config, fdisk, and
set the partition as bootable, it already was.
Since upgrading the hard disk, I have upgraded the system to 5.5 and
then to 6.2. This system has been working great for over a week now,
just have this boot problem.
--------------
/dev/ad0s1b none swap sw 0 0
/dev/ad0s1a / ufs rw 1 1
--------------
Output from bsdlabel
# bsdlabel ad0s1
# size offset fstype [fsize bsize bps/cpg]
a: 585018626 1048576 4.2BSD 2048 16384 28552
b: 1048576 0 swap
c: 586067202 0 unused 0 0 # "raw" part,
don't edit

Could it be because your root partition is not at offset 0?
Eric

I don't think so, but I certainly could be wrong. Here is another system:

# bsdlabel ad0s1
# /dev/ad0s1:
8 partitions:
# size offset fstype [fsize bsize bps/cpg]
a: 18932866 137363456 4.2BSD 2048 16384 28552
b: 1048576 0 swap
c: 156296322 0 unused 0 0 # "raw" part,
don't edit
d: 62914560 1048576 4.2BSD 2048 16384 28552
e: 73400320 63963136 4.2BSD 2048 16384 28544


here s1a is not at offset 0, yet the system boots fine. I create the
swap partition first usually, hence the offset being 0 for swap.

David

[Dag-Erling Smørgrav]

Post by David Cramblett

Post by Dag-Erling Smørgrav
What's in your /boot.config and /boot/loader.conf?

I have no boot.config.
# -- sysinstall generated deltas -- #
userconfig_script_load="YES"

Beats me... I can't even remember what userconfig_script is supposed to
do. Note that support for 5.2.1 ended on July 31, 2004.

DES

[Matthew D. Fuller]

On Tue, May 15, 2007 at 08:20:09AM +0200 I heard the voice of

Post by Dag-Erling Smørgrav

Post by David Cramblett

Post by Dag-Erling Smørgrav
What's in your /boot.config and /boot/loader.conf?

I have no boot.config.
# -- sysinstall generated deltas -- #
userconfig_script_load="YES"

Beats me... I can't even remember what userconfig_script is
supposed to do. Note that support for 5.2.1 ended on July 31, 2004.

Neither can I, but last weekend I upgraded some 4.x boxes to 5.x and
then 6.x, and with the 5.x loader they croaked before booting until I
took those userconfig_script_load lines out.

[Dag-Erling Smørgrav]

Post by Matthew D. Fuller

Post by Dag-Erling Smørgrav
Beats me... I can't even remember what userconfig_script is
supposed to do. Note that support for 5.2.1 ended on July 31, 2004.

Neither can I, but last weekend I upgraded some 4.x boxes to 5.x and
then 6.x, and with the 5.x loader they croaked before booting until I
took those userconfig_script_load lines out.

Ooh, now I remember; userconfig was the old hardware configuration
mechanism we had before device.hints. If you typed -c at the boot1
prompt, you would get an interactive display where you could assign IRQs
and ports to devices.

The script is basically a sequence of userconfig commands; you could
extract it from kenv after an interactive boot and store it in a file
which the loader would read the next time around.

AFAIR, userconfig was axed between FreeBSD 3 and 4, so there would be no
point in loading a userconfig script for a 5.x kernel.

DES

[David Cramblett]

Post by Dag-Erling Smørgrav

Post by Matthew D. Fuller

Post by Dag-Erling Smørgrav
Beats me... I can't even remember what userconfig_script is
supposed to do. Note that support for 5.2.1 ended on July 31, 2004.

Neither can I, but last weekend I upgraded some 4.x boxes to 5.x and
then 6.x, and with the 5.x loader they croaked before booting until I
took those userconfig_script_load lines out.

Ooh, now I remember; userconfig was the old hardware configuration
mechanism we had before device.hints. If you typed -c at the boot1
prompt, you would get an interactive display where you could assign IRQs
and ports to devices.
The script is basically a sequence of userconfig commands; you could
extract it from kenv after an interactive boot and store it in a file
which the loader would read the next time around.
AFAIR, userconfig was axed between FreeBSD 3 and 4, so there would be no
point in loading a userconfig script for a 5.x kernel.
DES

Thanks for the tip. I did remove the line, however my issue did not go
away. I am still getting dropped to the manual mount root prompt in the
middle of the boot sequence.

David