[SOLVED] Slackware 14.1 raid install Kernel Panic

Eeel · 01-25-2015, 04:29 PM

Hello,

I'm having a Kernel Panic on reboot after installed Slackware 14.1 on two sata drive with Raid 1 and 5, this is exactly what i have done:

Run Slackware 14.1 64bit install from USB, partitioning with cfdisk:

Code:

/boot	100Mo	sda1	DA	RAID 1	md3	
/	25Go	sda2	DA	RAID 5	md0
/home	50G	sda3	DA	RAID 5	md1
swap	4Go	sda4	DA	RAID 1	md2
free	~100mo

partition type DA

Duplicate partition to second 80Go drive with sfdisk:

Code:

sfdisk -d /dev/sda | sfdisk /dev/sdb

Create new array with mdadm:

Code:

/:	mdadm --create /dev/md0 --level 5 --raid-devices 2 /dev/sda2 /dev/sdb2 --metadata=1.2
/home:	mdadm --create /dev/md1 --level 5 --raid-devices 2 /dev/sda3 /dev/sdb3 --metadata=1.2
swap:	mdadm --create /dev/md2 --level 1 --raid-devices 2 /dev/sda4 /dev/sdb4 --metadata=1.2
/boot:	mdadm --create /dev/md3 --level 1 --raid-devices 2 /dev/sda1 /dev/sdb1 --metadata=0.90

Format swap partition with mkswap:

Code:

mkswap -f /dev/md2

Format other array with mkfs:

Code:

mkfs.ext4 /dev/md0
mkfs.ext4 /dev/md1
mkfs.ext4 /dev/md3

Install Slackware using "setup":

- Install LILO "simple" setup
- OS/2 Boot Manager found: NO
- Additional append parameter: root=/dev/md0
- Install LILO in MBR
- EXIT setup

Switch to installation directory with chroot:

Code:

chroot /mnt/

Edit lilo.conf:

Code:

# Start LILO global section
raid-extra-boot = mbr-only
# Append any additional kernel parameters:
append="root=/dev/md0 vt.default_utf8=0"
boot = /dev/md3

# End LILO global section
# Linux bootable partition config begins
image = /boot/vmlinuz
  initrd = /boot/initrd.gz
  root = /dev/md0
  label = Linux
  read-only
# Linux bootable partition config ends

Using the Generic Kernel or huge kernel result the same error, generic kernel:

Code:

 - cd /boot
 - rm vmlinuz System.map config
 - ln -s vmlinuz-generic-*     vmlinuz
 - ln -s System.map-generic-*  System.map
 - ln -s config-generic-*      config

huge kernel(i use this setup with previous Slackware 14):

Code:

 - cd /boot
 - rm vmlinuz System.map config
 - ln -s vmlinuz-huge-*     vmlinuz
 - ln -s System.map-huge-*  System.map
 - ln -s config-huge-*      config

Create mkinitrd.conf:

Code:

cp /etc/mkinitrd.conf.sample /etc/mkinitrd.conf

Edit mkinitrd.conf:

Code:

# mkinitrd.conf.sample
# See "man mkinitrd.conf" for details on the syntax of this file
#
SOURCE_TREE="/boot/initrd-tree"
#CLEAR_TREE="0"
OUTPUT_IMAGE="/boot/initrd.gz"
KERNEL_VERSION="$(uname -r)"
KEYMAP="fr"
MODULE_LIST="mbcache:jbd2:ext4"
#LUKSDEV="/dev/sda2"
ROOTDEV="/dev/md0"
ROOTFS="ext4"
#RESUMEDEV="/dev/sda2"
RAID="1"
#LVM="1"
#UDEV="1"
#MODCONF="0"
#WAIT="1"

i also have tested to create mkinitrd.conf with the following command:

Code:

/usr/share/mkinitrd/mkinitrd_command_generator.sh > /etc/mkinitrd.conf

Create the initrd with mkinitrd:

Code:

mkinitrd -F

Run Lilo:

Code:

lilo

Declare email and Raid array in /etc/mdadm.conf:

Code:

echo MAILADDR my-mail@somewhere.com >> /etc/mdadm.conf
mdadm -Es >> /etc/mdadm.conf

I have attached the screenshot of kernel Panic error, i continue to search on google with various quote of the error message, help appreciated

Richard Cranium · 01-25-2015, 06:09 PM

Re-run mkinitrd and lilo now that you've initialized /etc/mdadm.conf with the current configuration.

The default /etc/mdadm.conf contains nothing but comments. The init script in the initrd does...

Code:

  # Initialize RAID:
  if [ -x /sbin/mdadm ]; then
    # If /etc/mdadm.conf is present, udev should DTRT on its own;
    # If not, we'll make one and go from there:
    if [ ! -r /etc/mdadm.conf ]; then
      /sbin/mdadm -E -s >/etc/mdadm.conf
      /sbin/mdadm -S -s
      /sbin/mdadm -A -s
      # This seems to make the kernel see partitions more reliably:
      fdisk -l /dev/md* 1> /dev/null 2> /dev/null
    fi
  fi

...which means that udev will mount the software RAID devices with names such as /dev/md128 versus what you chose.

If you *don't* have an /etc/mdadm.conf file in your initrd, the above code will force a rescan of the RAID devices. If you have a correct /etc/mdadm.conf, udev will get it right the first time it sees the drives.

Eeel · 01-26-2015, 02:58 AM

Hello Richard,

Thanks for your answer, i'm at work at the moment, will be able to test in ~10 hours

From what i know i have things written in /etc/mdadm.conf and same in the /boot/initrd-tree/etc/mdadm.conf, my mdamdm.conf contain my 4 ARRAY and my mail, after reading a thread yesterday about the md128 and md129 problem, i also added 4 more lines, one for each md* (i don't remember the exact syntaxe i have used but it was something like that):

Code:

ARRAY /dev/md0 devices=sda2,sdb2
ARRAY /dev/md1 devices=sda3,sdb3
....

While i was doing some various test yesterday i have re run several time the mkinirtd -F and lilo commands with a mdadm.conf file containing my 4 ARRAYS.

Just to be shure i'm not misunderstanding:

Quote:

Originally Posted by Richard Cranium

The default /etc/mdadm.conf contains nothing but comments. The init script in the initrd does...

If you *don't* have an /etc/mdadm.conf file in your initrd

the initrd you are talking is the inird-tree or the initrd.gz file?

Richard Cranium · 01-26-2015, 07:36 AM

Quote:

Originally Posted by Eeel

Hello Richard,

Thanks for your answer, i'm at work at the moment, will be able to test in ~10 hours

From what i know i have things written in /etc/mdadm.conf and same in the /boot/initrd-tree/etc/mdadm.conf, my mdamdm.conf contain my 4 ARRAY and my mail, after reading a thread yesterday about the md128 and md129 problem, i also added 4 more lines, one for each md* (i don't remember the exact syntaxe i have used but it was something like that):

Code:

ARRAY /dev/md0 devices=sda2,sdb2
ARRAY /dev/md1 devices=sda3,sdb3
....

While i was doing some various test yesterday i have re run several time the mkinirtd -F and lilo commands with a mdadm.conf file containing my 4 ARRAYS.

Just to be shure i'm not misunderstanding:

the initrd you are talking is the inird-tree or the initrd.gz file?

They are the same. initrd.gz is built from the contents of /boot/initrd-tree. Take a look at /boot/init-tree/etc/mdadm.conf; it should match the one you have on the root file system or not exist at all for this stuff to work.

FWIW, my /etc/mdadm.conf looks like...

Code:

ARRAY /dev/md/2  metadata=1.2 UUID=77c6da43:3f4dc424:53420c4e:5806e34c name=darkstar:2
ARRAY /dev/md/0  metadata=1.2 UUID=c7868644:607a78cc:c2458ba0:0bd9416b name=darkstar:0 spares=1
ARRAY /dev/md/1  metadata=1.2 UUID=b5f4c53c:27ea290e:b00a2908:b621ba08 name=darkstar:1

Eeel · 01-26-2015, 08:38 AM

Thanks Richard, i will soon be able to do further test and give some feedback

Eeel · 01-26-2015, 02:47 PM

This what i just tested:

- Boot the slackware install USB key
- re run SETUP to map partitions then EXIT
- chroot /mnt/
- Check my /etc/mdadm.conf (see attached screenshot), my ARRAYS are there, note that the ARRAY /dev/md3 have a different syntax doesn't seems to be a problem my Slackware 14.0 box have the same
- Check if i have same mdadm.conf in /boot/initrd-tree/etc Ok
- Re run Mkinitrd -F and lilo command, see attached screenshot, there is error here.

no luck i have the same Kernel panic

j_v · 01-26-2015, 03:35 PM

One idea I have about this issue is that I've had issues with the naming of the raid drives to be off if the mdadm.conf wasn't included in the initrd.

Eeel · 01-27-2015, 02:21 AM

Quote:

Originally Posted by j_v

One idea I have about this issue is that I've had issues with the naming of the raid drives to be off if the mdadm.conf wasn't included in the initrd.

can you explain a little bit more your idea, i'm stuck, i have installed yesterday from scratch but no luck.

Richard Cranium · 01-27-2015, 03:23 AM

Quote:

Originally Posted by Eeel

This what i just tested:

- Boot the slackware install USB key
- re run SETUP to map partitions then EXIT
- chroot /mnt/
- Check my /etc/mdadm.conf (see attached screenshot), my ARRAYS are there, note that the ARRAY /dev/md3 have a different syntax doesn't seems to be a problem my Slackware 14.0 box have the same
- Check if i have same mdadm.conf in /boot/initrd-tree/etc Ok
- Re run Mkinitrd -F and lilo command, see attached screenshot, there is error here.

no luck i have the same Kernel panic

Before you chroot /mnt, you should also do...

Code:

mount --bind /proc /mnt/proc
mount --bind /sys /mnt/sys
mount --bind /run /mnt/run
mount --bind /dev /mnt/dev

How about renaming /boot/initrd-tree/etc/mdadm.conf to /boot/initrd-tree/etc/mdadm-nouse.conf, re-run the mkinitrd command with no parameters, and then lilo. Reboot.

The initrd init script should be able to find your raid arrays by itself. Give it a shot.

Labinnah · 01-27-2015, 06:19 AM

Use UUID instead of /dev/mdX in /etc/fstab. You can use it also in /etc/lilo.conf.

I found, that some kernels (for sure 3.14) have problem with RAID naming. I have /dev/md127 and /dev/md126 despite they are named differently in my /etc/mdadm.conf.

Eeel · 01-27-2015, 03:36 PM

Hi all,

Good news it work, thank you all for suggestion and help, atm i don't really know where was the problem, i will check tomorrow to re create the error.

j_v · 01-27-2015, 03:54 PM

Glad to hear you have success. Hope you discover the origin of the problem, as I think this would be a big help for others who boot to raid drives.

kikinovak · 01-27-2015, 05:50 PM

Quote:

Originally Posted by Eeel

Hi all,

Good news it work, thank you all for suggestion and help, atm i don't really know where was the problem, i will check tomorrow to re create the error.

I do know. Two things.

1. Wait until your disks are in sync. Keep an eye on them like this:

Code:

# watch cat /proc/mdstat

2. You have to invoke mdadm -E -s > /etc/mdadm.conf before building your initrd, not after.

3. Extra hint. You can speed up the sync process significantly with this little trick:

Code:

# echo 50000 > /proc/sys/dev/raid/speed_limit_min

Whole process for RAID 1 and RAID 5 is documented here:

http://www.microlinux.fr/slackware/L...rver-HOWTO.txt

Cheers,

Niki

Richard Cranium · 01-27-2015, 08:36 PM

Quote:

Originally Posted by kikinovak

2. You have to invoke mdadm -E -s > /etc/mdadm.conf before building your initrd, not after.

Or remove mdadm.conf from your initrd. The init script in the initrd will make a new one at boot.

Eeel · 01-28-2015, 12:37 PM

I haven't been able to reproduce the error today, i started a complete installation except that i didn't remove all partitions with cfdisk before installation. For me the only difference in the installation of yesterday that finally work was the two things Kikinova point:

1. Wait until your disks are in sync. Keep an eye on them like this:

Code:

watch cat /proc/mdstat

2. You have to invoke mdadm -E -s > /etc/mdadm.conf before building your initrd, not after.

Many thanks all for help. And as someone suggest in this forum that i don't remember the name i have made a donation equal to the price of the Slackware 14.1 DvD since i'm now using it

For anyone interested this is exactly the steps that i have followed.

Run Slackware 14.1 64 bit install from USB, partitioning with cfdisk:

Code:

/boot	100Mo	sda1	DA	RAID 1	md3	
/	25Go	sda2	DA	RAID 5	md0
/home	50G	sda3	DA	RAID 5	md1
swap	4Go	sda4	DA	RAID 1	md2
free	~100mo

partition type DA

Duplicate partition to second 80Go drive with sfdisk:

Code:

sfdisk -d /dev/sda | sfdisk /dev/sdb

Stop eventually old raid array from previous install with mdadm:

- see wich "mdx" array may be up from a previous install

Code:

cat /proc/mdstat

- stop old "mdx" array with mdadm:

Code:

mdadm --stop /dev/mdx

Create new array with mdadm:

Code:

/:	mdadm --create /dev/md0 --level 5 --raid-devices 2 /dev/sda2 /dev/sdb2 --metadata=1.2
/home:	mdadm --create /dev/md1 --level 5 --raid-devices 2 /dev/sda3 /dev/sdb3 --metadata=1.2
swap:	mdadm --create /dev/md2 --level 1 --raid-devices 2 /dev/sda4 /dev/sdb4 --metadata=1.2
/boot:	mdadm --create /dev/md3 --level 1 --raid-devices 2 /dev/sda1 /dev/sdb1 --metadata=0.90

Wait until your disks are in sync. Keep an eye on them like this (thx Kikinovak):

Code:

watch cat /proc/mdstat

Format swap partition with mkswap:

Code:

mkswap -f /dev/md2

Format other array with mkfs:

Code:

mkfs.ext4 /dev/md0
mkfs.ext4 /dev/md1
mkfs.ext4 /dev/md3

Install Slackware using "setup":

- Install LILO "simple" setup
- Additional append parameter: root=/dev/md0
- Install LILO in MBR
- EXIT setup

Switch to installation directory with chroot:

Code:

chroot /mnt/

Edit lilo.conf:

Code:

# Start LILO global section
raid-extra-boot = mbr-only
# Append any additional kernel parameters:
append="root=/dev/md0 vt.default_utf8=0"
boot = /dev/md3

# End LILO global section
# Linux bootable partition config begins
image = /boot/vmlinuz
  initrd = /boot/initrd.gz
  root = /dev/md0
  label = Linux
  read-only
# Linux bootable partition config ends

Declare email and Raid array in /etc/mdadm.conf:

Code:

echo MAILADDR my-mail@somewhere.com >> /etc/mdadm.conf
mdadm -Es >> /etc/mdadm.conf

Create mkinitrd.conf:

Code:

cp /etc/mkinitrd.conf.sample /etc/mkinitrd.conf

Edit mkinitrd.conf:

Code:

# mkinitrd.conf.sample
# See "man mkinitrd.conf" for details on the syntax of this file
#
SOURCE_TREE="/boot/initrd-tree"
#CLEAR_TREE="0"
OUTPUT_IMAGE="/boot/initrd.gz"
KERNEL_VERSION="$(uname -r)"
KEYMAP="us"
MODULE_LIST="ext4"
#LUKSDEV="/dev/sda2"
ROOTDEV="/dev/md0"
ROOTFS="ext4"
#RESUMEDEV="/dev/sda2"
RAID="1"
#LVM="1"
#UDEV="1"
#MODCONF="0"
#WAIT="1"

Create the initrd with mkinitrd:

Code:

mkinitrd -F

Run Lilo:

Code:

lilo

Exit chroot

Code:

exit

reboot computer