View Issue Details

IDProjectCategoryView StatusLast Update
0005970CentOS-6dracutpublic2013-02-12 09:48
ReporterLutz Vieweg 
PrioritynormalSeveritymajorReproducibilityalways
Status newResolutionopen 
Platformx86_64OSCentOSOS Version6.3
Product Version6.3 
Target VersionFixed in Version 
Summary0005970: Software RAID1 boot failure (Kernel Panic) on failed disk
DescriptionThe issue of bug report #5400 seems to be back with a vengeance - a fresh installation of CentOS 6.3 (from CentOS-6.3-x86_64-bin-DVD1.iso) on RAID1 partitions on two SATA drives (using an unencrypted /boot RAID1 partition on md0 = sda1,sdb1 plus one encrypted RAID1 partition on md1 = sda2,sdb2) fails to boot if the sda disk is removed from the chassis.

Symptoms are the same as were reported in #5400 (I could not find a method to re-open that report) - booting stalls after encryption phrase has been entered, no root device found, kernel panic.

Symptoms remain after yum update to dracut-004-284.el6_3.noarch and re-installation of kernel (despite that http://rhn.redhat.com/errata/RHBA-2012-0839.html reported this to be fixed from dracut-004-284 onwards).
Steps To ReproduceInstall CentOS 6.3 with encrypted root partition on md RAID 1, try to reboot with the first drive removed.
TagsNo tags attached.

Activities

Lutz Vieweg

Lutz Vieweg

2012-09-20 11:43

reporter   ~0015805

BTW: In a posting at https://www.centos.org/modules/newbb/viewtopic.php?topic_id=34988&viewmode=flat&order=ASC&start=20 a user writes:

 i had similar problem with 6.3
 what i did was :
 1 boot to rescue mode
 2 chroot /mnt/sysimage/
 3 grub-install /dev/md0
 then it works...

I have tried that, but it does not work for me - the rescue mode of the CentOS 6.3 install DVD does not succeed to mount /mnt/sysimage for unknown reasons, it just states "you have no linux partitions".
globalsi

globalsi

2012-10-01 11:59

reporter   ~0015863

I've got same pbm (post #3 : https://www.centos.org/modules/newbb/viewtopic.php?topic_id=39498&forum=55).

Steps :
 - Install from CentOS-6.3-x86_64-minimal
 - 2 partions : swap (8GB) & / (rest of the disk except at the end 3MB)
 - build the raid, chroot, dracut ...
=> boot with 2 disks : ok
boot with 1 disk :
md: bind<sda>
dracut warning : No root device "block:/dev/md_d0p1"
dracut warning : boot has failed. To debug this issue add "rdshell" to the kernel command line
dracut warning : Signal caught!
dracut warning : boot has failed. To debug this issue add "rdshell" to the kernel command line
Kernel panic - not syncing: Attempted to kill init!
(as http://bugs.centos.org/view.php?id=5400)

 - yum update change nothing.
globalsi

globalsi

2012-10-01 12:45

reporter   ~0015864

dropind to rdshell :
Dropping to debug shell.

sh: can't access tty; job control turned off
dracut:/# dmsetup ls --tree
No devices found
dracut:/# blkid -p
The low-level probing mode requires a device
dracut:/# dmesg|grep dracut
dracut: dracut-004-284.el6
dracut: rd_NO_LUKS: removing cryptoluks activation
dracut: rd_NO_LVM: removing LVM activation
dracut: Starting plymouth daemon
dracut Warning: No root device "block:/dev/md_d0p1" found
dracut:/# cat /proc/partitions
major minor #blocks name

   8 0 976762584 sda
dracut:/# cat /proc/mdstat
Personalities :
md0 : inactive sda1[1](S)
      976762496 blocks

unused devices: <none>
dracut:/# mdadm --run /dev/md_d0
md: raid1 personality registered for level 1
bio: create slab <bio-1> at 1
md/raid1:md_md0: active with 1 out of 2 mirrors
md_d0: detected capacity change from 0 to 1000204795904
mdadm: started /dev/md_d0
md_d0: p1 p2

dracut:/# exit
EXT4-fs (md_d0p1): mounted filesystem with ordered data mode. Opts:
dracut: Mounted root filesystem /dev/md_d0p1
dracut: Loading SELinux policy
type=1404 audit(1349094337.889:2): enforcing=1 old_enforcing=0 auid=4294967295 ses=4294967295
SELinux: 2048 avtab hash slots, 250818 rules.
SELinux: 2048 avtab hash slots, 250818 rules.
SELinux: 9 users, 12 roles, 3761 types, 187 bools, 1 sens, 1024 cats
SELinux: 81 classes, 250818 rules
SELinux: Completing initialization.
SELinux: Setting up existing superblocks.
SELinux: initialized (dev md_d0p1, type ext4), uses xattr
SELinux: initialized (dev tmpfs, type tmpfs), uses transition SIDs
....
type=1403 audit(1349094338.430:3): policy loaded auid=4294967295 ses=4294967295
dracut:
dracut: Switching root
udev : starting version 147
[root@machinename]# :

Seem to be ok after this.
I'll test and do feedback
Lutz Vieweg

Lutz Vieweg

2012-10-02 10:01

reporter   ~0015868

@globalsi: When you say "Seem to be ok after this.", you mean that the system is booting for once - but not that it would boot without manual intervention the next time it is restarted with one disk missing, right?

(At least I would not expect "mdadm --run /dev/md_d0" to have any permanent effect on the installation.)
globalsi

globalsi

2012-10-05 08:50

reporter   ~0015891

Exemple :
1/ build raid with sda & sdb. boots correct
2/ remove sda (crash) : boot impossible with only sdb
3/ rdshell & 'mdadm --run /dev/md_d0' => next boots (not only once) ok with only sdb
4/ add a new sda ; rebuild raid : next boots ok
BUT after
5/ crash sdb : boot impossible with only sda
6/ must do again rdshell & 'mdadm --run /dev/md_d0'

info : yesterday, there was a update of dracut ... maybe the bug is corrected.
stsimb

stsimb

2013-02-12 09:48

reporter   ~0016481

I manually appended "rd_retry=20" in the kernel line of /boot/grub/menu.lst and it fixed the problem.

Issue History

Date Modified Username Field Change
2012-09-20 11:39 Lutz Vieweg New Issue
2012-09-20 11:43 Lutz Vieweg Note Added: 0015805
2012-10-01 11:59 globalsi Note Added: 0015863
2012-10-01 12:45 globalsi Note Added: 0015864
2012-10-02 10:01 Lutz Vieweg Note Added: 0015868
2012-10-05 08:50 globalsi Note Added: 0015891
2013-02-12 09:48 stsimb Note Added: 0016481