Installation on RAID or LVM fails to boot

[quequotion]

Booting from RAID is often broken after upgrades or fresh installations of Ubuntu.

initramfs is not activating RAID devices.

Steps to reproduce:
1. Upgrade or Install Ubuntu on a RAID
2. Reboot
3. Observe failure to find root drive and drop to busybox

To work around this, when given the busybox prompt enter:

dmraid -ay

Verify that all of your raid devices are initialized, then enter:

exit

Booting should then continue.

See comment #2 for a persistent (rebootable) workaround.

[Launchpad Janitor (janitor)]

Status changed to 'Confirmed' because the bug affects multiple users.

[Thomas Schiex (thomas-schiex)]

To avoid having to type this at each boot, you may edit

  /usr/share/initramfs-tools/scripts/local-top/dmraid

and add a

dmraid -ay

there

Beyond this, I had another problem:

GRUB2 seems to lack the "part_msdos" preload module. I had to

grub-install --modules="part_msdos" --boot-directory=... /dev/mapper/...

after booting on a liveCD and chrooting to the original partition to avoid the grub rescue screen.

[quequotion (quequotion)]

>>Thomas

Done.

Even more deja vu!

I think we used the same work around back in Jaunty.

For more detail, here's that script:

---SCRIPT BEGIN---
#!/bin/sh

# local-top script for dmraid.

PREREQS=""
prereqs()
{
 echo $PREREQS
}

case $1 in
# get pre-requisites
prereqs)
 prereqs
 exit 0
 ;;
esac

# Activate any dmraid arrays that were not identified by udev and vol_id.

#if devices=$(dmraid -r -c); then
# for dev in $devices; do
# dmraid-activate $dev
# done
#fi

dmraid -ay
---SCRIPT END---

When udev fails, the "if" clause I've commented out should activate the raid drives, but it doesn't...

[quequotion (quequotion)]

Is this going to be fixed?

[quequotion (quequotion)]

Added udev because it should have these arrays activated before the bad script has a chance to fail.

[Christian Odhner (strask)]

I can confirm this bug is present on a fresh install (not upgrade) of 12.04 beta 2.

[Kevin Johnson (poettone)]

I'm having this activity happen to me when installing on a Windows 7 64bit machine using wubi.exe on an IBM T400. I've ran wubi.exe on other Windows 7 64 machines without issue, so I'm not sure, but I will try some of the fixes above and see if that helps.

[Ir0nsh007er (ir0n)]

Confirm bug still on. Whith fresh install not upgrade 12.10

[Alexandre Paradis (ag-ap)]

I wanted to add additional information to this bug.

The same problem exists for LVM VG groups not being activated before initramfs-tools tries to mount an LVM root partition.

Specifically my issue occurred after reboot, after the update to Precise (Ubuntu Server). I got the "ALERT! /dev/mapper/my-root does not exist Dropping to a shell!" error and sent back to initramfs prompt.

To fix the problem, I had to manually edit /usr/share/initramfs-tools/scripts/init-premount/lvm2 , and add the following line:
vgchange -ay

[fransvn (frans-vanniekerk)]

I am experiencing this exact problem on 13.04 after upgrading from 12.10. The work around described above allows me to boot into Ubuntu

[Hadrien Mary (hadim)]

Got the same problem here. I need to execute dmraid -ay in the busybox when boot fails. So I added dmraid -ay at the end of /usr/share/initramfs-tools/scripts/local-top/dmraid file then sudo update-initframfs -c -k all && sudo update-grub. This trick fix the problem but I will need to do it after each kernel upgrade I guess... Someone could say if the bug will be fixed or not ?

(I don't know if it's related but I am using Ubuntu Gnome 13.04 with staging ppa which include systemd package...)

I can provide more info if you need.

Thank you for your help

[Hadrien Mary (hadim)]

Note that I added dmraid to affected package because apt-file search /usr/share/initramfs-tools/scripts/local-top/dmraid returns dmraid

[quequotion (quequotion)]

>>HadiM

Thanks for the update! What version of ubuntu was that?

[Hadrien Mary (hadim)]

Last version of Ubuntu 13.04 with both gnome3 and gnome3-ppa staging. Do
you need more details ?

--
HadiM

On Fri, Jun 7, 2013 at 6:49 AM, quequotion <email address hidden>wrote:

> >>HadiM
>
> Thanks for the update! What version of ubuntu was that?
>
> --
> You received this bug notification because you are subscribed to the bug
> report.
> https://bugs.launchpad.net/bugs/931929
>
> Title:
> root on RAID:0 fails to boot
>
> Status in dmraid:
> New
> Status in tools for generating an initramfs:
> Confirmed
> Status in udev - /dev/ management daemon:
> New
> Status in “initramfs-tools” package in Ubuntu:
> Confirmed
>
> Bug description:
> Once again booting from RAID is broken.
>
> There were similar bugs in Jaunty, Karmic, and Lucid.
> I don't recall encountering this error in Natty.
> It was definetly not present in Oneiric.
>
> Steps to reproduce:
> 1. Upgrade to Precise (fresh install not tested, but likely has the same
> problem)
> 2. Reboot
> 3. Observe failure to find root and drop to busybox
>
> initramfs is not loading dmraid and therefore fails to boot.
>
> To work around this, when given the busybox prompt enter:
>
> dmraid -ay
>
> and verify that your raid devices are initialized, thn enter:
>
> exit
>
> and booting should continue.
>
> To manage notifications about this bug go to:
> https://bugs.launchpad.net/dmraid/+bug/931929/+subscriptions
>

[quequotion (quequotion)]

Does anyone know why this code:

if devices=$(dmraid -r -c); then
 for dev in $devices; do
  dmraid-activate $dev
 done
fi

is being used when

dmraid -ay

actually works (correctly, quickly, reliably)?

[quequotion (quequotion)]

I've updated the description, since the problem isn't limited to upgrades to precise.

[GLehnhoff (glehnhoff)]

I got this problem suddenly today with 14.04 on SDD/brtfs, shortly after I switched to NVIDIA driver, as the standard one crashes (known bug). As my disk is not an array, dmraid -ay does not help (command does not exist). Disk mounts perfectly in another Ubuntu. Current Ubuntu does not start anymore.

[cracket (cracket)]

Here same problem 3 years after releasing new version of Ubuntu. Today upgrading from 10.04 LTS to 12.04. Apart broken grub2 update, which I already worked around - system does not boot any more. System configuration: / at REVO Drive PCIe SSD, visible as fake raid (2x 115 GB) sdb & sdc, x86_64. Before upgrade in 10.04 it was recognised during boot as /dev/mapper/sil_biajabcecafc device, and root filesystem was mounted on it.
Unfortunately after update to 12.04 LTS , and fixing broken grub2 "grub_xputs" failure from pendrive my system is still unusable, probably due to invalid/broken initrd/udevd:
ALERT /dev/disk/by-uuid/3a11ac88-b86a-4500-a94a-737adf8b989e does not exist.
Dropping to a shell!
And your hints are not very helpful, because my new initrd upgraded by fresh ubuntu installation contains only dmsetup and mdadm binaries. I would add dmraid happily to initrd if my first boot was successful!

So problem #1 - how to create manually required device /dev/mapper/sil_biajabcecafc from initrd
Problem #2 - how to update broken installation 12.04 so it would have proper fakeraid autodetection in initrd
Problem #3 - how to inform people willing to install/update to 12.04 that they will have serious problems?