FAQ Search Today's Posts Mark Forums Read
» Video Reviews

» Linux Archive

Linux-archive is a website aiming to archive linux email lists and to make them easily accessible for linux users/developers.


» Sponsor

» Partners

» Sponsor

Go Back   Linux Archive > Gentoo > Gentoo User

 
 
LinkBack Thread Tools
 
Old 04-04-2012, 01:09 AM
walt
 
Default "lvm failed to start"

This is an ~amd64 machine, up to date as of today. The strange thing
is that lvm did *not* fail to start -- it's working perfectly.

Now, being an Incorrigible Old Fart(TM) I'm still using openrc, and
who knows what evil lurks in that paleolithic package?

Anyone else getting this (false) alarm during boot?
 
Old 04-04-2012, 02:51 AM
Dale
 
Default "lvm failed to start"

walt wrote:
> This is an ~amd64 machine, up to date as of today. The strange thing
> is that lvm did *not* fail to start -- it's working perfectly.
>
> Now, being an Incorrigible Old Fart(TM) I'm still using openrc, and
> who knows what evil lurks in that paleolithic package?
>
> Anyone else getting this (false) alarm during boot?
>
>

Well, for me its not a false alarm, sort of. Mine fails to start and
then works later on or after I restart lvm, sort of iffy here. This is
what mine looks like:

Setting up the Logical Volume Manager . . .
File-based locking initialisation failed
*Failed to setup the LVM [!!]
* Error: lvm failed to start

After all that, it's downhill sort of. Me, I get that with or without a
init thingy. I'm on Gentoo kernel 3.2.11. I'm using lvm2-2.02.95-r1.

From what I found, it does this because / is mounted ro. It can't write
the locking file. I thought those were in /var but . . .

Also, If I restart lvm after it boots, it starts fine and all the lvm
stuff shows up. Maybe this one isn't just me. o_O

Dale

:-) :-)


--
I am only responsible for what I said ... Not for what you understood or
how you interpreted my words!

Miss the compile output? Hint:
EMERGE_DEFAULT_OPTS="--quiet-build=n"
 
Old 04-04-2012, 08:39 AM
Neil Bothwick
 
Default "lvm failed to start"

On Tue, 03 Apr 2012 18:09:26 -0700, walt wrote:

> This is an ~amd64 machine, up to date as of today. The strange thing
> is that lvm did *not* fail to start -- it's working perfectly.
>
> Now, being an Incorrigible Old Fart(TM) I'm still using openrc, and
> who knows what evil lurks in that paleolithic package?
>
> Anyone else getting this (false) alarm during boot?

It is not really a false alarm, LVM does fail here but tries again later
on and succeeds. https://bugs.gentoo.org/show_bug.cgi?id=409921 covers it
and the workaround in comment 2 does it for me.

Basically, LVM is trying to write its lockfile to a readonly filesystem.
I'm going to try mounting /run on tmpfs in my initramfs as that is where
it is mounted now, but apparently too late for the first attempt at
starting LVM.


--
Neil Bothwick

"If Micro built cars, the worlds population would be in decline"
 
Old 04-04-2012, 10:22 AM
 
Default "lvm failed to start"

walt <w41ter@gmail.com> wrote:

> This is an ~amd64 machine, up to date as of today. The strange thing
> is that lvm did *not* fail to start -- it's working perfectly.
>
> Now, being an Incorrigible Old Fart(TM) I'm still using openrc, and
> who knows what evil lurks in that paleolithic package?
>
> Anyone else getting this (false) alarm during boot?
>
I get lvm failed to start because it says file based locking does not
work because the root file system is read only, but when it gets to
default runlevel it tries again and starts it OK.]

Weird.


--
Your life is like a penny. You're going to lose it. The question is:
How do
you spend it?

John Covici
covici@ccs.covici.com
 
Old 04-04-2012, 04:58 PM
Dan Johansson
 
Default "lvm failed to start"

On Tuesday 03 April 2012 18.09:26 walt wrote:
> This is an ~amd64 machine, up to date as of today. The strange thing
> is that lvm did *not* fail to start -- it's working perfectly.
>
> Now, being an Incorrigible Old Fart(TM) I'm still using openrc, and
> who knows what evil lurks in that paleolithic package?
>
> Anyone else getting this (false) alarm during boot?
I have a similar thing on my ~x86, but the difference is that I know why I get it.
I have successfully be able to create an initramfs that does a vgscan, vgchange -a y and mounts /usr (which is on LVM).
But now I (naturally) I get "LVM failed to start" (and of cause "failed to mount /usr) when openrc processes the init-scripts.
So, my question is really - should I remove lvm from the init-process and put noauto in /etc/fstab for /usr?
Even if it is on my playground box I still want to be able to boot it.
When these two "issues" are solved I'm going to let my system update udev.

--
Dan Johansson, <http://www.dmj.nu>
************************************************** *
This message is printed on 100% recycled electrons!
************************************************** *
 
Old 04-04-2012, 11:36 PM
Neil Bothwick
 
Default "lvm failed to start"

On Wed, 04 Apr 2012 18:58:38 +0200, Dan Johansson wrote:

> I have a similar thing on my ~x86, but the difference is that I know
> why I get it. I have successfully be able to create an initramfs that
> does a vgscan, vgchange -a y and mounts /usr (which is on LVM). But now
> I (naturally) I get "LVM failed to start" (and of cause "failed to
> mount /usr) when openrc processes the init-scripts.

This is not down to your initramfs, see the previously linked bug.


--
Neil Bothwick

A closed mouth gathers no foot.
 
Old 04-05-2012, 03:59 PM
Dan Johansson
 
Default "lvm failed to start"

ectrons!
************************************************** *
On Thursday 05 April 2012 00.36:33 Neil Bothwick wrote:
> On Wed, 04 Apr 2012 18:58:38 +0200, Dan Johansson wrote:
>
> > I have a similar thing on my ~x86, but the difference is that I know
> > why I get it. I have successfully be able to create an initramfs that
> > does a vgscan, vgchange -a y and mounts /usr (which is on LVM). But now
> > I (naturally) I get "LVM failed to start" (and of cause "failed to
> > mount /usr) when openrc processes the init-scripts.
>
> This is not down to your initramfs, see the previously linked bug.

What do you mean by "This is not down to your initramfs"?
With the new "C:" concept of udev /usr needs to be mounted before /sbin/init is run and as I am using LVM, LVM needs to be started before /sbin/init as well - or have I missed something here?

--
Dan Johansson, <http://www.dmj.nu>
************************************************** *
This message is printed on 100% recycled el
 
Old 04-05-2012, 09:42 PM
Neil Bothwick
 
Default "lvm failed to start"

On Thu, 05 Apr 2012 17:59:42 +0200, Dan Johansson wrote:

> > > I have a similar thing on my ~x86, but the difference is that I know
> > > why I get it. I have successfully be able to create an initramfs
> > > that does a vgscan, vgchange -a y and mounts /usr (which is on
> > > LVM). But now I (naturally) I get "LVM failed to start" (and of
> > > cause "failed to mount /usr) when openrc processes the
> > > init-scripts.
> >
> > This is not down to your initramfs, see the previously linked bug.
>
> What do you mean by "This is not down to your initramfs"?
> With the new "C:" concept of udev /usr needs to be mounted
> before /sbin/init is run and as I am using LVM, LVM needs to be started
> before /sbin/init as well - or have I missed something here?

You are getting this error from openrc, which means the initramfs
has already done its stuff and passed control to init on the real root
partition. At this point /usr is already mounted, it is the openrc
startup of LVM that is failing, because it is trying to write a lockfile
to a read-only filesystem, which would be the case even if you were not
using an initramfs.

The problem is that it is trying to write to /var/lock, which is on / at
this point, rather than /run/lock, which is on a writeable tmpfs.


--
Neil Bothwick

Bus: (n.) a connector you plug money into, something like a slot machine.
 
Old 04-07-2012, 08:47 AM
Carlos Sura
 
Default "lvm failed to start"

On 5 April 2012 15:42, Neil Bothwick <neil@digimed.co.uk> wrote:

On Thu, 05 Apr 2012 17:59:42 +0200, Dan Johansson wrote:



> > > I have a similar thing on my ~x86, but the difference is that I know

> > > why I get it. I have successfully be able to create an initramfs

> > > that does a vgscan, vgchange -a y and mounts /usr (which is on

> > > LVM). But now I (naturally) I get "LVM failed to start" (and of

> > > cause "failed to mount /usr) when openrc processes the

> > > init-scripts.

> >

> > This is not down to your initramfs, see the previously linked bug.

>

> What do you mean by "This is not down to your initramfs"?

> With the new "C:" concept of udev /usr needs to be mounted

> before /sbin/init is run and as I am using LVM, LVM needs to be started

> before /sbin/init as well - or have I missed something here?



You are getting this error from openrc, which means the initramfs

has already done its stuff and passed control to init on the real root

partition. At this point /usr is already mounted, it is the openrc

startup of LVM that is failing, because it is trying to write a lockfile

to a read-only filesystem, which would be the case even if you were not

using an initramfs.



The problem is that it is trying to write to /var/lock, which is on / at

this point, rather than /run/lock, which is on a writeable tmpfs.





--

Neil Bothwick



Bus: (n.) a connector you plug money into, something like a slot machine.


I got this error, and saw a bugfile on Gentoo, just downgrade to your previous working version and you will get it working again.--
Carlos Sura.-
www.carlossura.com
 
Old 04-07-2012, 10:38 AM
Neil Bothwick
 
Default "lvm failed to start"

On Sat, 7 Apr 2012 02:47:41 -0600, Carlos Sura wrote:

> > The problem is that it is trying to write to /var/lock, which is on /
> > at this point, rather than /run/lock, which is on a writeable tmpfs.

> I got this error, and saw a bugfile on Gentoo, just downgrade to your
> previous working version and you will get it working again.

That bug report also showed the correct solution. There's no need to
downgrade, just change the locking_dir location from /var/lock
to /run/lock. Not only is this simpler and faster, it means another
update can't cause the same problem.


--
Neil Bothwick

When there's a will, I want to be in it.
 

Thread Tools




All times are GMT. The time now is 05:25 AM.

VBulletin, Copyright ©2000 - 2014, Jelsoft Enterprises Ltd.
Content Relevant URLs by vBSEO ©2007, Crawlability, Inc.
Copyright 2007 - 2008, www.linux-archive.org