[linux-lvm] vgscan -- invalid i/o protocol version

[linux-lvm] vgscan -- invalid i/o protocol version

[James Pattinson]

Anyone seen this error msg before?

I'm running RH6 with kernel 2.2.10 and the LVM patches applied. I
originially tried to follow the HOWTO but on the first command i got:

pvcreate -- ERROR: "/etc/lvmtab" doesn't exist; please run vgscan

so i ran vgscan and got the message in the subject line. It's possible I'm
being really thick and I don't really know what I'm doing. Do I need the
devfs patches too?

I'm new to lvm-on-linux but I love it on HP-UX.

Cheers!

James

Re: [linux-lvm] vgscan -- invalid i/o protocol version

[Rolf Jakob]

> Anyone seen this error msg before?
Yes :-)

> I'm running RH6 with kernel 2.2.10 and the LVM patches applied. I
> originially tried to follow the HOWTO but on the first command i got:
> 
> pvcreate -- ERROR: "/etc/lvmtab" doesn't exist; please run vgscan
> 
> so i ran vgscan and got the message in the subject line. It's possible I'm

Did you build LVM as a module ? Is the module loaded ? I get the same
message when I run vgscan without modprobe lvm.o.

> being really thick and I don't really know what I'm doing. Do I need the
> devfs patches too?

Not that I know of. I never tried it without devfs, but that has other
reasons (having to remove the first SCSI disk is a strong argument for
devfs :-)

Rolf
-- 
Rolf Jakob at home (rjakob@duffy1.franken.de)
WWW : http://www.franken.de/users/duffy1/rjakob (KDE-Utils and CCS)

Re: [linux-lvm] vgscan -- invalid i/o protocol version

[Heinz Mauelshagen]

> 
> Anyone seen this error msg before?
> 
> I'm running RH6 with kernel 2.2.10 and the LVM patches applied. I
> originially tried to follow the HOWTO but on the first command i got:
> 
> pvcreate -- ERROR: "/etc/lvmtab" doesn't exist; please run vgscan

Have you patched lvm_0.6-1503199a.gz into the 0.6 LVM source tree?

> 
> so i ran vgscan and got the message in the subject line. It's possible I'm
> being really thick and I don't really know what I'm doing. Do I need the
> devfs patches too?
> 
> I'm new to lvm-on-linux but I love it on HP-UX.
> 

Regards,
Heinz

--

=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-

Systemmanagement C/S                             Deutsche Telekom AG
                                                 Entwicklungszentrum Darmstadt
Heinz Mauelshagen                                Otto-Roehm-Strasse 71c
Senior Systems Engineer                          Postfach 10 05 41
                                                 64205 Darmstadt
mge@ez-darmstadt.telekom.de                      Germany
                                                 +49 6151 886-425
                                                          FAX-386
=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-

Re: [linux-lvm] vgscan -- invalid i/o protocol version

[Michael Marxmeier]

> so i ran vgscan and got the message in the subject line. It's possible I'm
> being really thick and I don't really know what I'm doing. Do I need the
> devfs patches too?

AFAIR this error message has appeared previously when the /dev/lvm
device file was missing.

$ ls -l /dev/lvm
crw-r-----   1 root     root     109,   0 Mar 31 05:13 /dev/lvm

If it's not created, try a mknod /dev/lvm c 109 0

> I'm new to lvm-on-linux but I love it on HP-UX.

You're welcome.


Hope this helps
Michael

--
Michael Marxmeier           Marxmeier Software GmbH
E-Mail: mike@msede.com      Besenbruckstrasse 9
Voice : +49 202 2431440     42285 Wuppertal, Germany
Fax   : +49 202 2431420     http://www.msede.com/

Re: [linux-lvm] vgscan -- invalid i/o protocol version

[James Pattinson]

That's exactly the problem. In fact I found out that was the problem after
I sent the mail by looking at the output from `strace vgscan' but I
couldn't find the require major and minor numbers for the device file
anywhere in the docs.

I'm a bit stumped for now though as I don't have any spare partitions
(yet) - I tried to be cunning with loopback devices but it doesn't seem to
work:

[root@fuse /root]# dd if=/dev/zero bs=1024k count=32 of=loop1
[root@fuse /root]# dd if=/dev/zero bs=1024k count=32 of=loop2
[root@fuse /root]# dd if=/dev/zero bs=1024k count=32 of=loop3
[root@fuse /root]# losetup /dev/loop1 loop1
[root@fuse /root]# losetup /dev/loop2 loop2
[root@fuse /root]# losetup /dev/loop3 loop3
[root@fuse /root]# pvcreate /dev/loop1 /dev/loop2 /dev/loop3
pvcreate -- invalid physical volume name "/dev/loop1"
pvcreate -- invalid physical volume name "/dev/loop2"
pvcreate -- invalid physical volume name "/dev/loop3"

Is there any way I can bodge it like this to try it out?

One more thing - I am a big fan of the raid patches that allow you to
autodetect raid partitions on boot - currently my machine has it's root
partition striped across two disks. What is the striping like on LVM? Is
it feasible to boot from an LVM?

Cheers,

James

On Wed, 30 Jun 1999, Michael Marxmeier wrote:

> > so i ran vgscan and got the message in the subject line. It's possible I'm
> > being really thick and I don't really know what I'm doing. Do I need the
> > devfs patches too?
> 
> AFAIR this error message has appeared previously when the /dev/lvm
> device file was missing.

Re: [linux-lvm] vgscan -- invalid i/o protocol version

[Heinz Mauelshagen]

> 
> That's exactly the problem. In fact I found out that was the problem after
> I sent the mail by looking at the output from `strace vgscan' but I
> couldn't find the require major and minor numbers for the device file
> anywhere in the docs.

The patch lvm_0.6-patch-15031999a.gz (sorry for my typo in my previous
mail :*( ) is needed to take care of automatic creation of the /dev/lvm special.

> 
> I'm a bit stumped for now though as I don't have any spare partitions
> (yet) - I tried to be cunning with loopback devices but it doesn't seem to
> work:
> 
> [root@fuse /root]# dd if=/dev/zero bs=1024k count=32 of=loop1
> [root@fuse /root]# dd if=/dev/zero bs=1024k count=32 of=loop2
> [root@fuse /root]# dd if=/dev/zero bs=1024k count=32 of=loop3
> [root@fuse /root]# losetup /dev/loop1 loop1
> [root@fuse /root]# losetup /dev/loop2 loop2
> [root@fuse /root]# losetup /dev/loop3 loop3
> [root@fuse /root]# pvcreate /dev/loop1 /dev/loop2 /dev/loop3
> pvcreate -- invalid physical volume name "/dev/loop1"
> pvcreate -- invalid physical volume name "/dev/loop2"
> pvcreate -- invalid physical volume name "/dev/loop3"
> 
> Is there any way I can bodge it like this to try it out?

Sorry, no loop device support with 0.6.

It will be in 0.7 as an evaluation support for people who don't
have lots of disk partitions to play with.

I should better throw out 0.7 soon 8*)))

> 
> One more thing - I am a big fan of the raid patches that allow you to
> autodetect raid partitions on boot - currently my machine has it's root
> partition striped across two disks. What is the striping like on LVM? Is
> it feasible to boot from an LVM?
> 

That's one issue i'm working on with 0.7 right know.
After i asked for input of people who claimed already to be able to
have root in a LV but never got a return i wanted to bring support with 0.7.

Maybe i better release 0.7 and give that support later?!

It bassically works like this:

LILO -> INITRD (in very small partition) -> activate volume groups ->
change root to root LV -> copy /etc/lcmtab* to real LV root -> start as usual

The INITRD only contains some needed executables, shared libraries, dirs
and /linuxrc.


Regards,
Heinz


> Cheers,
> 
> James
> 
> On Wed, 30 Jun 1999, Michael Marxmeier wrote:
> 
> > > so i ran vgscan and got the message in the subject line. It's possible I'm
> > > being really thick and I don't really know what I'm doing. Do I need the
> > > devfs patches too?
> > 
> > AFAIR this error message has appeared previously when the /dev/lvm
> > device file was missing.
> 
> 
> 


--

=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-

Systemmanagement C/S                             Deutsche Telekom AG
                                                 Entwicklungszentrum Darmstadt
Heinz Mauelshagen                                Otto-Roehm-Strasse 71c
Senior Systems Engineer                          Postfach 10 05 41
                                                 64205 Darmstadt
mge@ez-darmstadt.telekom.de                      Germany
                                                 +49 6151 886-425
                                                          FAX-386
=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-

Re: [linux-lvm] vgscan -- invalid i/o protocol version

[Heinz Mauelshagen]

Hi Dave.

> 
> Can you send me the contents of the INITRD partition? Is this in the LVM
> documentation?
> 

No it's not in the 0.6 documentation.

Here are lvmcreate_initrd which creates an initial ram disk image
and an example /boot/lilo.conf to boot from a boot partition and change
root to a LV based root filesystem.

You have to have a /boot fs on a bios adressable disk (saw in /dev/hda5).
16 MB is sufficient, if you want to make this boot partition standalone
bootable and not only want to use it to store an INITRD image.

Copy vmlinux, System.map, lilo.conf, boot.b and chain.b to that boot fs.
Make /etc/lilo.conf a symlink to /boot/lilo.conf.

Set up LVs for /, /usr etc. and copy existing partition bases filesystem
contents to the temporary mounted LV based new filesystems.

The lvmcreate_initrd script below assumes /boot as your mountpoint of
the boot filesystem and that you load LVM as a module.
Don't forget to configure RAM disk and Initial RAM disk support in the kernel
and to copy any files to /boot you need for standalone emergency boot of
your boot fs (see end of the lilo.conf example).

Regards and good luck,
Heinz



A lilo.conf example assuming you would set up a vg00 as the "boot" VG,
your boot partition is on /dev/hda5 and your LV based root is on /dev/vg00/root:

#########################################################################
# LILO Configuration

# Start LILO global Section
boot = /dev/hda
install = /boot/boot.b	# must be redable by bios
read-only
prompt
timeout = 100
vga = normal    # force sane state
message = /boot/message
map = /boot/map	# must be readable by bios
# End LILO global Section

#
image = /boot/vmlinuz
  initrd = /boot/initrd.gz
  root = /dev/vg00/root
  alias = 1
  label = linuxinitrd

#
image = /boot/vmlinuz.old
  initrd = /boot/initrd.old.gz
  root = /dev/vg00/root
  alias = 2
  label = lnxoldinitrdold

#
image = /boot/vmlinuz
  append = "init=/bin/sh"
  root = /dev/hda5
  alias = 3
  label = linuxhda5
# End ###################################################################




#!/bin/sh
#
#  Copyright (C) 1997 - 1999  Heinz Mauelshagen, Germany
# 
#  June 1999
# 
#  LVM is free software; you can redistribute it and/or modify
#  it under the terms of the GNU General Public License as published by
#  the Free Software Foundation; either version 2, or (at your option)
#  any later version.
#  
#  LVM is distributed in the hope that it will be useful,
#  but WITHOUT ANY WARRANTY; without even the implied warranty of
#  MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
#  GNU General Public License for more details.
#  
#  You should have received a copy of the GNU General Public License
#  along with LVM; see the file COPYING.  If not, write to
#  the Free Software Foundation, 59 Temple Place - Suite 330,
#  Boston, MA 02111-1307, USA. 
# 

#
# make it devfs clean
#

cmd=`basename $0`
DEVRAM=/dev/ram1
INITRD=/boot/initrd.gz
INITRDSIZE=8192
INITRDFILES="/sbin/modprobe /sbin/vgchange /sbin/vgscan /bin/bash /bin/sh /bin/rm /sbin/insmod /dev/* /lib/modules/`uname -r`/modules.dep /lib/modules/`uname -r`/block/lvm.o"
TMP=/tmp/mnt.$$

trap "
  cd /
  umount $DEVRAM
  freeramdisk $DEVRAM
  echo -e 'Bye bye...\n'
" 1 2 3 15


function create_linuxrc {
   echo "#!/bin/sh
   /sbin/insmod lvm
   /sbin/vgscan
   /sbin/vgchange -a y" > linuxrc
   chmod 555 linuxrc
}

#
# Main
#
echo -e "\nLogical Volume Manager 0.7 by Heinz Mauelshagen  01/07/1999\n"
echo -e "$cmd -- this script creates a LVM initial ram disk in $INITRD\n"

echo "$cmd -- making ram filesystem"
mke2fs -m0 -N 4096 $DEVRAM $INITRDSIZE >/dev/null 2>&1
if [ $? -ne 0 ]; then
   echo -e "$cmd -- ERROR making ram disk filesystem\n"
   exit 1
fi

mkdir -p $TMP/ram >/dev/null 2>&1
if [ $? -ne 0 ]; then
   echo -e "$cmd -- ERROR making $TMP/ram\n"
   exit 1
fi

echo "$cmd -- mounting ram filesystem"
mount -t ext2 $DEVRAM $TMP/ram >/dev/null 2>&1
if [ $? -ne 0 ]; then
   echo -e "$cmd -- ERROR mounting $DEVRAM on $TMP/ram\n"
   exit 1
fi

cd $TMP/ram
[ ! -d etc ] && mkdir etc

#
# create new conf.modules to avoid kmod complaining about nonexsisting modules.
#
echo "$cmd -- creating etc/conf.modules"
i=0
while [ $i -lt 256 ]; do
   echo "alias	block-major-$i	off"
   echo "alias	char-major-$i	off"
   let i=i+1
done > etc/conf.modules

# to ensure, that modprobe doesn complain about timestamps
echo "$cmd -- creating new modules.dep"
depmod -a >/dev/null 2>&1
if [ $? -ne 0 ]; then
   echo -e "$cmd -- ERROR depmod\n"
   exit 1
fi

# copy necessary files to ram disk
echo "$cmd -- copying files to ram disk"
find $INITRDFILES|cpio -pdm $TMP/ram >/dev/null 2>&1
if [ $? -ne 0 ]; then
   echo -e "$cmd -- ERROR cpio to ram disk\n"
   exit 1
fi

# figure out which actual shared libraries we need in our initrd
echo "$cmd -- figuring out shared libraries"
SHLIBS=`ldd sbin/vg* bin/sh|awk '{if(/=>/){print $3}}'|sort -u 2>/dev/null`
if [ $? -ne 0 ]; then
   echo -e "$cmd -- ERROR figuring out needed shared libraries\n"
   exit 1
fi

echo "$cmd -- copying shared libraries to ram disk"
find $SHLIBS|cpio -Lpdm $TMP/ram >/dev/null 2>&1
if [ $? -ne 0 ]; then
   echo -e "$cmd -- ERROR copying needed shared libraries to ram disk\n"
   exit 1
fi

echo "$cmd -- creating linuxrc"
create_linuxrc >/dev/null 2>&1
if [ $? -ne 0 ]; then
   echo -e "$cmd -- ERROR creating linuxrc\n"
   exit 1
fi

cd /
echo "$cmd -- ummounting ram disk"
umount $DEVRAM
if [ $? -ne 0 ]; then
   echo -e "$cmd -- ERROR umounting $DEVRAM\n"
   exit 1
fi

echo "$cmd -- creating compressed initrd in $INITRD"
gzip -9 < $DEVRAM > $INITRD 2>/dev/null
if [ $? -ne 0 ]; then
   echo -e "$cmd -- ERROR creating $INITRD\n"
   exit 1
fi

freeramdisk $DEVRAM

--

=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-

Systemmanagement C/S                             Deutsche Telekom AG
                                                 Entwicklungszentrum Darmstadt
Heinz Mauelshagen                                Otto-Roehm-Strasse 71c
Senior Systems Engineer                          Postfach 10 05 41
                                                 64205 Darmstadt
mge@ez-darmstadt.telekom.de                      Germany
                                                 +49 6151 886-425
                                                          FAX-386
=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-