[linux-lvm] one PV crashed

[linux-lvm] one PV crashed

[Turbo Fredriksson]

I have (had) 6 PV's in my VG. The last one crashed (the whole disk
seems to have broken totaly).

It don't seem possible to restore the information (not without spending
$$$ to a data recovery company), so how do I get the data stored on the
OTHER disks?


[papadoc.root]# pvscan -u
pvscan -- reading all physical volumes (this may take a while...)
pvscan -- inactive PV "/dev/sdb1 " with UUID "ofZCeT-O16H-Df0h-4KoB-xQhm-lc24-67dZvk" of VG "vg1" [5.87 GB / 0 free]
pvscan -- inactive PV "/dev/sdc1 " with UUID "xN0MF7-xKU0-Uk3L-g8Pw-0YfH-fS3V-zgFA66" of VG "vg1" [1004.00 MB / 0 free]
pvscan -- inactive PV "/dev/sdd10" with UUID "5V0XNB-gA87-OU66-1iza-cvsz-xQ2q-6DEbyH" of VG "vg1" [2.37 GB / 0 free]
pvscan -- inactive PV "/dev/sde1 " with UUID "Bdm7PX-50Dq-hC6R-yXmv-DDq1-n6Nx-4ohaKK" of VG "vg1" [8.50 GB / 0 free]
pvscan -- inactive PV "/dev/sdf1 " with UUID "QnqfyA-fOTB-J1Pp-4be5-t6QU-bGjS-i2BKf2" of VG "vg1" [17.09 GB / 0 free]
pvscan -- total: 5 [34.84 GB] / in use: 5 [34.84 GB] / in no VG: 0 [0]
[papadoc.root]# /etc/init.d/lvm start
Setting up LVM Volume Groups...
vgchange -- ERROR: physical volume "/dev/sdg1" of volume group "vg1" is offline
vgchange -- run vgscan
[papadoc.root]# vgscan 
vgscan -- reading all physical volumes (this may take a while...)
vgscan -- "/etc/lvmtab" and "/etc/lvmtab.d" successfully created
vgscan -- WARNING: This program does not do a VGDA backup of your volume group
[papadoc.root]# pvscan -u
pvscan -- reading all physical volumes (this may take a while...)
pvscan -- inactive PV "/dev/sdb1 " with UUID "ofZCeT-O16H-Df0h-4KoB-xQhm-lc24-67dZvk"  is associated to an unknown VG (run vgscan)
pvscan -- inactive PV "/dev/sdc1 " with UUID "xN0MF7-xKU0-Uk3L-g8Pw-0YfH-fS3V-zgFA66"  is associated to an unknown VG (run vgscan)
pvscan -- inactive PV "/dev/sdd10" with UUID "5V0XNB-gA87-OU66-1iza-cvsz-xQ2q-6DEbyH"  is associated to an unknown VG (run vgscan)
pvscan -- inactive PV "/dev/sde1 " with UUID "Bdm7PX-50Dq-hC6R-yXmv-DDq1-n6Nx-4ohaKK"  is associated to an unknown VG (run vgscan)
pvscan -- inactive PV "/dev/sdf1 " with UUID "QnqfyA-fOTB-J1Pp-4be5-t6QU-bGjS-i2BKf2"  is associated to an unknown VG (run vgscan)
pvscan -- total: 5 [34.84 GB] / in use: 5 [34.84 GB] / in no VG: 0 [0]


This was the info the sg1 disk had:
PV#     PV Name    VG Name PV UUID                                 PV Size
6       /dev/sdg1  vg1     7x1CdK-50u6-NAmJ-6bHY-Ho09-U3hC-K7c18w  15.55 GB 0 free

World Trade Center FSF spy Nazi terrorist supercomputer cryptographic
toluene [Hello to all my fans in domestic surveillance] assassination
counter-intelligence Ft. Bragg subway Rule Psix security
[See http://www.aclu.org/echelonwatch/index.html for more about this]

Re: [linux-lvm] one PV crashed

[Heinz J . Mauelshagen]

On Fri, Feb 01, 2002 at 11:01:22AM +0100, Turbo Fredriksson wrote:
> I have (had) 6 PV's in my VG. The last one crashed (the whole disk
> seems to have broken totaly).
> 
> It don't seem possible to restore the information (not without spending
> $$$ to a data recovery company), so how do I get the data stored on the
> OTHER disks?

You don't spend a single dollar :-)


There's 2 solutions:

A:
--

Buy a replacemeent drive (say /dev/sdg1 ;-) and run

fdisk # to partition it the same way the broken one was
pvcreate /dev/sdg1
vgcfgrestore -f /etc/lvmconf/vg1.conf -n vg1 /dev/sdg1
vgscan


B:
--

If you don't have a replacement drive at hand you can even be faster by
faking one (you must have loop configured in the kernel and use LVM 1.0.2):

dd if=/dev/zero of=/tmp/my_sdg1.vgda bs=1k count=512
losetup /dev/loop0 /tmp/my_sdg1.vgda
pvcreate /dev/loop0
vgcfgrestore -f /etc/lvmconf/vg1.conf -n vg1 /devloop0
vgscan


After either of the above aproaches you need to lvreduce/lvremove those
LVs which had extents allocated on the failed drive.

You can figure out which those are by

pvdisplay -v /dev/sdg1

- or -

pvdisplay -v /dev/loop0

respectively.

If you had a replacement drive (case A), you can continue to use it because
it will jsut be an empty PV which belonngs to your VG after the
lvreduce/lvremove runs recommended above.

If you used the loop0 fake (case B) run

vgreduce vg1 /dev/loop0
losetup -d /dev/loop0
rm /tmp/my_sdg1.vgda


Hope that helps :-)

Regards,
Heinz    -- The LVM Guy --

> 
> 
> [papadoc.root]# pvscan -u
> pvscan -- reading all physical volumes (this may take a while...)
> pvscan -- inactive PV "/dev/sdb1 " with UUID "ofZCeT-O16H-Df0h-4KoB-xQhm-lc24-67dZvk" of VG "vg1" [5.87 GB / 0 free]
> pvscan -- inactive PV "/dev/sdc1 " with UUID "xN0MF7-xKU0-Uk3L-g8Pw-0YfH-fS3V-zgFA66" of VG "vg1" [1004.00 MB / 0 free]
> pvscan -- inactive PV "/dev/sdd10" with UUID "5V0XNB-gA87-OU66-1iza-cvsz-xQ2q-6DEbyH" of VG "vg1" [2.37 GB / 0 free]
> pvscan -- inactive PV "/dev/sde1 " with UUID "Bdm7PX-50Dq-hC6R-yXmv-DDq1-n6Nx-4ohaKK" of VG "vg1" [8.50 GB / 0 free]
> pvscan -- inactive PV "/dev/sdf1 " with UUID "QnqfyA-fOTB-J1Pp-4be5-t6QU-bGjS-i2BKf2" of VG "vg1" [17.09 GB / 0 free]
> pvscan -- total: 5 [34.84 GB] / in use: 5 [34.84 GB] / in no VG: 0 [0]
> [papadoc.root]# /etc/init.d/lvm start
> Setting up LVM Volume Groups...
> vgchange -- ERROR: physical volume "/dev/sdg1" of volume group "vg1" is offline
> vgchange -- run vgscan
> [papadoc.root]# vgscan 
> vgscan -- reading all physical volumes (this may take a while...)
> vgscan -- "/etc/lvmtab" and "/etc/lvmtab.d" successfully created
> vgscan -- WARNING: This program does not do a VGDA backup of your volume group
> [papadoc.root]# pvscan -u
> pvscan -- reading all physical volumes (this may take a while...)
> pvscan -- inactive PV "/dev/sdb1 " with UUID "ofZCeT-O16H-Df0h-4KoB-xQhm-lc24-67dZvk"  is associated to an unknown VG (run vgscan)
> pvscan -- inactive PV "/dev/sdc1 " with UUID "xN0MF7-xKU0-Uk3L-g8Pw-0YfH-fS3V-zgFA66"  is associated to an unknown VG (run vgscan)
> pvscan -- inactive PV "/dev/sdd10" with UUID "5V0XNB-gA87-OU66-1iza-cvsz-xQ2q-6DEbyH"  is associated to an unknown VG (run vgscan)
> pvscan -- inactive PV "/dev/sde1 " with UUID "Bdm7PX-50Dq-hC6R-yXmv-DDq1-n6Nx-4ohaKK"  is associated to an unknown VG (run vgscan)
> pvscan -- inactive PV "/dev/sdf1 " with UUID "QnqfyA-fOTB-J1Pp-4be5-t6QU-bGjS-i2BKf2"  is associated to an unknown VG (run vgscan)
> pvscan -- total: 5 [34.84 GB] / in use: 5 [34.84 GB] / in no VG: 0 [0]
> 
> 
> This was the info the sg1 disk had:
> PV#     PV Name    VG Name PV UUID                                 PV Size
> 6       /dev/sdg1  vg1     7x1CdK-50u6-NAmJ-6bHY-Ho09-U3hC-K7c18w  15.55 GB 0 free
> 
> World Trade Center FSF spy Nazi terrorist supercomputer cryptographic
> toluene [Hello to all my fans in domestic surveillance] assassination
> counter-intelligence Ft. Bragg subway Rule Psix security
> [See http://www.aclu.org/echelonwatch/index.html for more about this]
> 
> _______________________________________________
> linux-lvm mailing list
> linux-lvm@sistina.com
> http://lists.sistina.com/mailman/listinfo/linux-lvm
> read the LVM HOW-TO at http://www.sistina.com/lvm/Pages/howto.html

*** Software bugs are stupid.
    Nevertheless it needs not so stupid people to solve them ***

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

Heinz Mauelshagen                                 Sistina Software Inc.
Senior Consultant/Developer                       Am Sonnenhang 11
                                                  56242 Marienrachdorf
                                                  Germany
Mauelshagen@Sistina.com                           +49 2626 141200
                                                       FAX 924446
=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-

Re: [linux-lvm] one PV crashed

[Turbo Fredriksson]

>>>>> "Heinz" == Heinz J Mauelshagen <mauelshagen@sistina.com> writes:

    Heinz> On Fri, Feb 01, 2002 at 11:01:22AM +0100, Turbo Fredriksson
    Heinz> wrote:
    >> I have (had) 6 PV's in my VG. The last one crashed (the whole
    >> disk seems to have broken totaly).
    >> 
    >> It don't seem possible to restore the information (not without
    >> spending $$$ to a data recovery company), so how do I get the
    >> data stored on the OTHER disks?

    Heinz> You don't spend a single dollar :-)

I was referring to restore the lost data on the failed disk ..

    Heinz> There's 2 solutions:

    Heinz> A: --
    Heinz> Buy a replacemeent drive (say /dev/sdg1 ;-) and run

    Heinz> fdisk # to partition it the same way the broken one was
    Heinz> pvcreate /dev/sdg1 vgcfgrestore -f /etc/lvmconf/vg1.conf -n
    Heinz> vg1 /dev/sdg1 vgscan

I'll get a replacement from IBM (the disk is still on warranty), so I'd
go for this...

Since I have data on the vg, will this still let me keep this, or will
it be destroyed when doing the pvcreate/vgcfgrestore?

    Heinz> B: --
    Heinz> If you don't have a replacement drive at hand you can even
    Heinz> be faster by faking one (you must have loop configured in
    Heinz> the kernel and use LVM 1.0.2):

    Heinz> dd if=/dev/zero of=/tmp/my_sdg1.vgda bs=1k count=512
    Heinz> losetup /dev/loop0 /tmp/my_sdg1.vgda pvcreate /dev/loop0
    Heinz> vgcfgrestore -f /etc/lvmconf/vg1.conf -n vg1 /devloop0
    Heinz> vgscan

I'll do this as a temporary solution, while waiting for the replacement.
Knowing IBM, it can take a while :)


Same question here, will the content on the VG be destroyed?

    Heinz> Hope that helps :-)

It does (at least in the sense that I'll won't loose MORE data :) Thanx.

spy NORAD Peking domestic disruption Kennedy subway Ortega president
quiche class struggle [Hello to all my fans in domestic surveillance]
AK-47 CIA Ft. Bragg attack
[See http://www.aclu.org/echelonwatch/index.html for more about this]

Re: [linux-lvm] one PV crashed

[Heinz J . Mauelshagen]

On Fri, Feb 01, 2002 at 11:52:08AM +0100, Turbo Fredriksson wrote:
> >>>>> "Heinz" == Heinz J Mauelshagen <mauelshagen@sistina.com> writes:
> 
>     Heinz> On Fri, Feb 01, 2002 at 11:01:22AM +0100, Turbo Fredriksson
>     Heinz> wrote:
>     >> I have (had) 6 PV's in my VG. The last one crashed (the whole
>     >> disk seems to have broken totaly).
>     >> 
>     >> It don't seem possible to restore the information (not without
>     >> spending $$$ to a data recovery company), so how do I get the
>     >> data stored on the OTHER disks?
> 
>     Heinz> You don't spend a single dollar :-)
> 
> I was referring to restore the lost data on the failed disk ..

Well, that's a different story :)

> 
>     Heinz> There's 2 solutions:
> 
>     Heinz> A: --
>     Heinz> Buy a replacemeent drive (say /dev/sdg1 ;-) and run
> 
>     Heinz> fdisk # to partition it the same way the broken one was
>     Heinz> pvcreate /dev/sdg1 vgcfgrestore -f /etc/lvmconf/vg1.conf -n
>     Heinz> vg1 /dev/sdg1 vgscan
> 
> I'll get a replacement from IBM (the disk is still on warranty), so I'd
> go for this...

Ok.

> 
> Since I have data on the vg, will this still let me keep this, or will
> it be destroyed when doing the pvcreate/vgcfgrestore?
> 
>     Heinz> B: --
>     Heinz> If you don't have a replacement drive at hand you can even
>     Heinz> be faster by faking one (you must have loop configured in
>     Heinz> the kernel and use LVM 1.0.2):
> 
>     Heinz> dd if=/dev/zero of=/tmp/my_sdg1.vgda bs=1k count=512
>     Heinz> losetup /dev/loop0 /tmp/my_sdg1.vgda pvcreate /dev/loop0
>     Heinz> vgcfgrestore -f /etc/lvmconf/vg1.conf -n vg1 /devloop0
>     Heinz> vgscan
> 
> I'll do this as a temporary solution, while waiting for the replacement.
> Knowing IBM, it can take a while :)

No comment.

> 
> 
> Same question here, will the content on the VG be destroyed?
> 
>     Heinz> Hope that helps :-)
> 
> It does (at least in the sense that I'll won't loose MORE data :) Thanx.

Right.

Did you try it already with loop?

> 
> spy NORAD Peking domestic disruption Kennedy subway Ortega president
> quiche class struggle [Hello to all my fans in domestic surveillance]
> AK-47 CIA Ft. Bragg attack
> [See http://www.aclu.org/echelonwatch/index.html for more about this]
> 
> _______________________________________________
> linux-lvm mailing list
> linux-lvm@sistina.com
> http://lists.sistina.com/mailman/listinfo/linux-lvm
> read the LVM HOW-TO at http://www.sistina.com/lvm/Pages/howto.html

-- 

Regards,
Heinz    -- The LVM Guy --

*** Software bugs are stupid.
    Nevertheless it needs not so stupid people to solve them ***

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

Heinz Mauelshagen                                 Sistina Software Inc.
Senior Consultant/Developer                       Am Sonnenhang 11
                                                  56242 Marienrachdorf
                                                  Germany
Mauelshagen@Sistina.com                           +49 2626 141200
                                                       FAX 924446
=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-

Re: [linux-lvm] one PV crashed

[Turbo Fredriksson]

>>>>> "Heinz" == Heinz J Mauelshagen <mauelshagen@sistina.com> writes:

    >>  I'll do this as a temporary solution, while waiting for the
    >> replacement.  Knowing IBM, it can take a while :)

    Heinz> No comment.

Ahh, a diplomat... That's rare in these day's n' age :)

    >>  It does (at least in the sense that I'll won't loose MORE data
    >> :) Thanx.

    Heinz> Right.

    Heinz> Did you try it already with loop?

Not yet, I haven't dared :) I don't want to risk loosing what's left...

Ft. Bragg Iran Saddam Hussein cracking plutonium Peking pits kibo
Ortega SDI Semtex Panama DES Marxist Waco, Texas
[See http://www.aclu.org/echelonwatch/index.html for more about this]

Re: [linux-lvm] one PV crashed

[Turbo Fredriksson]

Quoting "Heinz J . Mauelshagen" <mauelshagen@sistina.com>:

> dd if=/dev/zero of=/tmp/my_sdg1.vgda bs=1k count=512
> losetup /dev/loop0 /tmp/my_sdg1.vgda
> pvcreate /dev/loop0
> vgcfgrestore -f /etc/lvmconf/vg1.conf -n vg1 /devloop0
> vgscan

----- s n i p -----
[papadoc.root]# dd if=/dev/zero of=/var/my_sdg1.vgda bs=1k count=512
512+0 records in
512+0 records out
[papadoc.root]# losetup /dev/loop0 /var/my_sdg1.vgda
[papadoc.root]# pvcreate /dev/loop0
pvcreate -- physical volume "/dev/loop0" successfully created

[papadoc.root]# vgcfgrestore -f /etc/lvmconf/vg1.conf -n vg1 /dev/loop0
vgcfgrestore -- physical volume "/dev/loop0" doesn't belong to volume group "vg1"

[papadoc.root]# vgcfgrestore -f /etc/lvmconf/vg1.conf -n vg1 -o /dev/sdg1 /dev/loop0
vgcfgrestore -- ERROR "pv_read(): pv_create_name_from_kdev_t" reading physical volume "/dev/loop0"

[papadoc.root]# vgscan 
vgscan -- reading all physical volumes (this may take a while...)
vgscan -- "/etc/lvmtab" and "/etc/lvmtab.d" successfully created
vgscan -- WARNING: This program does not do a VGDA backup of your volume group

----- s n i p -----

> pvdisplay -v /dev/loop0

----- s n i p -----
[papadoc.root]# pvdisplay -v /dev/loop0 
pvdisplay -- getting physical volume size
pvdisplay -- ERROR "pv_read(): pv_create_name_from_kdev_t" no VALID physical volume "/dev/loop0"


----- s n i p -----


Panama Soviet Iran North Korea Kennedy smuggle Cocaine Ft. Meade 747
toluene 767 Qaddafi kibo cryptographic Clinton
[See http://www.aclu.org/echelonwatch/index.html for more about this]

Re: [linux-lvm] one PV crashed

[Turbo Fredriksson]

Quoting Turbo Fredriksson <turbo@bayour.com>:

> [papadoc.root]# vgcfgrestore -f /etc/lvmconf/vg1.conf -n vg1 /dev/loop0
> vgcfgrestore -- physical volume "/dev/loop0" doesn't belong to volume group "vg1"
> 
> [papadoc.root]# vgcfgrestore -f /etc/lvmconf/vg1.conf -n vg1 -o /dev/sdg1 /dev/loop0
> vgcfgrestore -- ERROR "pv_read(): pv_create_name_from_kdev_t" reading physical volume "/dev/loop0"
>
> [papadoc.root]# pvdisplay -v /dev/loop0 
> pvdisplay -- getting physical volume size
> pvdisplay -- ERROR "pv_read(): pv_create_name_from_kdev_t" no VALID physical volume "/dev/loop0"


Does anyone have a solution to this? What am I doing wrong?


SEAL Team 6 Iran killed KGB Kennedy Treasury cracking Semtex quiche
strategic radar Honduras Legion of Doom nitrate PLO
[See http://www.aclu.org/echelonwatch/index.html for more about this]

Re: [linux-lvm] one PV crashed

[Heinz J . Mauelshagen]

On Wed, Feb 06, 2002 at 12:01:29PM +0100, Turbo Fredriksson wrote:
> Quoting Turbo Fredriksson <turbo@bayour.com>:
> 
> > [papadoc.root]# vgcfgrestore -f /etc/lvmconf/vg1.conf -n vg1 /dev/loop0
> > vgcfgrestore -- physical volume "/dev/loop0" doesn't belong to volume group "vg1"
> > 
> > [papadoc.root]# vgcfgrestore -f /etc/lvmconf/vg1.conf -n vg1 -o /dev/sdg1 /dev/loop0
> > vgcfgrestore -- ERROR "pv_read(): pv_create_name_from_kdev_t" reading physical volume "/dev/loop0"
> >
> > [papadoc.root]# pvdisplay -v /dev/loop0 
> > pvdisplay -- getting physical volume size
> > pvdisplay -- ERROR "pv_read(): pv_create_name_from_kdev_t" no VALID physical volume "/dev/loop0"
> 
> 
> Does anyone have a solution to this? What am I doing wrong?

Turbo,

you seem to be using an older LVM version with broken loop support :-(
Found a bug in vgcfgrestore(8) which will likely cause problems with this too.
More to follow...


> 
> 
> SEAL Team 6 Iran killed KGB Kennedy Treasury cracking Semtex quiche
> strategic radar Honduras Legion of Doom nitrate PLO
> [See http://www.aclu.org/echelonwatch/index.html for more about this]
> 
> _______________________________________________
> linux-lvm mailing list
> linux-lvm@sistina.com
> http://lists.sistina.com/mailman/listinfo/linux-lvm
> read the LVM HOW-TO at http://www.sistina.com/lvm/Pages/howto.html

-- 

Regards,
Heinz    -- The LVM Guy --

*** Software bugs are stupid.
    Nevertheless it needs not so stupid people to solve them ***

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

Heinz Mauelshagen                                 Sistina Software Inc.
Senior Consultant/Developer                       Am Sonnenhang 11
                                                  56242 Marienrachdorf
                                                  Germany
Mauelshagen@Sistina.com                           +49 2626 141200
                                                       FAX 924446
=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-

Re: [linux-lvm] one PV crashed

[Turbo Fredriksson]

Quoting "Heinz J . Mauelshagen" <mauelshagen@sistina.com>:

> On Wed, Feb 06, 2002 at 12:01:29PM +0100, Turbo Fredriksson wrote:
> > Quoting Turbo Fredriksson <turbo@bayour.com>:
> > 
> > > [papadoc.root]# vgcfgrestore -f /etc/lvmconf/vg1.conf -n vg1 /dev/loop0
> > > vgcfgrestore -- physical volume "/dev/loop0" doesn't belong to volume group "vg1"
> > > 
> > > [papadoc.root]# vgcfgrestore -f /etc/lvmconf/vg1.conf -n vg1 -o /dev/sdg1 /dev/loop0
> > > vgcfgrestore -- ERROR "pv_read(): pv_create_name_from_kdev_t" reading physical volume "/dev/loop0"
> > >
> > > [papadoc.root]# pvdisplay -v /dev/loop0 
> > > pvdisplay -- getting physical volume size
> > > pvdisplay -- ERROR "pv_read(): pv_create_name_from_kdev_t" no VALID physical volume "/dev/loop0"
> > 
> > 
> > Does anyone have a solution to this? What am I doing wrong?
> 
> Turbo,
> 
> you seem to be using an older LVM version with broken loop support :-(
> Found a bug in vgcfgrestore(8) which will likely cause problems with this too.
> More to follow...

Dang! My LVM version is: 1.0.1-rc4

I'll try the other solution, with a separate harddisk. I borrowed a 18Gb disk to at least hold
SOME of the data I and my users are collecting (mail etc) :)

smuggle Honduras Iran class struggle Treasury PLO Serbian colonel spy
nuclear president FSF assassination Soviet kibo
[See http://www.aclu.org/echelonwatch/index.html for more about this]

Re: [linux-lvm] one PV crashed

[Turbo Fredriksson]

Quoting "Heinz J . Mauelshagen" <mauelshagen@sistina.com>:

> Buy a replacemeent drive (say /dev/sdg1 ;-) and run
> 
> fdisk # to partition it the same way the broken one was
> pvcreate /dev/sdg1
> vgcfgrestore -f /etc/lvmconf/vg1.conf -n vg1 /dev/sdg1
> vgscan

Oki, since I had a LVM with broken loop support, I'll try with a spare disk...

----- s n i p -----
[papadoc.root]# pvcreate /dev/sdh1
pvcreate -- physical volume "/dev/sdh1" successfully created

[papadoc.root]# vgcfgrestore -f /etc/lvmconf/vg1.conf -n vg1 /dev/sdh1
vgcfgrestore -- physical volume "/dev/sdh1" doesn't belong to volume group "vg1"

[papadoc.root]# vgcfgrestore -f /etc/lvmconf/vg1.conf -n vg1 -o /dev/sdg1 /dev/sdh1
vgcfgrestore -- size of physical volume /dev/sdh1 differs from backup
----- s n i p -----

I'm not sure exactly how big my previous (now crashed) sdg disk was, or how big
the sdg1 partition was...

How do I:

        1. Find out _exactly_ how big the sdg1 partition was (the only
           number I have is '15.5Gb')?
        or
        2. Use the  existing sdh1 even though the size don't match?


> After either of the above aproaches you need to lvreduce/lvremove those
> LVs which had extents allocated on the failed drive.
>
> You can figure out which those are by
>
> pvdisplay -v /dev/sdg1

----- s n i p ----
[papadoc.root]# pvdisplay -v /dev/sdh1
pvdisplay -- getting physical volume size
pvdisplay -- "/dev/sdh1" is a new physical volume of 2.22 GB
----- s n i p ----

Not nearly the same size as the original sdg1, but....

cracking Saddam Hussein Albanian class struggle smuggle BATF Legion of
Doom Cuba ammonium DES 747 767 pits Serbian plutonium
[See http://www.aclu.org/echelonwatch/index.html for more about this]

Re: [linux-lvm] one PV crashed

[Turbo Fredriksson]

>>>>> "Turbo" == Turbo Fredriksson <turbo@bayour.com> writes:

    Turbo> Oki, since I had a LVM with broken loop support, I'll try
    Turbo> with a spare disk...

    Turbo> [papadoc.root]# vgcfgrestore -f /etc/lvmconf/vg1.conf -n
    Turbo> vg1 -o /dev/sdg1 /dev/sdh1 vgcfgrestore -- size of physical
    Turbo> volume /dev/sdh1 differs from backup ----- s n i p -----

    Turbo> I'm not sure exactly how big my previous (now crashed) sdg
    Turbo> disk was, or how big the sdg1 partition was...

Oki, I managed to pvcreate my new sdh1, and then a vgcfgrestore.

        vgcfgrestore -f /etc/lvmconf/vg1.conf -n vg1 -o /dev/sdg1 /dev/sdh1

This went ok, but 'pvscan -u' is behaving (?) strangly...

----- s n i p -----
[papadoc.root]# pvscan -u
pvscan -- reading all physical volumes (this may take a while...)
pvscan -- inactive PV "/dev/sdb1 " with UUID "ofZCeT-O16H-Df0h-4KoB-xQhm-lc24-67dZvk"  is associated to an unknown VG (run vgscan)
pvscan -- inactive PV "/dev/sdc1 " with UUID "xN0MF7-xKU0-Uk3L-g8Pw-0YfH-fS3V-zgFA66"  is associated to an unknown VG (run vgscan)
pvscan -- inactive PV "/dev/sdd10" with UUID "5V0XNB-gA87-OU66-1iza-cvsz-xQ2q-6DEbyH"  is associated to an unknown VG (run vgscan)
pvscan -- inactive PV "/dev/sde1 " with UUID "Bdm7PX-50Dq-hC6R-yXmv-DDq1-n6Nx-4ohaKK"  is associated to an unknown VG (run vgscan)
pvscan -- inactive PV "/dev/sdf1 " with UUID "QnqfyA-fOTB-J1Pp-4be5-t6QU-bGjS-i2BKf2"  is associated to an unknown VG (run vgscan)
pvscan -- inactive PV "/dev/sdh1 " with UUID "ZNODti-EFFz-rGID-HYyh-kg7Z-yDCe-LLbXBv" is in no VG  [15.55 GB]
pvscan -- total: 6 [50.39 GB] / in use: 5 [34.84 GB] / in no VG: 1 [15.55 GB]
----- s n i p -----

Running vgscan (as directed) gives me:

----- s n i p -----
[papadoc.root]# vgscan 
vgscan -- reading all physical volumes (this may take a while...)
vgscan -- "/etc/lvmtab" and "/etc/lvmtab.d" successfully created
vgscan -- WARNING: This program does not do a VGDA backup of your volume group
----- s n i p -----

Ideas how to get my VG up and running again?

cryptographic Ft. Bragg NORAD plutonium 747 Legion of Doom Serbian
Saddam Hussein munitions Noriega supercomputer colonel fissionable
smuggle kibo
[See http://www.aclu.org/echelonwatch/index.html for more about this]

Re: [linux-lvm] one PV crashed

[Turbo Fredriksson]

>>>>> "Turbo" == Turbo Fredriksson <turbo@bayour.com> writes:

    Turbo> This went ok, but 'pvscan -u' is behaving (?) strangly...

Shutting down (vgchange -a n), doing the vgcfgrestore, and then starting
it again (vgchange -a y) gave me this:

----- s n i p -----
[papadoc.root]# /etc/init.d/lvm stop
Shutting down LVM Volume Groups...
vgchange -- no volume groups found

[papadoc.root]# vgcfgrestore -f /etc/lvmconf/vg1.conf -n vg1 -o /dev/sdg1 /dev/sdh1
vgcfgrestore -- VGDA for "vg1" successfully restored to physical volume "/dev/sdh1"
vgcfgrestore -- you may not have an actual backup of restored volume group "vg1"

[papadoc.root]# /etc/init.d/lvm start
Setting up LVM Volume Groups...
vgchange -- volume group "vg1" successfully activated

[papadoc.root]# pvscan -u
pvscan -- reading all physical volumes (this may take a while...)
pvscan -- ACTIVE   PV "/dev/sdb1 " with UUID "ofZCeT-O16H-Df0h-4KoB-xQhm-lc24-67dZvk" of VG "vg1" [5.87 GB / 0 free]
pvscan -- ACTIVE   PV "/dev/sdc1 " with UUID "xN0MF7-xKU0-Uk3L-g8Pw-0YfH-fS3V-zgFA66" of VG "vg1" [1004.00 MB / 0 free]
pvscan -- ACTIVE   PV "/dev/sdd10" with UUID "5V0XNB-gA87-OU66-1iza-cvsz-xQ2q-6DEbyH" of VG "vg1" [2.37 GB / 0 free]
pvscan -- ACTIVE   PV "/dev/sde1 " with UUID "Bdm7PX-50Dq-hC6R-yXmv-DDq1-n6Nx-4ohaKK" of VG "vg1" [8.50 GB / 0 free]
pvscan -- ACTIVE   PV "/dev/sdf1 " with UUID "QnqfyA-fOTB-J1Pp-4be5-t6QU-bGjS-i2BKf2" of VG "vg1" [17.09 GB / 0 free]
pvscan -- inactive PV "/dev/sdh1 " with UUID "ZNODti-EFFz-rGID-HYyh-kg7Z-yDCe-LLbXBv" is in no VG  [15.55 GB]
pvscan -- total: 6 [50.39 GB] / in use: 5 [34.84 GB] / in no VG: 1 [15.55 GB]
----- s n i p -----


I'm getting so close, I can SMELL the coffey! :)

Why is sdh1 inactive?

Qaddafi NSA Treasury NORAD FBI smuggle Saddam Hussein Kennedy Soviet
jihad bomb genetic Khaddafi pits quiche
[See http://www.aclu.org/echelonwatch/index.html for more about this]

Re: [linux-lvm] one PV crashed

[Heinz J . Mauelshagen]

Turbo,

did you run "vgscan" *before* the "vgchange -ay"?
Not running it could explain the inactive PV on /dev/sdh1.


On Fri, Feb 15, 2002 at 01:01:10PM +0100, Turbo Fredriksson wrote:
> >>>>> "Turbo" == Turbo Fredriksson <turbo@bayour.com> writes:
> 
>     Turbo> This went ok, but 'pvscan -u' is behaving (?) strangly...
> 
> Shutting down (vgchange -a n), doing the vgcfgrestore, and then starting
> it again (vgchange -a y) gave me this:
> 
> ----- s n i p -----
> [papadoc.root]# /etc/init.d/lvm stop
> Shutting down LVM Volume Groups...
> vgchange -- no volume groups found
> 
> [papadoc.root]# vgcfgrestore -f /etc/lvmconf/vg1.conf -n vg1 -o /dev/sdg1 /dev/sdh1
> vgcfgrestore -- VGDA for "vg1" successfully restored to physical volume "/dev/sdh1"
> vgcfgrestore -- you may not have an actual backup of restored volume group "vg1"
> 
> [papadoc.root]# /etc/init.d/lvm start
> Setting up LVM Volume Groups...
> vgchange -- volume group "vg1" successfully activated
> 
> [papadoc.root]# pvscan -u
> pvscan -- reading all physical volumes (this may take a while...)
> pvscan -- ACTIVE   PV "/dev/sdb1 " with UUID "ofZCeT-O16H-Df0h-4KoB-xQhm-lc24-67dZvk" of VG "vg1" [5.87 GB / 0 free]
> pvscan -- ACTIVE   PV "/dev/sdc1 " with UUID "xN0MF7-xKU0-Uk3L-g8Pw-0YfH-fS3V-zgFA66" of VG "vg1" [1004.00 MB / 0 free]
> pvscan -- ACTIVE   PV "/dev/sdd10" with UUID "5V0XNB-gA87-OU66-1iza-cvsz-xQ2q-6DEbyH" of VG "vg1" [2.37 GB / 0 free]
> pvscan -- ACTIVE   PV "/dev/sde1 " with UUID "Bdm7PX-50Dq-hC6R-yXmv-DDq1-n6Nx-4ohaKK" of VG "vg1" [8.50 GB / 0 free]
> pvscan -- ACTIVE   PV "/dev/sdf1 " with UUID "QnqfyA-fOTB-J1Pp-4be5-t6QU-bGjS-i2BKf2" of VG "vg1" [17.09 GB / 0 free]
> pvscan -- inactive PV "/dev/sdh1 " with UUID "ZNODti-EFFz-rGID-HYyh-kg7Z-yDCe-LLbXBv" is in no VG  [15.55 GB]
> pvscan -- total: 6 [50.39 GB] / in use: 5 [34.84 GB] / in no VG: 1 [15.55 GB]
> ----- s n i p -----
> 
> 
> I'm getting so close, I can SMELL the coffey! :)
> 
> Why is sdh1 inactive?
> 
> Qaddafi NSA Treasury NORAD FBI smuggle Saddam Hussein Kennedy Soviet
> jihad bomb genetic Khaddafi pits quiche
> [See http://www.aclu.org/echelonwatch/index.html for more about this]
> 
> _______________________________________________
> linux-lvm mailing list
> linux-lvm@sistina.com
> http://lists.sistina.com/mailman/listinfo/linux-lvm
> read the LVM HOW-TO at http://www.sistina.com/lvm/Pages/howto.html

-- 

Regards,
Heinz    -- The LVM Guy --

*** Software bugs are stupid.
    Nevertheless it needs not so stupid people to solve them ***

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

Heinz Mauelshagen                                 Sistina Software Inc.
Senior Consultant/Developer                       Am Sonnenhang 11
                                                  56242 Marienrachdorf
                                                  Germany
Mauelshagen@Sistina.com                           +49 2626 141200
                                                       FAX 924446
=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-