linux-lvm.redhat.com archive mirror
 help / color / mirror / Atom feed
* [linux-lvm] Missing 1 of 3 PVs and all partition tables are good
@ 2005-11-09 17:37 Cory Zerwas
  2005-11-09 21:03 ` Cory Zerwas
  0 siblings, 1 reply; 2+ messages in thread
From: Cory Zerwas @ 2005-11-09 17:37 UTC (permalink / raw)
  To: linux-lvm

[-- Attachment #1: Type: text/plain, Size: 1826 bytes --]

Hi all,

I'm in a pinch. I've got about 500Gb of data trapped in a non-functioning
VG... Here's the deal:

I upgraded from FC3 to FC4, or tried to. The whole update bombed out but
during this time the LVM disks were unplugged from the system. When I got
the new system up on the old FS my LVM conf still appears to be in place. I
powered up the 3 drives and reset the system and I found the VG won't
activate.
I did not export the volume first. I realize now that this was a mistake.

I have 3 whole disks for this VG, each with 1 partition, type 8e, occupying
the entire disk. These are hdb1, hdc1, hdd1.

I'm noticing that an "fdisk -l" lists hdb, hdc, and hdd as all having
partitions of type 8e from beginning to end of disk. However, when I go into
LVM and run "pvscan" I only see two PVs belonging to my VG "bloaty" and it
shows:

PV /dev/hdc1 VG bloaty lvm2 [111.78 GB / 0 free]
PV /dev/hdd1 VG bloaty lvm2 [115.03 GB / 64.00 MB free]
Total: 2 [266.81 GB] / in use: 2 [226.81 GB] / in no VG: 0 [0 ]

LVM doesn't seem to notice /dev/hdb1? It's a 232.88 Gb PV, how can this be?
And why isn't PVSCAN complaining about a missing PV? It should be, shouldn't
it?

When I perform a "lvmdiskscan" I get (the parenthesis outline my input):

/dev/hda1 [ 101.94 MB]
(^^this is the /boot partition)
/dev/hdc1 [ 111.79 GB] LVM physical volume
/dev/hda2 [ 17.54 GB]
(^^this is the / partition)
/dev/hda3 [ 1.00 GB]
(^^this is the swap partition)
/dev/hdb1 [ 232.88 GB]
(^^THIS SHOULD BE A PV)
/dev/hdd1 [ 115.04 GB] LVM physical volume

Is there some trick here I don't know about? Do I have to restore the LVM
conf?

I am stumped. I'm pretty sure I'm not a newb, but I am definately not a
wizard. I've got my LPIC Level 1 cert if that means anything.

Heh, thanks for any help,
Cory

[-- Attachment #2: Type: text/html, Size: 2368 bytes --]

^ permalink raw reply	[flat|nested] 2+ messages in thread

* Re: [linux-lvm] Missing 1 of 3 PVs and all partition tables are good
  2005-11-09 17:37 [linux-lvm] Missing 1 of 3 PVs and all partition tables are good Cory Zerwas
@ 2005-11-09 21:03 ` Cory Zerwas
  0 siblings, 0 replies; 2+ messages in thread
From: Cory Zerwas @ 2005-11-09 21:03 UTC (permalink / raw)
  To: linux-lvm

[-- Attachment #1: Type: text/plain, Size: 2819 bytes --]

Whoops, where's my head at... I forgot the critical bits:

LVM ver 2.01.08 (2005-03-22)
Library ver 1.01.02 (2005-05-17)
Driver ver 4.4.0
Fedora Core 4 with 2.6.13-1.1526 kernel

hda are the system essentials; /boot, /, and swap
hdb, hdc, hdd are the LVM disks. Each disk contains 1 primary type 8e
partition from beginning to end of disk.

Also, recently while I was tinkering with it I attempted a 'vgcfgrestore
bloaty' and received:

Couldn't find device with uuid 'tfkC8Z-sWLe-09A5-1Yow-qKw1-0DZz-x73mWt'.
Couldn't find all physical volumes for volume group videovg.
Restore failed.

I surmise that that uuid belongs to hdb1. Would this be the appropriate
point to restore pv metadata?
http://www.tldp.org/HOWTO/html_single/LVM-HOWTO/#recovermetadata

If I've omitted anything please let me know.

Thanks again,
Cory

On 11/9/05, Cory Zerwas <gannas@gmail.com> wrote:
>
> Hi all,
>
> I'm in a pinch. I've got about 500Gb of data trapped in a non-functioning
> VG... Here's the deal:
>
> I upgraded from FC3 to FC4, or tried to. The whole update bombed out but
> during this time the LVM disks were unplugged from the system. When I got
> the new system up on the old FS my LVM conf still appears to be in place. I
> powered up the 3 drives and reset the system and I found the VG won't
> activate.
> I did not export the volume first. I realize now that this was a mistake.
>
> I have 3 whole disks for this VG, each with 1 partition, type 8e,
> occupying the entire disk. These are hdb1, hdc1, hdd1.
>
> I'm noticing that an "fdisk -l" lists hdb, hdc, and hdd as all having
> partitions of type 8e from beginning to end of disk. However, when I go into
> LVM and run "pvscan" I only see two PVs belonging to my VG "bloaty" and it
> shows:
>
> PV /dev/hdc1 VG bloaty lvm2 [111.78 GB / 0 free]
> PV /dev/hdd1 VG bloaty lvm2 [115.03 GB / 64.00 MB free]
> Total: 2 [266.81 GB] / in use: 2 [226.81 GB] / in no VG: 0 [0 ]
>
> LVM doesn't seem to notice /dev/hdb1? It's a 232.88 Gb PV, how can this
> be? And why isn't PVSCAN complaining about a missing PV? It should be,
> shouldn't it?
>
> When I perform a "lvmdiskscan" I get (the parenthesis outline my input):
>
> /dev/hda1 [ 101.94 MB]
> (^^this is the /boot partition)
> /dev/hdc1 [ 111.79 GB] LVM physical volume
> /dev/hda2 [ 17.54 GB]
> (^^this is the / partition)
> /dev/hda3 [ 1.00 GB]
> (^^this is the swap partition)
> /dev/hdb1 [ 232.88 GB]
> (^^THIS SHOULD BE A PV)
> /dev/hdd1 [ 115.04 GB] LVM physical volume
>
> Is there some trick here I don't know about? Do I have to restore the LVM
> conf?
>
> I am stumped. I'm pretty sure I'm not a newb, but I am definately not a
> wizard. I've got my LPIC Level 1 cert if that means anything.
>
> Heh, thanks for any help,
> Cory
>

[-- Attachment #2: Type: text/html, Size: 3748 bytes --]

^ permalink raw reply	[flat|nested] 2+ messages in thread

end of thread, other threads:[~2005-11-09 21:03 UTC | newest]

Thread overview: 2+ messages (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
2005-11-09 17:37 [linux-lvm] Missing 1 of 3 PVs and all partition tables are good Cory Zerwas
2005-11-09 21:03 ` Cory Zerwas

This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).