[linux-lvm] PVs not found (kernel 2.4<->devfs?) [was: lvm lost after reboot!?]
Constantin Mercier
admin at kitchhiking.org
Tue Jul 19 00:00:13 UTC 2005
Hi,
I'm still trying to solve the problem explained earlier (see
quote below) and maybe understood better now what might be
relevant information after digging a bit deeper.
-----
What happened:
System: Debian 3.1 (Sarge) / 2.4.27-2 (standard 2.4 kernel from
netinstall)
# dmesg|grep -B1 -i lvm
device-mapper: 4.1.1-ioctl (2004-04-07) initialised: dm-devel at redhat.com
LVM version 1.0.8(17/11/2003) module loaded
2 VGs, each on an own partition/disk, so no spanning
Disks are working fine (/ and /boot are conventional partions on /dev/
hda,
VG WD2000JB-182GB is on /dev/hda4, machine boots fine)
I rebooted the system today and afterwards the LVs did not mount:
cserver:~# cat /var/log/boot|grep -i "block device"
Tue Jul 19 03:16:56 2005: mount: /dev/WD2000JB-182GB/mld is not a
valid block device
Tue Jul 19 03:16:56 2005: mount: /dev/WD2000JB-182GB/srv is not a
valid block device
Tue Jul 19 03:16:56 2005: mount: /dev/IBM80G2/costa is not a valid
block device
I tried a hint from Heinz:
# cserver:/# cp /etc/lvmconf/WD2000JB-182GB.conf /etc/lvmtab.d/
# cserver:/# echo -ne "WD2000JB-182GB\0" > /etc/lvmtab
but then:
# vgchange -ay -v WD2000JB-182GB
vgchange -- locking logical volume manager
vgchange -- using volume group(s) on command line
vgchange -- checking volume group name "WD2000JB-182GB"
vgchange -- checking existence of volume group "WD2000JB-182GB"
vgchange -- reading volume group data for "WD2000JB-182GB" from lvmtab
vgchange -- checking volume group consistency of "WD2000JB-182GB"
vgchange -- checking if all physical volumes of volume group
"WD2000JB-182GB" are available
vgchange -- ERROR: can't get name(s) of physical volumes
vgchange -- Please check, if /proc is mounted
-----
Hints:
/proc/lvm _is_ mounted
I read a post in [linux-lvm] *) (similar trouble with LVM after
mounting a SCSI cd-writer) and recalled I had attached
a USB-stick before rebooting.
*) https://www.redhat.com/archives/linux-lvm/2002-January/msg00318.html
devfs seems to be enabled:
# dmesg|grep -i devfs
devfs: v1.12c (20020818) Richard Gooch (rgooch at atnf.csiro.au)
devfs: boot_options: 0x0
but not used. But when I booted with the usb stick attached it might
have
mounted (usbdevfs)?
my /proc/partitions looks like
major minor #blocks name rio rmerge rsect ruse wio wmerge wsect
wuse running use aveq
33 0 80418240 ide/host2/bus0/target0/lun0/disc
33 1 80413326 ide/host2/bus0/target0/lun0/part1
22 0 712618 hdc
3 0 195360984 ide/host0/bus0/target0/lun0/disc
3 1 48163 ide/host0/bus0/target0/lun0/part1
3 2 489982 ide/host0/bus0/target0/lun0/part2
3 3 4883760 ide/host0/bus0/target0/lun0/part3
3 4 189936495 ide/host0/bus0/target0/lun0/part4
3 64 80418240 ide/host0/bus0/target1/lun0/disc
3 65 80418208 ide/host0/bus0/target1/lun0/part1
/proc/partitions looks like:
major minor #blocks name rio rmerge rsect ruse wio wmerge wsect
wuse running use aveq
33 0 80418240 ide/host2/bus0/target0/lun0/disc 5084 46140
102448 5800 0 0 0 0 -2 10974830 21005752
33 1 80413326 ide/host2/bus0/target0/lun0/part1 2 6 16 0 0 0
0 0 0 0 0
22 0 712618 hdc 23 6 112 4540 0 0 0 0 -158 10963830 41273884
3 0 195360984 ide/host0/bus0/target0/lun0/disc 25141 147207
504590 78200 6251 13538 158714 72850 -2 10971270 21150982
3 1 48163 ide/host0/bus0/target0/lun0/part1 29 553 1164
240 12 5 34 850 0 740 1090
3 2 489982 ide/host0/bus0/target0/lun0/part2 1 0 8 10 0 0
0 0 0 10 10
3 3 4883760 ide/host0/bus0/target0/lun0/part3 10688 16077
213418 32450 6239 13533 158680 72000 0 28160 104450
3 4 189936495 ide/host0/bus0/target0/lun0/part4 9 27 72 0 0
0 0 0 0 0 0
3 64 80418240 ide/host0/bus0/target1/lun0/disc 19 61 226 100
3 0 24 10 -2 10974860 21000062
3 65 80418208 ide/host0/bus0/target1/lun0/part1 12 0 90 30 3
0 24 10 0 40 40
Alan was so kind to give a couple of advises but I admittedly think
I could not provide enough detail to make him able to help me out.
Sorry for being so unstructured in my eralier descriptions,
I was a bit panicking earlier the day.
I'd be really grateful for any further advice.
TIA for your efforts,
Constantin
> Constantin Mercier wrote:
> Hi,
>
> I'm using LVM only for about a month now and it ran flawless until
> today.
>
> After rebooting the system all my PVs/VGs/LVs were gone, "mount"
> saying
> e.g.:
>
> > mount: /dev/WD2000JB-182GB/srv is not a valid block device
>
> /etc/lvmtab now is an empty file, /etc/lvmtab.d, /etc/lvm/backup and
> /etc/lvm/archive are empty as well.
>
> This is strange because etc/lvm/lvm.conf says that backups and
> archives
> should be created but i can't find anything.
>
> What remains are some files in /etc/lvmconf named
> "WD2000JB-182GB.conf" and
> "IBM80G2.conf" and some older versions of those.
>
> I read on the web *) about recovering the data using pvcreate/
> vgcfgrestore,
> but since "/etc/lvm/backup" and "/etc/lvm/archive" are empty I
> don't know
> how to retrieve the UUIDs of my PVs.
>
> *) http://codeworks.gnomedia.com/archives/2005/general/lvm_recovery/
>
> Of course I don't have any further backup of these, since I first
> read about their
> importance doing that when digging for hints howto recover my data.
>
> (stupid me (tm))
>
> The System is running Debian Sarge stable (2.4.27) on a
> P3/Intel BX with 3 harddrives attached to it. LVM is Version 1.08
>
> from /etc/fstab:
>
> #lvm
> /dev/WD2000JB-182GB/mld /home/mld auto
> suid,dev,exec 0 0
> /dev/WD2000JB-182GB/srv /srv/foo auto
> suid,noexec,nodev 0 0
> /dev/IBM80G2/costa_home /home/costa auto
> suid,noexec,nodev 0 0
>
> where:
>
> /dev/WD2000JB-182GB/ is situated on /dev/hda (a 200GB Harddisk)
> /dev/IBM80G2/costa_home sits on /dev/hde (an 80GB Harddisk attached
> to an onboard
> HPT366 ATA66 Controller)
>
> pvdisplay /dev/WD2000JB-182GB says:
>
> > cserver:/etc/lvm# pvdisplay /dev/hde
> > pvdisplay -- ERROR "pv_read(): pv_create_name_from_kdev_t" no
> VALID physical volume "/dev/hde"
>
> vgdisplay just returns:
>
> > cserver:/etc/lvm# vgdisplay
> > vgdisplay -- no volume groups found
>
> If any of you has a clou how i could recover my data (or just the
> UUIDs) or
> maybe rebuild the LVM I'd be really grateful.
>
> Hope to hear from you,
> Constantin
>
> _______________________________________________
> linux-lvm mailing list
> linux-lvm at redhat.com
> https://www.redhat.com/mailman/listinfo/linux-lvm
> read the LVM HOW-TO at http://tldp.org/HOWTO/LVM-HOWTO/
>
More information about the linux-lvm
mailing list