lvm2 migration issue

I'm trying to upgrade one of my Fedora Core 1 boxes to FC2, which has the root filesystem on LVM. The installer won't allow me to ugprade because, it seems, vgscan is returning an error.

The output of "lvm vgscan -v -d" from a VC whilst running the installer is:

1 PV(s) found for VG mainvg: expected 2
Volume group "mainvg" not found
Wiping cache of LVM-capable devices
Wiping internal cache
Reading all physical volumes. This may take a while...
Finding all volume groups
Finding volume group "mainvg"

"lvm pvscan" gives:

1 PV(s) found for VG mainvg: expected 2
Logical volume (usrvol) contains an incomplete mapping table.
Wiping cache of LVM-capable devices
Wiping internal cache
Walking through all physical volumes
PV /dev/hda7 VG mainvg lvm1 [19.82 GB / 0 free]
Total: 1 [19.82 GB] / in use: 1 [19.82 GB] / in no VG: 0 [0 ]

so it doesn't recognise /dev/hda2 as a PV, for some reason.

When booted into FC1, /proc/lvm/global has:

LVM module LVM version 1.0.5+(22/07/2002)

Total: 1 VG 2 PVs 2 LVs (1 LV open 1 times)

Global: 105960 bytes malloced IOP version: 10 0:30:18 active

VG: mainvg [2 PV, 2 LV/1 open] PE Size: 4096 KB
Usage [KB/PE]: 25907200 /6325 total 25907200 /6325 used 0 /0 free
PVs: [AA] hda2 5120000 /1250 5120000 /1250 0 /0

       [AA] hda7                  20787200 /5075    20787200 /5075           0 /0     
    LVs: [AWDL  ] usrvol                     5120000 /1250     1x open
         [AWDL  ] storevol                  20787200 /5075     close

and back booted into the FC2 installer,

# lvm pvdisplay  /dev/hda2
  Failed to read physical volume "/dev/hda2"

Any ideas how I can get round this?

Regards,

joe

Depends what's causing it: try turning up the verbosity (eg. vgscan -vvv) and see if it gives a reason for avoiding hda2.

Alasdair

[email protected]

Thanks a lot, that did the trick. It skipped that partition because it had a old raid superblock on it, so I ran "mdadm --zero-superblock /dev/hda2" from the installer shell and vgscan was happy again.

Regards,

joe