0

I'm struggling with the weirdest LVM problem ever.

I'm running a small server at home with a Software Raid 5 from 4 3 TB Disks. (There is no space for a Hardware Raid Controller in the Case).

From Technology it is an Intel Rapid Storage Technology.

The Storage mount of this Server looks like:

Phys Disk -> Software Raid 5 -> LVM2 -> EXT4

After a Power outage last week, the Bios was in a faulty state, so i reconfigured the Bios to handle the Disks as RAID again.

But then, after a first look in the Raid Controller Interface i found the following:

enter image description here

Status Failed because one Disk was completely Removed from Raid Info?!

So as there where no other options then to Delete and Recreate the Raid array, i could boot up the System after that but without my LVM Storage Space inside.

The mdadm.conf changed from:

ARRAY metadata=imsm UUID=3870907f:d7fafe1a:593e95f1:3496eea9
ARRAY /dev/md/Volume1 container=3870907f:d7fafe1a:593e95f1:3496eea9 member=0 UUID=0429e388:74a0f3ea:8ca3b4a3:5f976ec2

to:

ARRAY metadata=imsm UUID=52065345:de8d583c:8fa56cc8:55d93798
ARRAY /dev/md/Volume1 container=52065345:de8d583c:8fa56cc8:55d93798 member=0 UUID=b8f3dcf7:0e95fede:e6224238:b2426835

and

blkid /dev/md126

now returns

/dev/md126: UUID="Wzcm2T-H2pt-c41w-dbi5-K0IP-O7FU-GRfoA3" TYPE="LVM2_member"

Due to the recreation of the softwareraid, I changed the UUID for PV0 from mowPbj-0RMX-Glr2-e6ar-vOgP-mcHH-k4sAmc to Wzcm2T-H2pt-c41w-dbi5-K0IP-O7FU-GRfoA3 in /etc/lvm/backup/LogVol0:

# Generated by LVM2 version 2.02.168(2) (2016-11-30): Mon Oct 14 00:05:24 2019

contents = "Text Format Volume Group"
version = 1

description = "Created *after* executing 'vgchange -ay'"

creation_host = "home"  # Linux home 4.15.17-1-pve #1 SMP PVE 4.15.17-9 (Wed, 9 May 2018 13:31:43 +0200) x86_64
creation_time = 1571004324  # Mon Oct 14 00:05:24 2019

LogVol0 {
    id = "Ch3etp-GYit-9qVW-nhEg-bv3K-IQnW-0jCjgM"
    seqno = 16
    format = "lvm2"         # informational
    status = ["RESIZEABLE", "READ", "WRITE"]
    flags = []
    extent_size = 8192      # 4 Megabytes
    max_lv = 0
    max_pv = 0
    metadata_copies = 0

    physical_volumes {

        pv0 {
            id = "Wzcm2T-H2pt-c41w-dbi5-K0IP-O7FU-GRfoA3"
            device = "/dev/md126"   # Hint only

            status = ["ALLOCATABLE"]
            flags = []
            dev_size = 17581580288  # 8.18706 Terabytes
            pe_start = 2048
            pe_count = 2146188  # 8.18706 Terabytes
        }
    }

    logical_volumes {

        Logical_Volume_00 {
            id = "BWUgvn-24uy-XXf5-aFVx-o9YA-RF2H-2uD3OV"
            status = ["READ", "WRITE", "VISIBLE"]
            flags = []
            creation_time = 1395998316  # 2014-03-28 10:18:36 +0100
            creation_host = "debian"
            segment_count = 1

            segment1 {
                start_extent = 0
                extent_count = 2146188  # 8.18706 Terabytes

                type = "striped"
                stripe_count = 1    # linear

                stripes = [
                    "pv0", 0
                ]
            }
        }
    }

}

Current output of 'pvdisplay':

pvdisplay 
  --- Physical volume ---
  PV Name               /dev/md126
  VG Name               LogVol0
  PV Size               8.19 TiB / not usable 4.00 MiB
  Allocatable           yes (but full)
  PE Size               4.00 MiB
  Total PE              2146188
  Free PE               0
  Allocated PE          2146188
  PV UUID               Wzcm2T-H2pt-c41w-dbi5-K0IP-O7FU-GRfoA3

(The pvdisk was missing first as i tried to restore VolumeGroup with 'vgcfgrestore LogVol0 LogVol0', so i created this new pvdisk)

After creating a "new" PVdisk and restoring the Backup File, lvdisplay looks like:

  --- Logical volume ---
  LV Path                /dev/LogVol0/Logical_Volume_00
  LV Name                Logical_Volume_00
  VG Name                LogVol0
  LV UUID                BWUgvn-24uy-XXf5-aFVx-o9YA-RF2H-2uD3OV
  LV Write Access        read/write
  LV Creation host, time debian, 2014-03-28 10:18:36 +0100
  LV Status              NOT available
  LV Size                8.19 TiB
  Current LE             2146188
  Segments               1
  Allocation             inherit
  Read ahead sectors     auto

pvs:

  PV         VG      Fmt  Attr PSize   PFree 
  /dev/md126 LogVol0 lvm2 a--    8.19t     0 

lvscan:

  inactive          '/dev/LogVol0/Logical_Volume_00' [8.19 TiB] inherit

lsblk -f:

sdb                isw_raid_member                                              
└─md126            LVM2_member           Wzcm2T-H2pt-c41w-dbi5-K0IP-O7FU-GRfoA3 
sdc                isw_raid_member                                              
└─md126            LVM2_member           Wzcm2T-H2pt-c41w-dbi5-K0IP-O7FU-GRfoA3 
sdd                isw_raid_member                                              
└─md126            LVM2_member           Wzcm2T-H2pt-c41w-dbi5-K0IP-O7FU-GRfoA3 
sde                isw_raid_member                                              
└─md126            LVM2_member           Wzcm2T-H2pt-c41w-dbi5-K0IP-O7FU-GRfoA3 

so if I activate that VG with 'vgchange -ay' the output changes to:

sdb                           isw_raid_member                                              
└─md126                       LVM2_member           Wzcm2T-H2pt-c41w-dbi5-K0IP-O7FU-GRfoA3 
  └─LogVol0-Logical_Volume_00 LVM2_member           mowPbj-0RMX-Glr2-e6ar-vOgP-mcHH-k4sAmc 
sdc                           isw_raid_member                                              
└─md126                       LVM2_member           Wzcm2T-H2pt-c41w-dbi5-K0IP-O7FU-GRfoA3 
  └─LogVol0-Logical_Volume_00 LVM2_member           mowPbj-0RMX-Glr2-e6ar-vOgP-mcHH-k4sAmc 
sdd                           isw_raid_member                                              
└─md126                       LVM2_member           Wzcm2T-H2pt-c41w-dbi5-K0IP-O7FU-GRfoA3 
  └─LogVol0-Logical_Volume_00 LVM2_member           mowPbj-0RMX-Glr2-e6ar-vOgP-mcHH-k4sAmc 
sde                           isw_raid_member                                              
└─md126                       LVM2_member           Wzcm2T-H2pt-c41w-dbi5-K0IP-O7FU-GRfoA3 
  └─LogVol0-Logical_Volume_00 LVM2_member           mowPbj-0RMX-Glr2-e6ar-vOgP-mcHH-k4sAmc  

And type of /dev/LogVol0/Logical_Volume_00 is:

/dev/LogVol0/Logical_Volume_00: LVM2 PV (Linux Logical Volume Manager), UUID: mowPbj-0RMX-Glr2-e6ar-vOgP-mcHH-k4sAmc, size: 9001769107456

So I, can not mount it, because I've created the "OLD" LVM within a new LVM?

With Testdisk I can see the ext4 partition inside the inner LVM, but how can I get back my mow... Volume?

Thx for any help...

(Oh yeah, I know I need more Backups...)

1 Answer 1

0

When a LV shows as inactive, I'ld try something like:

lvchange -ay

Side note: something's weird ... Either you setup your RAID from your bios, and then that's a hardware raid, your system only sees a virtual device (/dev/sdX). Even if you do not have some dedicated card dealing with this: your motherboard can play that role.

Or you have a software raid, your OS sees all disks, you have some special device representing your raid (/dev/mdX). From the look of it, this is your case. Though I don't understand what's going on in your BIOS.

2
  • It's called a FakeRaid and so i have both /dev/sd* and /dev/md* in system available ...
    – CyberLine
    Oct 15, 2019 at 14:02
  • more importantly, did you get your LV back? looking into fakeraid, I see very few cases for it to make sense (dual-booting with non-linux systems, or linux without raid?). Sounds like risky anyway, if somehow based on motherboard. Hope you have spare parts.
    – SYN
    Oct 15, 2019 at 14:51

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .