PDA

View Full Version : Fedora core 7 LVM issues


srinivasvaradar
20th June 2007, 06:33 PM
I recently upgrade from FC6 to FC7, upgrade went fine, but problem started after that. Here is my setup.
/dev/sda1 + /dev/hdd1 = /dev/md0 (raid1)
/dev/sda2 + /dev/hdd2 = /dev/md1 (raid1)

An LV group called storage sits on top of md0 and md1.
The problem is that the lv is displayed on startup as follows
Locking /var/lock/lvm/V_homegroup RB
Finding volume group "homegroup"
/dev/sda1: lvm2 label detected
/dev/sda2: lvm2 label detected
/dev/sda1: lvm2 label detected
/dev/sda1: size is 312576642 sectors
Fixing up missing format1 size (149.05 GB) for PV /dev/sda1
/dev/sda2: lvm2 label detected
/dev/sda2: size is 175815360 sectors
Fixing up missing format1 size (83.84 GB) for PV /dev/sda2
/dev/sda1: lvm2 label detected
/dev/sda1: size is 312576642 sectors
Fixing up missing format1 size (149.05 GB) for PV /dev/sda1
/dev/sda2: lvm2 label detected
/dev/sda2: size is 175815360 sectors
Fixing up missing format1 size (83.84 GB) for PV /dev/sda2
1 logical volume(s) in volume group "homegroup" already active
Finding volume group for uuid x5qSPOu9uNvCmnUnFE0y57tLwvE5Xl3gHqV7VC9d6P6Br8U7a1 cpubP9xOGsIsM3
/dev/sda1: lvm2 label detected
/dev/sda2: lvm2 label detected
/dev/sda1: lvm2 label detected
/dev/sda1: size is 312576642 sectors
Fixing up missing format1 size (149.05 GB) for PV /dev/sda1
/dev/sda2: lvm2 label detected
/dev/sda2: size is 175815360 sectors
Fixing up missing format1 size (83.84 GB) for PV /dev/sda2
/dev/sda1: lvm2 label detected
/dev/sda1: size is 312576642 sectors
Fixing up missing format1 size (149.05 GB) for PV /dev/sda1
/dev/sda2: lvm2 label detected
/dev/sda2: size is 175815360 sectors
Fixing up missing format1 size (83.84 GB) for PV /dev/sda2
Found volume group "homegroup"
Locking memory
Setting activation/missing_stripe_filler to /dev/ioerror
Getting target version for linear
Getting target version for striped
Loading homegroup-storage table
device-mapper: reload ioctl failed: Invalid argument
Unlocking memory
1 logical volume(s) in volume group "homegroup" now active
Unlocking /var/lock/lvm/V_homegroup
Dumping persistent device cache to /etc/lvm/.cache
Locking /etc/lvm/.cache (F_WRLCK, 1)
Unlocking fd 3
Wiping internal VG cache

The /dev/mapper/ shows homegroup-storage
which is the VG.
There is no /dev/homegroup available.
At the same time the error message in the syslog shows
Jun 20 12:05:01 pdc kernel: device-mapper: table: 253:0: linear: dm-linear: Device lookup failed
Jun 20 12:05:01 pdc kernel: device-mapper: ioctl: error adding target to table

So there is a problem, I am not sure if this is a bug or an error in the setup. Any help is appreciated, including comments. Thanks again

JEO
20th June 2007, 10:25 PM
Drives that used to be called hdd1 and hdd2 in FC6 have changed to the new libata format in F7 and would show up as sdX1 and sdX2. Try booting up the rescue cd noting which pata_ or ata modules are loaded with the lsmod |grep ata command. Then use mkinitrd witht the --preload=<that kernel module name> option. That should fix the syslog errors during bootup.
If you are not experiencing the syslog errors during bootup then edit your raid configuration to include the new /dev/sdX format drive names.

srinivasvaradar
20th June 2007, 11:27 PM

Thanks for the post, from the /sbin/lvm.static command, I think its already recognizing the drives as /dev/sdX, and I was able to rebuild my raid1s, without a problem. The problem is the lv groups is not being mapped to be mounted.

JEO
21st June 2007, 03:20 AM
Try a vgscan command and a vgchange -ay command and see what those produce.

srinivasvaradar
21st June 2007, 10:45 PM
nice catch, vgscan results in a segmentation fault.
Reading all physical volumes. This may take a while...
Finding all volume groups
/dev/ramdisk: size is 32768 sectors
/dev/ramdisk: size is 32768 sectors
/dev/ramdisk: No label detected
/dev/loop0: size is 0 sectors
/dev/sda: size is 488397168 sectors
/dev/md0: size is 312576512 sectors
/dev/md0: size is 312576512 sectors
/dev/md0: lvm2 label detected
Segmentation fault

JEO
22nd June 2007, 02:11 AM
in lvm2 vgscan is ran by the init scripts automatically. Now the problem is what to do about it? Since you created the volume under Fedora Core 6, you might try accessing it from the FC6 rescue CD. Then you could rsync your data to a backup drive or whatever.