lvm with lfs

Quoted 6 lines · Jens Benecke

Hello,

I upgraded to 2.2.4 (0.9.1-beta2) from 2.2.18 (LVM 0.9) and this seems to have killed one of my PVs. I tried the 0.9 utilities and updated to 0.9.1-beta7 when I noticed the problems. No change. Booting back to 2.2.18 didn't help, the PV was gone there as well.

follow-up:

vgcfgrestore doesn't work, because LVM thinks my hdc1 is no PV. Here's the output:

vgcfgrestore -n data -ll:
--- Volume group ---
VG Name data
VG Access read/write
VG Status NOT available/resizable
VG # 0
MAX LV 256
Cur LV 1
Open LV 0
MAX LV Size 255.99 GB
Max PV 256
Cur PV 2
Act PV 2
VG Size 95.44 GB
PE Size 4 MB
Total PE 24432
Alloc PE / Size 24432 / 95.44 GB
Free PE / Size 0 / 0
VG UUID ec0e1J-hf9c-2DVg-VMGE-Qn1c-HPgl-z1jMGO

--- Logical volume ---
LV Name /dev/data/archiv
VG Name data
LV Write Access read/write
LV Status NOT available
LV # 1
# open 0
LV Size 95.44 GB
Current LE 24432
Allocated LE 24432
Allocation next free
Read ahead sectors 120
Block device 58:0

--- Physical volume ---
PV Name /dev/hdc1
VG Name data
PV Size 76.33 GB / NOT usable 3.96 MB [LVM: 197 KB]
PV# 1
PV Status NOT available
Allocatable yes (but full)
Cur LV 1
PE Size (KByte) 4096
Total PE 19539
Free PE 0
Allocated PE 19539
PV UUID Re4KFN-ES3b-K2zz-EIer-iYKv-i7wX-lHTPGN

--- Physical volume ---
PV Name /dev/hda1
VG Name data
PV Size 19.12 GB / NOT usable 2.27 MB [LVM: 140 KB]
PV# 2
PV Status NOT available
Allocatable yes (but full)
Cur LV 1
PE Size (KByte) 4096
Total PE 4893
Free PE 0
Allocated PE 4893
PV UUID qT529c-dwfb-utXE-1Ds4-18AV-0I6O-2cHdfR

vgcfgrestore -v -n data /dev/hdc1:
vgcfgrestore -- locking logical volume manager
vgcfgrestore -- restoring volume group "data" from "/etc/lvmconf/data.conf" vgcfgrestore -- checking existence of "/etc/lvmconf/data.conf"
vgcfgrestore -- reading volume group data for "data" from "/etc/lvmconf/data.conf" vgcfgrestore -- reading physical volume data for "data" from "/etc/lvmconf/data.conf" vgcfgrestore -- reading logical volume data for "data" from "/etc/lvmconf/data.conf" vgcfgrestore -- checking volume group consistency of "data"
vgcfgrestore -- checking volume group consistency of "data"
vgcfgrestore -- reading physical volume "/dev/hdc1"
vgcfgrestore -- ERROR "pv_read(): PV identifier invalid" reading physical volume "/dev/hdc1"

Thank you!

Signature · 9 lines

Jens Benecke             /* Nobody will ever see this error message :-) */
                            panic("Cannot initialize video hardware\n");
                                  -- Linux 2.0.38,  arch/m68k/atari/atafb.c
http://www.hitchhikers.de/ - Die kostenlose Mitfahrzentrale f�r ganz Europa
_______________________________________________
linux-lvm mailing list
[email protected]
http://lists.sistina.com/mailman/listinfo/linux-lvm
read the LVM HOW-TO at http://www.sistina.com/lvm/Pages/howto.html

Quoted 3 lines · Heinz J. Mauelshagen

Jens,
you need to run pvcreate on the invalid PV before running vgcfgrestore.

I did so.

Now I have a PV that is associated to an unknown VG, and vgcfgrestore says

vgcfgrestore -- size of physical volume /dev/hdc1 differs from backup

although 'vgcfgrestore -n data -ll' and 'pvdata -PP /dev/hdc1' show EXACTLY the same amount of extents (used/free/..), and GB. And hdc1 certainly didn't change in size.

Strange: BOTH vgcfgrestore and pvdata -PP know that the PVs belong to the VG "data", but vgscan doesn't find it, and pvscan says both belong to an unknown VG and I should run vgscan.

So, I am a little further but not yet there =;)

Thanks for the help!

Signature · 9 lines

Jens Benecke             /* Nobody will ever see this error message :-) */
                            panic("Cannot initialize video hardware\n");
                                  -- Linux 2.0.38,  arch/m68k/atari/atafb.c
http://www.hitchhikers.de/ - Die kostenlose Mitfahrzentrale f�r ganz Europa
_______________________________________________
linux-lvm mailing list
[email protected]
http://lists.sistina.com/mailman/listinfo/linux-lvm
read the LVM HOW-TO at http://www.sistina.com/lvm/Pages/howto.html

Just an update on this:

Jens and I talked on IRC (irc.openprojects.net #lvm) about the problem. Turns out it was our old friend the missing UUID again. We fixed it by manually adding the missing UUIDs in using a hexeditor on a copy of the metadata, then using dd to copy the metadata back onto the disk. Nothing else seemed to work.

Regards,

Signature · 14 lines

AJ Lewis
Sistina Software Inc.                  Voice:  612-638-0500
1313 5th St SE, Suite 111              Fax:    612-638-0500
Minneapolis, MN 55414                  E-Mail: [email protected]
http://www.sistina.com

Current GPG fingerprint = 3B5F 6011 5216 76A5 2F6B  52A0 941E 1261 0029 2648
Get my key at: http://www.sistina.com/~lewis/gpgkey
 (Unfortunately, the PKS-type keyservers do not work with multiple sub-keys)

-----Begin Obligatory Humorous Quote----------------------------------------
Quitters never win, and winners never quit,
but those who never quit AND never win are idiots.
-----End Obligatory Humorous Quote------------------------------------------

Recompiling an XFS 2.4.5 tree, after patching with LVM 0.9.1b7, which went fine except for the drivers/md/Makefile( stuff already exists in this tree), no other changes, checked the .rej file, I get the following failure when trying to make the kernel:

drivers/md/mddev.o: In function `__update_hardblocksize':
drivers/md/mddev.o(.text+0xaba1): undefined reference to get_hardblocksize' drivers/md/mddev.o(.text+0xabed): undefined reference to get_hardblocksize'
make[1]: *** [kallsyms] Error 1
make[1]: Leaving directory `/usr/src/linux-2.4-xfs/linux'
make: *** [vmlinux] Error 2

Any ideas?

Signature · 12 lines

Austin Gonyou
Systems Architect, CCNA
Coremetrics, Inc.
Phone: 512-796-9023
email: [email protected]


_______________________________________________
linux-lvm mailing list
[email protected]
http://lists.sistina.com/mailman/listinfo/linux-lvm
read the LVM HOW-TO at http://www.sistina.com/lvm/Pages/howto.html

Quoted 1 line · Austin Gonyou

Recompiling an XFS 2.4.5 tree, after patching with LVM 0.9.1b7, which

went

Quoted 23 lines

fine except for the drivers/md/Makefile( stuff already exists in this tree), no other changes, checked the .rej file, I get the following failure when trying to make the kernel:

drivers/md/mddev.o: In function `__update_hardblocksize':
drivers/md/mddev.o(.text+0xaba1): undefined reference to get_hardblocksize' drivers/md/mddev.o(.text+0xabed): undefined reference to get_hardblocksize'
make[1]: *** [kallsyms] Error 1
make[1]: Leaving directory `/usr/src/linux-2.4-xfs/linux'
make: *** [vmlinux] Error 2

Any ideas?

--
Austin Gonyou
Systems Architect, CCNA
Coremetrics, Inc.
Phone: 512-796-9023
email: [email protected]

Quoted 5 lines

_______________________________________________
linux-lvm mailing list
[email protected]
http://lists.sistina.com/mailman/listinfo/linux-lvm
read the LVM HOW-TO at http://www.sistina.com/lvm/Pages/howto.html

I had the same problem a couple of nights ago.

Replace get_hardblocksize with get_hardsect_size there should be 2 instances of it in lvm.c

Enjoy

Quoted 9 lines · Jens Benecke

Quoted 5 lines · Jens Benecke

Hello,
I upgraded to 2.2.4 (0.9.1-beta2) from 2.2.18 (LVM 0.9) and this seems to have killed one of my PVs. I tried the 0.9 utilities and updated to 0.9.1-beta7 when I noticed the problems. No change. Booting back to 2.2.18 didn't help, the PV was gone there as well.

follow-up:
vgcfgrestore doesn't work, because LVM thinks my hdc1 is no PV. Here's the
output: (...)

I think I found a way: I have found a hex-dump backup of the first megabyte of the disk in - apparently - working order.

So: hdc1 was the broken PV, it's the only partition on hdc, and I have a file called hdc-first-MB-hexdump.

Does this help? ;)

The question is, does this already contain file system data or is this still all LVM structure data?

I'm quite sure that the LVM structure didn't
change since this file was created, only (of course) the data changed on the filesystem contained in the one (single) LV on data=(hdc1,hda1).

Signature · 10 lines

-- 
Jens Benecke             /* Nobody will ever see this error message :-) */
                            panic("Cannot initialize video hardware\n");
                                  -- Linux 2.0.38,  arch/m68k/atari/atafb.c
http://www.hitchhikers.de/ - Die kostenlose Mitfahrzentrale f�r ganz Europa
_______________________________________________
linux-lvm mailing list
[email protected]
http://lists.sistina.com/mailman/listinfo/linux-lvm
read the LVM HOW-TO at http://www.sistina.com/lvm/Pages/howto.html