Skip to contentSkip to conversation

64K limit for VG/LV

linux-lvm Dieter Stueken 2 participants 3 messages Jan 22, 2002 – Jan 23, 2002

lore
Share & tools
lore
#0 lore raw

I just merged two volume groups without any error message, but the result seems to have more than 65535 LE now. The only effect until now is, that the number of free PE becomes negative and is displayed as an unsigned huge value:

--- Volume group ---

VG Name               vg0
VG Access             read/write
VG Status             available/resizable
VG #                  0
MAX LV                256
Cur LV                9
Open LV               2
MAX LV Size           255.99 GB
Max PV                256
Cur PV                4
Act PV                4
VG Size               286.33 GB
PE Size               4.00 MB
Total PE              73300
Alloc PE / Size       97200 / 379.69 GB
Free  PE / Size       4294943396 / 1.91 TB
VG UUID               QDKBpV-xwqr-gO3b-VKiA-Jaf4-grmm-pWEY32

My VG consists of 4 x 75 GB IBM-DTLA-307075. Should I worry now? If there is any limit of 65536 I might suspect some 16bit value for PE addressing. If no overflow is checked, a wrong PE may be addressed beyond LE 65536 and my VG possibly gets destroyed :-(

But I found some hints in the mailing archive, saying that the 64K limit does not affect the VG itself but only each LV inside. Thus as long as each of my LV is <256G I don't have to worry, and the reported "Free PE" might be just cosmetics.

Besides, vgdisplay says: "Logical Volume Manager 1.0.1-rc2" On SuSE 7.3 running 2.4.16.

If I want to increase my LE size, I think I have to create a new VG. Is there any tool to copy my LVs to the new VG other than "dd"?

Dieter.

Signature · 4 lines
Dieter St�ken, con terra GmbH, M�nster
    [email protected]         [email protected]
    http://www.conterra.de/     http://qgp.uni-muenster.de/~stueken
    (0)251-980-2027             (0)251-83-334974
#1 lore raw
Quoted 24 lines · Dieter Stueken

I just merged two volume groups without any error message, but the result seems to have more than 65535 LE now. The only effect until now is, that the number of free PE becomes negative and is displayed as an unsigned huge value:

--- Volume group ---

VG Name               vg0
VG Access             read/write
VG Status             available/resizable
VG #                  0
MAX LV                256
Cur LV                9
Open LV               2
MAX LV Size           255.99 GB
Max PV                256
Cur PV                4
Act PV                4
VG Size               286.33 GB
PE Size               4.00 MB
Total PE              73300
Alloc PE / Size       97200 / 379.69 GB
Free  PE / Size       4294943396 / 1.91 TB
VG UUID               QDKBpV-xwqr-gO3b-VKiA-Jaf4-grmm-pWEY32
Quoted 7 lines

My VG consists of 4 x 75 GB IBM-DTLA-307075. Should I worry now? If there is any limit of 65536 I might suspect some 16bit value for PE addressing. If no overflow is checked, a wrong PE may be addressed beyond LE 65536 and my VG possibly gets destroyed :-(

But I found some hints in the mailing archive, saying that the 64K limit does not affect the VG itself but only each LV inside.

Dieter,
that's correct. It is just a per LV limit for the max amount of its Logical Extents (LEs).

Quoted 2 lines

Thus as long as each of my LV is <256G I don't have to worry, and the reported "Free PE" might be just cosmetics.

Cosmetics is a nice way to put it :-)
We probably have missed that one, because for a VG with such a large capacity we choosed larger PE sizes.

Quoted 6 lines

Besides, vgdisplay says: "Logical Volume Manager 1.0.1-rc2" On SuSE 7.3 running 2.4.16.

If I want to increase my LE size, I think I have to create a new VG. Is there any tool to copy my LVs to the new VG other than "dd"?

cat ;-)
No, sorry. None included in LVM.

Quoted 13 lines

Dieter.

Signature · 10 lines
Dieter Stüken, con terra GmbH, Münster
    [email protected]         [email protected]
    http://www.conterra.de/     http://qgp.uni-muenster.de/~stueken
    (0)251-980-2027             (0)251-83-334974

_______________________________________________
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
Signature · 15 lines
Regards,
Heinz    -- The LVM Guy --

*** Software bugs are stupid.
    Nevertheless it needs not so stupid people to solve them ***

=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-

Heinz Mauelshagen                                 Sistina Software Inc.
Senior Consultant/Developer                       Am Sonnenhang 11
                                                  56242 Marienrachdorf
                                                  Germany
[email protected]                           +49 2626 141200
                                                       FAX 924446
=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-
#2 lore raw
Quoted 9 lines · Heinz J . Mauelshagen

that's correct. It is just a per LV limit for the max amount of its Logical Extents (LEs).

Quoted 2 lines

Thus as long as each of my LV is <256G I don't have to worry, and the reported "Free PE" might be just cosmetics.

Cosmetics is a nice way to put it :-)
We probably have missed that one, because for a VG with such a large capacity we choosed larger PE sizes.

but I still worry a bit about my data. The reason for the wrong "Free PE"
reported seems not to be the 64k limit, but the fact that some of my LVs are reported/counted twice!

-> vgdisplay -v vg0 | fgrep "LV Name"

LV Name                /dev/vg0/root
LV Name                /dev/vg0/win2k
LV Name                /dev/vg0/ortho
LV Name                /dev/vg0/praktikum
LV Name                /dev/vg0/swap
LV Name                /dev/vg0/mirror
LV Name                /dev/vg0/disk_huetti
LV Name                /dev/vg0/SuSE-7.3
LV Name                /dev/vg0/mirror          <<<<<
LV Name                /dev/vg0/disk_huetti     <<<<<

Thus I still fear of some evil inconsistencies for my data. Is there any tool available like "e2fsck" for LVM to resolve such a problem?

Dieter.