Resizing PV
Is it possible to reduce the size of a PV, freeing some space at the end? I can make sure that the end is unused, of course.
Signature · 2 lines
M�ns Rullg�rd [email protected]
Quoted 3 lines · Måns Rullgård
Is it possible to reduce the size of a PV, freeing some space at the end? I can make sure that the end is unused, of course.
No such command implemented yet.
Hack:
you can use LVM2 and change the text format of the metadata by hand before running vgcfgrestore to store it back onto the PVs.
If you want to follow this non-supported path, you want to take a copy of the original metadata file in /etc/lvm/backup/ to be able to back out gain.
Quoted 11 lines
Signature · 10 lines
-- Måns Rullgård [email protected] _______________________________________________ linux-lvm mailing list [email protected] http://lists.sistina.com/mailman/listinfo/linux-lvm read the LVM HOW-TO at http://tldp.org/HOWTO/LVM-HOWTO/
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
=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-Quoted 5 lines · Heinz J . Mauelshagen
Quoted 2 lines
Is it possible to reduce the size of a PV, freeing some space at the end? I can make sure that the end is unused, of course.
No such command implemented yet.
Quoted 7 lines
Hack:
you can use LVM2 and change the text format of the metadata by hand before running vgcfgrestore to store it back onto the PVs.If you want to follow this non-supported path, you want to take a copy of the original metadata file in
/etc/lvm/backup/to be able to back out gain.
I'm already using lvm2. I noticed the pvresize command, but it just says it's not implemented yet. Any chance it will be some time soon?
As for editing the metadata, I'm a little confused about the meaning of some fields. After running vgcfgbackup, I have this in the file:
physical_volumes {
pv0 {
id = "eZV4mb-jk4z-pyfm-tGbJ-mBSW-drRT-ZHtgmn"
device = "/dev/discs/disc0/part2" # Hint only
status = ["ALLOCATABLE"]
pe_start = 65920
pe_count = 1189 # 37.1562 Gigabytes
}
}What is the meaning of the pe_start field? If I wanted to reduce the size by, say, 1 GB, would it be correct to reduce pe_count by 32, the PE size being 32 MB?
Signature · 2 lines
M�ns Rullg�rd [email protected]
Quoted 11 lines · Måns Rullgård
Quoted 9 lines · Heinz J . Mauelshagen
Quoted 2 lines
Is it possible to reduce the size of a PV, freeing some space at the end? I can make sure that the end is unused, of course.
No such command implemented yet.
Hack:
you can use LVM2 and change the text format of the metadata by hand before running vgcfgrestore to store it back onto the PVs.
Quoted 6 lines
Quoted 3 lines
If you want to follow this non-supported path, you want to take a copy of the original metadata file in
/etc/lvm/backup/to be able to back out gain.I'm already using lvm2. I noticed the pvresize command, but it just says it's not implemented yet. Any chance it will be some time soon?
We've got higher priority stuff in the works. Probably later this year.
Quoted 16 lines
As for editing the metadata, I'm a little confused about the meaning of some fields. After running vgcfgbackup, I have this in the file:
physical_volumes { pv0 { id = "eZV4mb-jk4z-pyfm-tGbJ-mBSW-drRT-ZHtgmn" device = "/dev/discs/disc0/part2" # Hint only status = ["ALLOCATABLE"] pe_start = 65920 pe_count = 1189 # 37.1562 Gigabytes } }
Quoted 3 lines
What is the meaning of the
pe_startfield? If I wanted to reduce the size by, say, 1 GB, would it be correct to reducepe_countby 32, the PE size being 32 MB?
pe_start is the sector offset into the device.
Yes, you need to reduce pe_count by 32 making it 1157 and vgcfgrestore. You need to have recent LVM2 code so that vgcfgrestore updates the pe_count properly (and of course, the 32 PEs need to be empty).
Quoted 11 lines
Signature · 10 lines
-- Måns Rullgård [email protected] _______________________________________________ linux-lvm mailing list [email protected] http://lists.sistina.com/mailman/listinfo/linux-lvm read the LVM HOW-TO at http://tldp.org/HOWTO/LVM-HOWTO/
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
=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-Quoted 1 line · Heinz J . Mauelshagen
pe_startis the sector offset into the device.
OK, I thought it might be, but I wasn't sure.
Quoted 2 lines
Yes, you need to reduce
pe_countby 32 making it 1157 and vgcfgrestore. You need to have recent LVM2 code so that vgcfgrestore updates thepe_count
How recent? Is this new enough?
LVM version: 2.00.03 (2003-07-13)
Library version: 1.00.02-ioctl (2003-07-12)
Driver version: 4.0.0
I'm using whatever device-mapper is in Linux 2.6.0-test5.
Quoted 1 line
properly (and of course, the 32 PEs need to be empty).
Signature · 2 lines
M�ns Rullg�rd [email protected]
Quoted 6 lines · Måns Rullgård
Quoted 1 line · Heinz J . Mauelshagen
pe_startis the sector offset into the device.OK, I thought it might be, but I wasn't sure.
Quoted 8 lines
Quoted 2 lines
Yes, you need to reduce
pe_countby 32 making it 1157 and vgcfgrestore. You need to have recent LVM2 code so that vgcfgrestore updates thepe_countHow recent? Is this new enough?
LVM version: 2.00.03 (2003-07-13)
Library version: 1.00.02-ioctl (2003-07-12)
Driver version: 4.0.0
Should do it.
Quoted 2 lines
I'm using whatever device-mapper is in Linux 2.6.0-test5.
That's ok. device-mapper isn't involved in such
userspace metadata updates to the PVs.
Quoted 13 lines
Quoted 2 lines
properly (and of course, the 32 PEs need to be empty).
Signature · 9 lines
Måns Rullgård [email protected] _______________________________________________ linux-lvm mailing list [email protected] http://lists.sistina.com/mailman/listinfo/linux-lvm read the LVM HOW-TO at http://tldp.org/HOWTO/LVM-HOWTO/
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
=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-Quoted 8 lines · Heinz J . Mauelshagen
Quoted 8 lines
Quoted 1 line
pe_startis the sector offset into the device.OK, I thought it might be, but I wasn't sure.
Quoted 3 lines
Yes, you need to reduce
pe_countby 32 making it 1157 and vgcfgrestore. You need to have recent LVM2 code so that vgcfgrestore updates thepe_count
Quoted 7 lines
Quoted 5 lines
How recent? Is this new enough?
LVM version: 2.00.03 (2003-07-13)
Library version: 1.00.02-ioctl (2003-07-12)
Driver version: 4.0.0Should do it.
It seemed to work. Thanks for the help.
Signature · 2 lines
M�ns Rullg�rd [email protected]