I bit-copied a backup LVM disk using dd from a Knoppix LiveCD (based on
Debian) and wound up with duplicate volume and uuids, meaning I couldn't
mount both the main and backup volumes simultaneously at the same time for
subsequent backups. Of course, vgrename is ineffective to allow
simultaneous mounting without changing the uuids.
My system hardware is:
Biostar Geforce 6100 AM2, with the usual IDE setup and 2 SATA channels.
Note that if a drive is plugged into the JSATA2 connector in the absence of
anything plugged into JSATA1, it appears to the OS as SDA, NOT SDB. I don't
know if that's normal for SATA or not. It is disconcerting.
My main and backup hard drives are Samsung SATA2 HD403LJ, the backup drive
is plugged into a mobile rack.
NOTE: I'd really like to see a specific Linux LVM command that will change
uuids for an LVM volume without manual intervention.
However, the workaround in the "HOW TO ALTER SATA LVM VOLUME UUIDs" section
below will keep SATA users going in the meantime, readers in a hurry should
scroll down to that part of this post.
When I used the recommended method described in
Re: Altering a Physical Volume's UUID [solved]
( http://www.spinics.net/lists/lvm/msg16090.html ) involving extracting the
uuids via vcfgbackup and incrementing them all by 1, creating a physical
volume via pvcreate, then writing it back via vgcfgrestore from a Knoppix
LiveCD on a pair of SATA hard drives.
It failed as below:
original terrarium.cfg file:
# Generated by LVM2: Thu Aug 16 16:24:03 2007
contents = "Text Format Volume Group"
version = 1
description = "vgcfgbackup -f terrarium.cfg terrarium"
creation_host = "terrarium" # Linux terrarium 2.6.21-2-k7 #1 SMP Wed Jul 11
04:29:08 UTC 2007 i686
creation_time = 1187306643 # Thu Aug 16 16:24:03 2007
terrarium {
id = "gX4iKn-9GCb-TKSG-mKQE-dqc0-Kt1Q-FAo1Ne"
seqno = 3
status = ["RESIZEABLE", "READ", "WRITE"]
extent_size = 8192 # 4 Megabytes
max_lv = 0
max_pv = 0
physical_volumes {
pv0 {
id = "5fmUbT-7liU-DUoI-6C0z-98Pq-uDE4-d3zkuL"
device = "/dev/sda5" # Hint only
status = ["ALLOCATABLE"]
pe_start = 384
pe_count = 95327 # 372.371 Gigabytes
}
}
logical_volumes {
root {
id = "hZBRED-nxBj-v1xU-8TFP-le4B-ye16-fWyn5e"
status = ["READ", "WRITE", "VISIBLE"]
segment_count = 1
segment1 {
start_extent = 0
extent_count = 94664 # 369.781 Gigabytes
type = "striped"
stripe_count = 1 # linear
stripes = [
"pv0", 0
]
}
}
swap_1 {
id = "dQ248n-d62X-G6Dn-E9iJ-aHSA-cFbp-4hFuYW"
status = ["READ", "WRITE", "VISIBLE"]
segment_count = 1
segment1 {
start_extent = 0
extent_count = 663 # 2.58984 Gigabytes
type = "striped"
stripe_count = 1 # linear
stripes = [
"pv0", 94664
]
}
}
}
}
After incrementing the uuids by one, I got:
root@1[knoppix]# pvcreate -ff --uuid
gX4iKn-9GCb-TKSG-mKQE-dqc0-Kt1Q-FAo1Nf --restorefile terrarium.cfg /dev/loop2
Found duplicate PV 5fmUbT7liUDUoI6C0z98PquDE4d3zkuL: using /dev/sda5 not
/dev/loop0
Found duplicate PV 5fmUbT7liUDUoI6C0z98PquDE4d3zkuL: using /dev/loop0
not /dev/sda5
Found duplicate PV 5fmUbT7liUDUoI6C0z98PquDE4d3zkuL: using /dev/sda5 not
/dev/loop0
Found duplicate PV 5fmUbT7liUDUoI6C0z98PquDE4d3zkuL: using /dev/loop0
not /dev/sda5
Found duplicate PV 5fmUbT7liUDUoI6C0z98PquDE4d3zkuL: using /dev/sda5 not
/dev/loop0
Couldn't find device with uuid '5fmUbT-7liU-DUoI-6C0z-98Pq-uDE4-d3zkuM'.
Can't find uuid gX4iKn-9GCb-TKSG-mKQE-dqc0-Kt1Q-FAo1Nf in backup file
terrarium.cfg
While "can't find" is what I expected, nothing wrote back when I tried
vgcfgrestore, all I got was an error message.
There's at least one other variation of this basic method that doesn't
involve loop mounting, and it also fails. What appears to be going on is
that at least in an SATA vgcfgrestore , the copy one tried to modify is
checked against the original and fails.
HOW TO ALTER SATA LVM VOLUME UUIDs
I got this working by using LVM commands to change uuids for PV
(physical_volume) and for VG (volume group), then manually incremented the
LV (logical_volume) uuids. As far as I can see, this should work for either
IDE or SATA HDs. The original uuids are different from the above because I
did some experimentation with this before submitting this solution to the list.
In the example below, replace "terrarium" with the actual volume name
you're using, /dev/sda5 with the actual disk partition you're trying to
use. In-line comments marked with *** .
Only the drive you are trying to change UUIDs for should be plugged in at
this point.
DO NOT LVM-MOUNT IT USING vgchange -ay .
root@1[knoppix]# pvchange -u /dev/sda5
Physical volume "/dev/sda5" changed
1 physical volume changed / 0 physical volumes not changed
root@1[knoppix]# vgchange -u terrarium
Volume group "terrarium" successfully changed
root@1[knoppix]# vgcfgbackup -f terrarium.cfg terrarium
Volume group "terrarium" successfully backed up.
*** This creates a LVM configuration file that can be manually edited
containing the above uuid changes produced by pvchange -u and vgchange -u.
root@1[knoppix]# nano terrarium.cfg
*** At this point, open with a text editor (in this example, nano) and
increment the last character in each of the "logical_volumes" uuids by 1 ,
f becomes g, 4 becomes 5, etc. In the example configuration files below,
this would be the uuids for root and swap_1, e.g.
root {
id = "hZBRED-nxBj-v1xU-8TFP-le4B-ye16-fWyn5f"
becomes id = "hZBRED-nxBj-v1xU-8TFP-le4B-ye16-fWyn5g"
DO NOTHING WITH THE ID and PV uuids at this stage. Apparently, only the
volume ID and PV ID uuids are checked against the original.
root@1[knoppix]# pvcreate -ff --uuid jEzDgz-S7nH-oTeU-nmWk-kTSM-tmpo-7O4Ryp
--restorefile terrarium.cfg /dev/sda5
Can't find uuid jEzDgz-S7nH-oTeU-nmWk-kTSM-tmpo-7O4Ryp in backup file
terrarium.cfg
*** This error message is usual, and not a problem as long as vgcfgrestore
returns as below:
root@1[knoppix]# vgcfgrestore -f terrarium.cfg terrarium
Restored volume group terrarium
*** This writes the LVM configuration file back to the disk's LVM
configuration.-
*** At this point, you can use vgrename:
# vgrename terrarium terrarium-back
*** At this point, you have successfully changed the uuids and volume name
now have a backup drive volume that will mount simultaneously with your
main drive.
*** original cfg
-----------------------------------------------------
# Generated by LVM2: Sat Aug 18 09:23:35 2007
segment1 {
contents = "Text Format Volume Group"
version = 1
description = "vgcfgbackup -f terrarium.cfg terrarium"
creation_host = "Knoppix" # Linux Knoppix 2.6.17 #4 SMP PREEMPT
Wed May 1$
creation_time = 1187443415 # Sat Aug 18 09:23:35 2007
terrarium {
id = "tOezwV-ZnrW-wPM9-I3D1-4zjj-KDY2-Ikz1O4"
seqno = 14
status = ["RESIZEABLE", "READ", "WRITE"]
extent_size = 8192 # 4 Megabytes
max_lv = 0
max_pv = 0
physical_volumes {
pv0 {
id = "GOe7U8-fU7p-xC0V-btvK-q1mh-sotC-0qqzht"
device = "/dev/sda5" # Hint only
status = ["ALLOCATABLE"]
pe_start = 384
pe_count = 95327 # 372.371 Gigabytes
}
}
logical_volumes {
root {
id = "hZBRED-nxBj-v1xU-8TFP-le4B-ye16-fWyn5f"
status = ["READ", "WRITE", "VISIBLE"]
segment_count = 1
segment1 {
start_extent = 0
extent_count = 94664 # 369.781 Gigabytes
type = "striped"
stripe_count = 1 # linear
stripes = [
"pv0", 0
]
}
}
swap_1 {
id = "dQ248n-d62X-G6Dn-E9iJ-aHSA-cFbp-4hFuYX"
status = ["READ", "WRITE", "VISIBLE"]
segment_count = 1
segment1 {
start_extent = 0
extent_count = 663 # 2.58984 Gigabytes
type = "striped"
stripe_count = 1 # linear
stripes = [
"pv0", 94664
]
}
}
}
}
-----------------------------------------------------
*** AFTER uuid change to verify that it worked (i.e. are the IDs changed
from the original?):
root@1[knoppix]# vgcfgbackup -f terrarium.cfg terrarium
Volume group "terrarium" successfully backed up.
nano terrarium.cfg
*** modified cfg
---------------------------------------------------
description = "vgcfgbackup -f terrarium.cfg terrarium"
creation_host = "Knoppix" # Linux Knoppix 2.6.17 #4 SMP PREEMPT
Wed May 1$
creation_time = 1187444082 # Sat Aug 18 09:34:42 2007
terrarium {
id = "jEzDgz-S7nH-oTeU-nmWk-kTSM-tmpo-7O4Ryp"
seqno = 17
status = ["RESIZEABLE", "READ", "WRITE"]
extent_size = 8192 # 4 Megabytes
max_lv = 0
max_pv = 0
physical_volumes {
pv0 {
id = "lk7nC4-13Bo-YbhF-PiPD-f146-MP6V-lcArz1"
device = "/dev/sda5" # Hint only
status = ["ALLOCATABLE"]
pe_start = 384
pe_count = 95327 # 372.371 Gigabytes
}
}
logical_volumes {
root {
id = "hZBRED-nxBj-v1xU-8TFP-le4B-ye16-fWyn5g"
status = ["READ", "WRITE", "VISIBLE"]
segment_count = 1
segment1 {
start_extent = 0
extent_count = 94664 # 369.781 Gigabytes
type = "striped"
stripe_count = 1 # linear
stripes = [
"pv0", 0
]
}
}
swap_1 {
id = "dQ248n-d62X-G6Dn-E9iJ-aHSA-cFbp-4hFuYY"
status = ["READ", "WRITE", "VISIBLE"]
segment_count = 1
segment1 {
start_extent = 0
extent_count = 663 # 2.58984 Gigabytes
type = "striped"
stripe_count = 1 # linear
stripes = [
"pv0", 94664
]
}
}
}
}
--------------------------------
at this point, you can rename the volume.
member The Internet Society (ISOC), The HTML Writers Guild.
"You can't have in a democracy various groups with arms - you have to have
the state with a monopoly on power." Condoleeza Rice, US Secretary of State
Personal Website http://www.ecis.com/~alizard
business Website http://www.reptilelabs.com
backup address (if ALL else fails) alizard@gmail.com
PGP 8.0 key available by request or keyserver. Download PGP from:
http://www.pgpi.org for e-mail privacy.
Disaster prep info: http://www.ecis.com/~alizard/y2k.html
***Looking for INTELLIGENT new technology public policy
alternatives?*** http://www.ecis.com/~alizard/technology.html
_______________________________________________
linux-lvm mailing list
linux-lvm@redhat.com
https://www.redhat.com/mailman/listinfo/linux-lvm
read the LVM HOW-TO at http://tldp.org/HOWTO/LVM-HOWTO/
