[linux-lvm] same PV detected twice -> VGs mix up
Eckhard Neber
eckhard.neber at web.de
Thu Aug 7 10:33:02 UTC 2003
Hello,
I have a strange problem here.
During installation we misconfigured our external raid somehow to
provide the same lodical drive as lun0 as well as lun1. Linux detects
both and maps them as sdb and sdc.
Lvm activates one of them and everything works well, so we decided to
leave it that way as the server was already in use when we discovered
this problem.
Now we wanted to add two more ATA discs to the server that were used in
a different computer as one logical volume (data120).
Here is the output of some of the tools:
5#e:~ pvscan
pvscan -- reading all physical volumes (this may take a while...)
pvscan -- inactive PV "/dev/sdb" of VG "raid" [1.01 TB / 173.03 GB free]
pvscan -- ACTIVE PV "/dev/sdc" of VG "raid" [1.01 TB / 173.03 GB free]
pvscan -- inactive PV "/dev/hdc" of VG "data120" [57.26 GB / 0 free]
pvscan -- inactive PV "/dev/hdd" of VG "data120" [57.26 GB / 0 free]
pvscan -- total: 4 [2.13 TB] / in use: 4 [2.13 TB] / in no VG: 0 [0]
5#e:~ vgscan
vgscan -- reading all physical volumes (this may take a while...)
vgscan -- found active volume group "raid"
vgscan -- found active volume group "data120"
vgscan -- "/etc/lvmtab" and "/etc/lvmtab.d" successfully created
vgscan -- WARNING: This program does not do a VGDA backup of your volume
groups
5#e:~ lvscan
lvscan -- ACTIVE "/dev/raid/soft" [200 GB]
lvscan -- ACTIVE "/dev/raid/mail" [10 GB]
lvscan -- ACTIVE "/dev/raid/backup" [600 GB]
lvscan -- ACTIVE "/dev/raid/users" [50 GB]
lvscan -- ACTIVE "/dev/raid/soft" [200 GB]
lvscan -- ACTIVE "/dev/raid/mail" [10 GB]
lvscan -- ACTIVE "/dev/raid/backup" [600 GB]
lvscan -- ACTIVE "/dev/raid/users" [50 GB]
lvscan -- 8 logical volumes with 1.68 TB total in 2 volume groups
lvscan -- 8 active logical volumes
It is remarkable that the LVs of raid show up twice. What is really
strange is that if I mount /dev/data120/backup, which is the LV of the
ATA disks, it works, but I get the contents of /dev/raid/soft.
How can it be that these two VGs get mixed up? Is it because there are
identical names?
A quick solution could probably be to restrict the kernel to probe only
for lun0. I'd like to avoid to change the configuration of the raid
because it would require to fully back it up (in case that something
goes wrong).
We are using a kernel that was build from SuSE's sorces
linux-2.4.20.SuSE-86.
I don't dare to mount and modify anything with data120 as I fear that it
could damage the other LVs.
Any help would be appreciated.
Thanks in advance
Eckhard Neber
--
-----------------------------------------------------------------
Eckhard Neber Tel: ++49-(0)731-50-26354
Abt. Mikrowellentechnik Fax: ++49-(0)731-50-26359
Universitaet Ulm e-mail: eckhard.neber at web.de
89069 Ulm
-----------------------------------------------------------------
More information about the linux-lvm
mailing list