Endless loop running vgscan

Hello,

I try to get my LVM1 array (hd[e,f,g,h]1) working with kernel 2.6.0
and LVM2. After installing LVM2.2.00.08 I try to run vgscan, it scans the devices and founds the LVM1 labels. But then it runs in a endless loop:
...........
Ignoring duplicate PV on /dev/hdg1
Found /dev/hdf1 in VG raid_vg
Ignoring duplicate PV on /dev/hdf1
Found /dev/hdh1 in VG raid_vg
Ignoring duplicate PV on /dev/hdh1
Found /dev/hde1 in VG raid_vg
Ignoring duplicate PV on /dev/hde1
Found /dev/hdg1 in VG raid_vg
Ignoring duplicate PV on /dev/hdg1
Found /dev/hdf1 in VG raid_vg
Ignoring duplicate PV on /dev/hdf1
...........

until it get killed by the kernel (Out of Memory).

The VG was created with LVM 1.0.5.

Thanks for any help,
Sebastian

Set up a filter in lvm.conf so that LVM2 only sees the RAID devices and ignores the constituent devices that make up the RAID devices.

The LVM2 logic isn't clever enough yet always to work out the correct device to use. This will get fixed.

Alasdair

[email protected]

Sorry, RAID is a little bit wrong in my case. The harddiscs running on
a Promise RAID-Controller but the RAID-function is disabled and it's used as a normal IDE controller. I appended my configuration.

I've also this problem if I use this filter:
filter = [ "a/hd[e-h]1$/", "r/.*/" ]

Regards,
Sebastian

--- Volume group ---

VG Name               raid_vg
VG Access             read/write
VG Status             available/resizable
VG #                  0
MAX LV                256
Cur LV                2
Open LV               2
MAX LV Size           255.99 GB
Max PV                256
Cur PV                4
Act PV                4
VG Size               38.15 GB
PE Size               4 MB
Total PE              9766
Alloc PE / Size       9760 / 38.12 GB
Free  PE / Size       6 / 24 MB
VG UUID               none

--- Logical volume ---

LV Name                /dev/raid_vg/ftp_lv
VG Name                raid_vg
LV Write Access        read/write
LV Status              available
LV #                   1
# open                 1
LV Size                25 GB
Current LE             6400
Allocated LE           6400
Stripes                4
Stripe size (KByte)    64
Allocation             next free
Read ahead sectors     1024
Block device           58:0

--- Logical volume ---

LV Name                /dev/raid_vg/raid_lv
VG Name                raid_vg
LV Write Access        read/write
LV Status              available
LV #                   2
# open                 1
LV Size                13.12 GB
Current LE             3360
Allocated LE           3360
Stripes                4
Stripe size (KByte)    64
Allocation             next free
Read ahead sectors     1024
Block device           58:1

--- Physical volumes ---

PV Name (#)           /dev/hde1 (1)
PV Status             available / allocatable
Total PE / Free PE    2440 / 0

PV Name (#)           /dev/hdf1 (2)
PV Status             available / allocatable
Total PE / Free PE    2442 / 2

PV Name (#)           /dev/hdg1 (3)
PV Status             available / allocatable
Total PE / Free PE    2442 / 2

PV Name (#)           /dev/hdh1 (4)
PV Status             available / allocatable
Total PE / Free PE    2442 / 2

Is that the wrong way round? Accept the md devices instead of the hd ones if your mds are formed from hds.

Alasdair

I have no md's. My LVM VG consists of four PVs, hd[e-h]1.

Sebastian

If you send me (not the mailing list) the debug output file (truncated after it's started looping and then compress it) from vgscan I'll try to see what's happening (later this week).
'pvs -v' output may also be useful.
[See WHATS_NEW file for how to edit lvm.conf to get a debug log file]

Alasdair