Skip to contentSkip to conversation

LVM can't detect raid-1 mirrors?

linux-lvm Hadmut Danisch 2 participants 4 messages Jul 18, 2005 – Jul 19, 2005

lore
Share & tools
lore
#0 lore raw

Hi,

I need some help with a strange lvm problem:

I want to put LVM on a Linux Raid-1 device. I already did this several times successfully. But this time I have a strange problem:

The raid device is /dev/md3, and it should consist of the two partitions /dev/hda3 and /dev/hdc3.

I created the raid device with just /dev/hdc3, second missing, then pvcreate /dev/md3 and created the volume group 'raid'. Works well.

pvscan correctly shows the raid device instead of /dev/hdc3:

sodom# pvscan

  PV /dev/hdc4   VG nord   lvm2 [58,59 GB / 33,59 GB free]
  PV /dev/md3    VG raid   lvm2 [54,35 GB / 21,35 GB free]

Total: 2 [112,94 GB] / in use: 2 [112,94 GB] / in no VG: 0 [0 ]

(ignore hdc4, it is just another lvm because hdc is larger than hda, there is no problem with hdc4)

But as soon as I add /dev/hda3 to the raid system (mdad /dev/md3 -a /dev/hda3), lvm confuses the partitions. It does not detect /dev/md3 anymore and complains that there are two different partitions with the same ID:

sodom# pvscan
Found duplicate PV 39jeBxsxpKtJrwP6FN3Yvq6zB7khU6Pa: using /dev/hdc3 not /dev/hda3

  PV /dev/hdc4   VG nord   lvm2 [58,59 GB / 33,59 GB free]
  PV /dev/hda3   VG raid   lvm2 [54,35 GB / 21,35 GB free]

Total: 2 [112,94 GB] / in use: 2 [112,94 GB] / in no VG: 0 [0 ]

The volume group can't then be used. As soon as I take out /dev/hda3 and wipe the first sectors, everything is ok again.

Any idea what's going wrong here?

Just in case it matters, the partition tables:

sodom# sfdisk -l /dev/hda

Disk /dev/hda: 116301 cylinders, 16 heads, 63 sectors/track
Warning: The partition table looks like it was made
for C/H/S=*/255/63 (instead of 116301/16/63).
For this listing I'll assume that geometry.
Units = cylinders of 8225280 bytes, blocks of 1024 bytes, counting from 0

   Device Boot Start     End   #cyls    #blocks   Id  System
/dev/hda1          0+    100     101-    811251   fd  Linux raid autodetect
/dev/hda2        101     201     101     811282+  fd  Linux raid autodetect
/dev/hda3        202    7296    7095   56990587+  fd  Linux raid autodetect
/dev/hda4          0       -       0          0    0  Empty

sodom# sfdisk -l /dev/hdc

Disk /dev/hdc: 238216 cylinders, 16 heads, 63 sectors/track
Warning: The partition table looks like it was made
for C/H/S=*/255/63 (instead of 238216/16/63).
For this listing I'll assume that geometry.
Units = cylinders of 8225280 bytes, blocks of 1024 bytes, counting from 0

   Device Boot Start     End   #cyls    #blocks   Id  System
/dev/hdc1          0+    100     101-    811251   fd  Linux raid autodetect
/dev/hdc2        101     201     101     811282+  fd  Linux raid autodetect
/dev/hdc3        202    7296    7095   56990587+  fd  Linux raid autodetect
/dev/hdc4       7297   14945    7649   61440592+  8e  Linux LVM

regards
Hadmut

#1 lore raw

BTW, lvm version is 2.01.12-2 (debian).

I just saw that this problem was already mentioned on the mailing list in February with the subject

Having duplicate PV problems, think there's a bug in LVM2 md component detection

but there wasn't any solution.

regards
Hadmut

#2 lore raw

Have you tried using the filters?

brassow

Quoted 20 lines · Hadmut Danisch

BTW, lvm version is 2.01.12-2 (debian).

I just saw that this problem was already mentioned on the mailing list in February with the subject

Having duplicate PV problems, think there's a bug in LVM2 md
component
detection

but there wasn't any solution.

regards
Hadmut

_______________________________________________
linux-lvm mailing list
[email protected]
https://www.redhat.com/mailman/listinfo/linux-lvm
read the LVM HOW-TO at http://tldp.org/HOWTO/LVM-HOWTO/

#3 lore raw
Quoted 3 lines · Jonathan E Brassow

Have you tried using the filters?

brassow

Setting md_component_detection = 0 to 1 solved the problem.

regards
Hadmut