LVM2 in a cluster and lvchange.

Hi,

I would like to use the LVM in a 2-node cluster. I know that LVM is not CLVM, which will be released next year or so, but I would like to use LVM2 for it.

What I did was follows:

node1:
pvcreate /dev/sdo
vgcreate vg10 /dev/sdo
lvcrete -n test -L20G vg10
mkreiserfs /dev/vg10/test

Now, I tried to access the newly created LV from node2:
node2# pvscan ---------------> is this necesarry?
PV /dev/sdo VG vg10 lvm2 [68.36 GB / 48.36 GB free]
Total: 1 [48.36 GB] / in use: 1 [48.36 GB] / in no VG: 0 [0 ]
node2# vgscan ---------------> is this necesarry?
Reading all physical volumes. This may take a while... Found volume group "vg10" using metadata type lvm2
node2# lvscan ---------> it found the lv, but
inactive '/dev/vg10/test' [20.00 GB] next free (default)
node2# mount /dev/vg10/test /opt ---> can not be accessable.
mount: special device /dev/vg10/test does not exist
node2# lvscan -D ------> read from the disk,
inactive '/dev/vg10/test' [20.00 GB] next free (default)
node2# mount /dev/vg10/test /opt ---> still not accessable.
mount: special device /dev/vg10/test does not exist
node2# lvscan -D -b ---> same with -b option.
inactive '/dev/vg10/test' [20.00 GB] next free (default)
node2# mount /dev/vg10/test /opt
mount: special device /dev/vg10/test does not exist
node2# lvchange -ay /dev/vg10/test ----> changing it to active node. (it is active in node1)
node2# mount /dev/vg10/test /opt --------> now works.

Again, after extending the lv from node1, even tough lv size is shown correctly in node2, it can not be accessable, unless 'doing lvchange -an /dev/vg10/test ; lvchange -ay /dev/vg10/test'

node1# lvextend -L+10G /dev/vg10/test
Extending logical volume test to 30.00 GB
Logical volume test successfully resized

node2:~ # lvdisplay
--- Logical volume ---

 LV Name                /dev/vg10/test
 VG Name                vg10
 LV UUID                zUjDCv-L774-opXv-fZz8-zZpT-omKv-csr8JM
 LV Write Access        read/write
 LV Status              available
 # open                 0
 LV Size                30.00 GB
 Current LE             7680
 Segments               1
 Allocation             next free (default)
 Read ahead sectors     0
 Block device           254:1

node2:~ # resize_reiserfs /dev/vg10/test
resize_reiserfs 3.6.2 (2002)
/dev/vg10/test already is of the needed size. Nothing to be done --------> means that block devices tools did not see the extended size yet.

node2:~ # lvchange -an /dev/vg10/test ; lvchange -ay /dev/vg10/test node2:~ # resize_reiserfs /dev/vg10/test
resize_reiserfs 3.6.2 (2002)
ReiserFS report:

blocksize             4096
block count           7864320 (5242880)
free blocks           7855869 (5234509)
bitmap block count    240 (160)

Syncing..done ------------------> now it sees the extended size...

My question is that 'doing lvchange -an /dev/vg10/test ; lvchage -ay /dev/vg10/test' in the other node make the kernel recognize the lv changes. Is there any better way or command, becase making inactive temporaryly does not make me feel right.

TIA

_________________________________________________________________
The new MSN 8: smart spam protection and 2 months FREE* http://join.msn.com/?page=features/junkmail

Your other option is to poke around with dmsetup.

Currently:

dmsetup suspend vg10-test
lvchange -ay vg10/test

should work, but mixing dmsetup with the LVM2 tools like that is not supported and so it could break as a side-effect of changes in future releases.

Alasdair

[email protected]