Single disk lock issue

[hattori]

I attempted to create a 2-node cluster on HP.

Here the procedure that I followed:

1) created 4 volumes and mapped to cluster host group which concluded both nodes
2) Both nodes see all these volumes

3) on node 1, performed the following steps for each volume:
Assuming cxtxdx is primary and cytydy is the secondary path
# pvcreate â f /dev/rdsk/cxtxdx
o # pvcreate â f /dev/rdsk/cytydy
o # mkdir /dev/vg##
o # mknod /dev/vg##/group c 64 0x##0000
o # vgcreate /dev/vg## /dev/dsk/cxtxdx /dev/dsk/cytydy
o # lvcreate â L 1024 /dev/vg##
o # vgchange â a n /dev/vg##
o # vgexport â p â s -m /tmp/vg##.map /dev/vg##
o # transfer this vg##.map file over node 2 at /tmp directory Repeat for all volume groups

4) On node 2, performed the following steps:
o # mkdir /dev/vg##
o # mknod /dev/vg##/group c 64 0x##0000
o # ioscan â f
o # insf -e
o # vgimport â s â m /tmp/vg##.map /dev/vg##
o # vgchange â a n /dev/vg#
Repeat for all volumes groups

lvmtab on node 1:
# strings /etc/lvmtab
/dev/vg00
/dev/dsk/c2t1d0
/dev/vg01
/dev/dsk/c34t0d0
/dev/dsk/c32t0d0
/dev/vg02
/dev/dsk/c32t0d1
/dev/dsk/c34t0d1
/dev/vg03
/dev/dsk/c34t0d2
/dev/dsk/c32t0d2
/dev/vg04
/dev/dsk/c32t0d3
/dev/dsk/c34t0d3

lvmtab on node 2:
# strings /etc/lvmtab
/dev/vg00
/dev/dsk/c2t1d0
/dev/vg01
/dev/dsk/c97t0d0
/dev/dsk/c95t0d0
/dev/vg02
/dev/dsk/c97t0d1
/dev/dsk/c95t0d1
/dev/vg03
/dev/dsk/c97t0d2
/dev/dsk/c95t0d2
/dev/vg04
/dev/dsk/c97t0d3
/dev/dsk/c95t0d3


6) Then I used SMH to create cluster:
a) added both nodes to cluster
b) defined parameters, network, lock, and volume groups
c) I selected sigle disk lock and selected all volume group in "cluster aware"
d) clicked ok
e) clicked "check configuration"

The operation log returned the following message: First cluster lock physical volume "/dev/dsk/c95t0d0" on node "powerhp" does not belong to first cluster lock volume group "/dev/vg01". Specify a physical volume that belongs to first volume group. (see attachment for more details)


So, it is the cluster clock setting that caused this problem. However, I am not able to change the lock device path since there is only one path in the drop down list.


Does anyone have any suggestions on what I might need to do to fix the problem?

[Steven E. Protter]

Shalom,

I don't think you have finished configuration thats all.

cmquerycl -n nodename1 nodename2

Followed by cmcheckconf and cmapplyconf.

You can set the lock disk in your configuration file before running cmquerycl

There are a number of errors in your post that make it difficult for me to understand the process.

ex: I used SMH (do you mean SAM) to create cluster.

clock instead of lock.

But I plowed through and tried to help anyway.

SEP

[hattori]

Steven,

Sorry for the confusion.

SMH stands for System Management Homepage.Basically, it's is same as Serviceguard Manager. HP SMH includes the integration of Web-based applications that provide a graphical user interface (GUI) for HP system administration tasks. You can also check it out here: http://docs.hp.com/en/B3935-90108/ch01s02.html#babhjjhd

"clock" is my typo. I meant lock

I also noticed the text in steps 3 and 4 are messed up. HEre are the better ones:

3) on node 1, performed the following steps for each volume:
Assuming cxtxdx is primary and cytydy is the secondary path
# pvcreate -f /dev/rdsk/cxtxdx
# pvcreate -f /dev/rdsk/cytydy
# mkdir /dev/vg##
# mknod /dev/vg##/group c 64 0x##0000
# vgcreate /dev/vg## /dev/dsk/cxtxdx /dev/dsk/cytydy
# lvcreate -L 1024 /dev/vg##
# vgchange -a n /dev/vg##
# vgexport -p -s -m /tmp/vg##.map /dev/vg##
# transfer this vg##.map file over node 2 at /tmp directory Repeat for all volume groups

4) On node 2, performed the following steps:
# mkdir /dev/vg##
# mknod /dev/vg##/group c 64 0x##0000
# ioscan -f
# insf -e
# vgimport -s -m /tmp/vg##.map /dev/vg##
# vgchange -a n /dev/vg#
Repeat for all volumes groups



By the way, what do these commands do. At what point should I execute them?

cmquerycl -n nodename1 nodename2

cmcheckconf
cmapplyconf

[David Bellamy]

hattori
cmquerycl - queries the nodes and get information about LVM and network interface ..
cmcheckcl - checks the configuration of your Service Guard cluster and will report any errors.

cmapplycl - applies the changes to the cluster.