topic Re: error with cluster lock disk in Operating System - HP-UX

error with cluster lock disk

Guillermo Barrera — Thu, 09 Jun 2011 14:06:06 GMT

I was trying to add a package to a production-running cluster. Fails with:

cmcheckconf -v -C /tmp/SVRBDE_cluster.txt -P /etc/cmcluster/form01/form01.conf

Error: First cluster lock volume group "/dev/vgVA_lock" is not available on all nodes. Choose a volume group that is available on all nodes.
Error: First cluster lock physical volume "/dev/dsk/c4t0d1" on node "svrabajo" does not belong to first cluster lock volume group "/dev/vgVA_lock". Specify a physical volume that belongs to first volume group.
Error: Unable to find find first cluster lock volume group "/dev/vgVA_lock" on node "svrariba" but found it in cluster "SVRBDE".

How can i make lock_disk available fro both nodes.... vgchange ---?????

Maybe some differences on lvmtab, but i am missing some point???

Thanks

Re: error with cluster lock disk

Guillermo Barrera — Thu, 09 Jun 2011 14:07:45 GMT

strings /etc/lvmtab output is included in attached file on first message....

Re: error with cluster lock disk

g3jza — Thu, 09 Jun 2011 15:11:36 GMT

Hi,
in the attachment, I include the correct procedure to build a cluster based on cluster lock disk...

Error: First cluster lock volume group "/dev/vgVA_lock" is not available on all nodes. Choose a volume group that is available on all nodes.
-->make sure the VG can be seen on all the nodes of the cluster...

Re: error with cluster lock disk

Guillermo Barrera — Fri, 10 Jun 2011 09:44:47 GMT

I was reading you attached file.Thanks.
A little bit better, but, well, no so much...

The situation now is:

i) lock vg is vgVA_lock, i can activate VG on both nodes using vgchange -a s

On node 1) I can activate VG executing:

[0.00 root@svrabajo] cmcluster > vgchange -a s vgVA_lock

Activated volume group in Shared Mode.
This node is the Server.
Volume group "vgVA_lock" has been successfully changed.
[0.00 root@svrabajo] cmcluster > vgdisplay -v vgVA_lock
--- Volume groups ---
VG Name /dev/vgVA_lock
VG Write Access read/write
VG Status available, shared, server
Max LV 255
Cur LV 0
Open LV 0
Max PV 16
Cur PV 1
Act PV 1
Max PE per PV 1016
VGDA 2
PE Size (Mbytes) 4
Total PE 4
Alloc PE 0
Free PE 4
Total PVG 0
Total Spare PVs 0
Total Spare PVs in use 0
--- ---

svrabajo Server

--- Physical volumes ---
PV Name /dev/dsk/c4t0d1
PV Name /dev/dsk/c7t0d1 Alternate Link
PV Status available
Total PE 4
Free PE 4
Autoswitch On
Proactive Polling On

On node 2) I can activate VG:

[0.02 root@svrariba] / > vgchange -a s vgVA_lock
Activated volume group in Shared Mode.
This node is the Server.
Volume group "vgVA_lock" has been successfully changed.
[0.02 root@svrariba] / > vgdisplay -v vgVA_lock
--- Volume groups ---
VG Name /dev/vgVA_lock
VG Write Access read/write
VG Status available, shared, server
Max LV 255
Cur LV 0
Open LV 0
Max PV 16
Cur PV 1
Act PV 1
Max PE per PV 1016
VGDA 2
PE Size (Mbytes) 4
Total PE 4
Alloc PE 0
Free PE 4
Total PVG 0
Total Spare PVs 0
Total Spare PVs in use 0
--- ---

svrariba Server

--- Physical volumes ---
PV Name /dev/dsk/c4t0d1
PV Name /dev/dsk/c7t0d1 Alternate Link
PV Status available
Total PE 4
Free PE 4
Autoswitch On

ii) cmcheckconf fails. I changed file "/tmp/cmclconfig.ascii" LOCK_PV section (node svrariba):

NODE_NAME svrabajo
NETWORK_INTERFACE lan0
HEARTBEAT_IP 10.241.128.30
NETWORK_INTERFACE lan1
NETWORK_INTERFACE lan2
HEARTBEAT_IP 10.38.64.11
FIRST_CLUSTER_LOCK_PV /dev/dsk/c4t0d1
# List of serial device file names
# For example:
# SERIAL_DEVICE_FILE /dev/tty0p0

# Possible standby Network Interfaces for lan0: lan1.
# Warning: There are no standby network interfaces for lan2.

NODE_NAME svrariba
NETWORK_INTERFACE lan0
HEARTBEAT_IP 10.241.128.31
NETWORK_INTERFACE lan1
NETWORK_INTERFACE lan2
HEARTBEAT_IP 10.38.64.10
FIRST_CLUSTER_LOCK_PV /dev/dsk/c9t0d1 ------->>Â¡Â¡changed for /dev/dsk/c4t0d1!!

And the error:

[0.02 root@svrariba] / > cmcheckconf -v -C /tmp/cmclconfig.ascii

Checking cluster file: /tmp/cmclconfig.ascii
Checking nodes ... Done
Checking existing configuration ... Done
Gathering configuration information ... Done
Gathering configuration information ... Done
Gathering configuration information ............. Done
Cluster SVRBDE is an existing cluster
FIRST_CLUSTER_LOCK_PV value for node svrariba cannot be modified from /dev/dsk/c9t0d1 to /dev/dsk/c4t0d1 while cluster is running.
cmcheckconf : Unable to verify cluster file: /tmp/cmclconfig.ascii.
[0.05 root@svrariba] / >

iii) node svrabajo cant no see PV /dev/dsk/c9t0d1

Thanks.

Re: error with cluster lock disk

g3jza — Fri, 10 Jun 2011 10:01:17 GMT

Hmm,
is there any reason that you need the cluster lock vg be activated on both nodes at the same time, using Shared-LVM? The cluster lock disk should be using simple VG activation (#vgchange -a y). I don't even know if the SLVM is supported as cluster lock VG...

Re: error with cluster lock disk

rariasn — Wed, 06 Jul 2011 17:35:06 GMT

Hola Guillermo:

4.) If your Serviceguard version is A.11.18 or older, halt the cluster now. ("cmhaltcl")
If your Serviceguard version is A.11.19 or newer but you're using Veritas CVM, halt the SG-CFS-pkg package.

http://h30499.www3.hp.com/t5/Serviceguard/How-to-change-Cluster-Lck-Disk-in-HP-UX/m-p/4784939#M19697

rgs,

Re: error with cluster lock disk

Guillermo Barrera — Fri, 10 Jun 2011 10:44:09 GMT

@g3jza: Yes, you are right, but i can not activate it with vgchane -a e (confilcts with configured mode!! and i have no idea what's happened, so i tried this)

it's old version:

[0.03 root@svrariba] / > cmversion
A.11.15.00

In case a halt the cluster software, and execute:

i)checkconf (without warnings, errors, whatever..)
ii)applyconf
iii) start software

everything will be OK?????

Thanks.

Re: error with cluster lock disk

g3jza — Fri, 10 Jun 2011 10:51:22 GMT

Try running:

#vgchange -c n -S n vgname

to put it out of shared mode

Re: error with cluster lock disk

g3jza — Fri, 10 Jun 2011 10:53:01 GMT

"FIRST_CLUSTER_LOCK_PV value for node svrariba cannot be modified from /dev/dsk/c9t0d1 to /dev/dsk/c4t0d1 while cluster is running."

yep, try with halted cluster...