Hp-UX 11.11 - VCS3.5 three node cluster with oracle 9i RAC ..

[Saravanan_6]

Hi,

I am trying to set up a HP-UX 11.11 - VCS 3.5 cluster with Oracle 9i RAC on 3 nodes. I am facing a problem in bringing up the cluster status UP.

Steps Followed :

I followed the dbac_icg.pdf procedures to setup my cluster.

1. Private IP's are configured on each node and all three nodes are pinging via public/private IP's
2. In each node, rlogin is working without prompting username/passwd.
3. All cluster prerequisites are made .

4. database_ac_for_oracle9i/installDBAC script is executed.
5. All required ( cluster name, unique id, IP's ..etc ) details are provided and it installed all the packages and three nodes are rebooted.
6. Executed the below commands after the node reboot..
/sbin/init.d/llt start
/sbin/init.d/gab start
/sbin/init.d/vcsmm start
/sbin/init.d/lmx start
/sbin/init.d/odm start
Hastart

then verified the cluster status ;


corniche# hastatus
attempting to connect....connected

group resource system message
--------------- -------------------- --------------- --------------------
corniche RUNNING
coronet RUNNING
zephyr RUNNING

corniche# gabconfig -a
GAB Port Memberships
===============================================================
Port a gen e450e407 membership 012
Port h gen d75ed703 membership 012
Port o gen e450e407 membership 012

corniche# hastatus -summary

-- SYSTEM STATE
-- System State Frozen

A corniche RUNNING 0
A coronet RUNNING 0
A zephyr RUNNING 0
corniche#

Odm port -d is not running and the cluster is not in UP status. Please pass me the hints to troubleshoot this problem and making the cluster UP.

~Saravanan

[Sameer_Nirmal · Accepted solution]

Refer the following document for the steps to configure/setup Veritas ODM as mentioned on the page 20.

http://ftp.support.veritas.com/pub/support/products/DBE_Advanced_Cluster_for_Oracle_RAC/286508.pdf

[whiteknight]

Saravana,

Have you tuned symantec kernel parameter mm_slave_max ? This is one of the required parameter to be tuned for VCS

#kmtune -s =2048

WK

[Saravanan_6]

Hi,

Thanks for a useful hint. I followed the steps mentioned in page 20 and but still i could not make the cluster UP.

corniche# ls -al /sbin/rc2.d/*odm
lrwxr-xr-x 1 root sys 16 Jan 3 08:25 /sbin/rc2.d/S980odm -> /sbin/init.d/odm
corniche# rm -f /sbin/rc2.d/S980odm
corniche# cat /dev/odm/cluster ==> This file itself missing in my setup.
cat: Cannot open /dev/odm/cluster: No such file or directory
corniche#

corniche# hagrp -state
#Group Attribute System Value
ClusterService State corniche |ONLINE|
ClusterService State coronet |OFFLINE|
ClusterService State zephyr |OFFLINE|
corniche#

I did uninstallation through uninstallDBAC script and tried to set the cluster freshly. At Every install, i am getting the same status.

Please throw some light to solve this problem.

With regards,
Saravanan

[melvyn burnard]

I suggest you visit the symantec wb support pages for issues with the VCS product:
http://www.symantec.com/enterprise/support/

[Sameer_Nirmal]

Looks like you didn't read the instructions properly.

There has to be /sbin/rc2.d/S980odm file which is needed to start the ODM driver during bootup. It is linked as
# ln -s /sbin/init.d/odm /sbin/rc2.d/S980odm

/dev/odm/cluster file will only show up if the /dev/odm is mounted.

Assuming if the odm is not started during bootup, you can start it using
# /sbin/init.d/odm start

If there is a problem, it will show you why ODM can't be up.