Problems in my SG setup

[shashi kanth]

I am doing 2-node SG 11.18 Oracle10g fail-over setup in HP-UX 11.23 PA MC-OE. I am facing several issues.

(1) When SG trying to import VxVM DGs, in the package control log file, it continuously throws bellow error, and it never imports DGs. How to overcome this problem ?

"Apr 29 20:42:51 - Node "corniche": KSTATE for "/dev/vx/dsk/Oracle10g_dg/Oracle10g_Oradata" is DISABLED."


(2) How to remove a disk/LUN completely from VxVM control ?

I have specified one LUN for "cluster lock", Initially this LUN is under VxVM control, i have removed it manually by using the commands "vxdisk destroy; vxdisk rm" , but whenever i reboot the nodes, this LUN is going back to the VxVM control. How to remove it permanently ?


(3) I have 2 LAN cards in each node , lan0 and lan1 . lan0 is stationary, lan1 is for heartbeat. I have specified one relocatable IP in the package configuration file ( IP_ADDRESS, IP_SUBNET, MONITORED_SUBNET). My question is, do i need to specify the same IP in package control file also(IP , SUBNET) ?


My intention/purpose is, when SG starts the package , SG should assign this virtual/relocatable IP to lan0:1 and starts the database and listener. I have modified the listener file with virtual/relocatable IP. But, my question is, where i have tell that SG should bind that IP to lan0 only ?

[Steven E. Protter]

Shalom,

1)
Check the logs, look for further information on:
"/dev/vx/dsk/Oracle10g_dg/Oracle10g_Oradata"

There is a problem with this disk device but more information is needed. Check syslog.log and the cluster logs, there should be more.

2) No idea, cluster lock disk needs to be on shared storage.

3) I'd need to see configuration files and logs to help with this.

SEP

[skt_skt]

the lan card selection is defined in the cluster cofiguration file

do a
#cmgetconf -c clustername outputfile.out

and see the outputfile.out

[shashi kanth]

I have attached my cluster configuration file, package control file, and package control file.

Where i can find the SG cluster logs ?

There is much information is the syslog file.

[shashi kanth]

Package configuration file

[shashi kanth]

Package control file

[shashi kanth]

haoracle.conf file

[Stephen Doud · Accepted solution]

You did not mention whether this disk group is part of a CFS (cluster file system) environment and packages.

If it is not, then treat the problem as a regular VxVM issue. VxVM disk groups are only defined in the package control script, so problems with the enabling of the disk group would be registered in the package control log and perhaps syslog.log.

Use this command to locate any log associated with vxvm:

# find / | grep vx | grep log


See http://seer.support.veritas.com/docs/231913.htm