PRIMECLUSTER Installation and Administration Guide 4.1 (for Solaris(TM) Operating System)

5.3 Setting Up Shared Disk Connection Confirmation

Shared disk connection confirmation is a function that performs the followings:

• Check that the shared disk is powered.
• Check that the cable is properly connected.
• Detect new shared disks.

If the paths of a disk array unit are multiplexed, and the paths are controlled by a multipath disk control function, shared disk connection confirmation will not detect an error even if an error occurs in part of the multiplexed paths.

If an error is detected in a shared disk unit or if a new shared disk unit is detected, shared disk connection confirmation outputs a message to syslogd(1M) and the CRM main window.

This function allows you to specify whether or not RMS activation is to be suppressed in the node in which the failure was detected. If a failure is detected in a shared disk unit, the userApplication might not operate normally and user resources in that shared disk unit might become damaged. It is therefore recommended that you specify that RMS activation should be suppressed if a failure is detected. After PRIMECLUSTER is installed, shared disk connection confirmation is initially disabled by default. To change this, follow the procedures described below.

• Set up shared disk connection confirmation after the initial setup of the Cluster Resource Management facility is completed.
• To view the manual pages of each command, add "/etc/opt/FJSVcluster/man" to the MANPATH variable.

Enabling shared disk connection confirmation

Take the following steps on each node. Shared disk connection confirmation will be executed from the next node startup.

RMS activation is to be suppressed if a failure is detected

1. Add the following to the "RELIANT_INITSCRIPT" environment variable of RMS:

   /etc/opt/FJSVcluster/bin/clinitscript

   Example: Add the following to /opt/SMAW/SMAWRrms/bin/hvenv.local.

   export RELIANT_INITSCRIPT=/etc/opt/FJSVcluster/bin/clinitscript

2. Execute the "clsetacfparam(1M)" command.

   Example:

   # /etc/opt/FJSVcluster/bin/clsetacfparam -p auto

If there is an error in the shared disk unit, it is detected by the registered script (/etc/opt/FJSVcluster/bin/clinitscript) and RMS stops. In this case, the following RMS message will be sent to syslogd(1M):

"(INI, 12): FATAL ERROR: InitScript returned non-zero exit code 1."

RMS activation is not to be suppressed if a failure is detected

1. Execute the "clsetacfparam(1M)" command.

   Example:

   # /etc/opt/FJSVcluster/bin/clsetacfparam -p auto

Disabling shared disk connection confirmation

To cancel the setting previously made for "Enabling shared disk connection confirmation," perform the following procedure on all the nodes. After you next boot the node, shared disk connection will not be confirmed.

"Suppress RMS activation if an error is detected" was previously configured

1. Delete the following from the RELIANT_INITSCRIPT environment variable of RMS.

   /etc/opt/FJSVcluster/bin/clinitscript

   Example: Delete the following from /opt/SMAW/SMAWRrms/bin/hvenv.local.

   export RELIANT_INITSCRIPT=/etc/opt/FJSVcluster/bin/clinitscript

2. Execute the "clsetacfparam(1M)" command.

   Example:

   # /etc/opt/FJSVcluster/bin/clsetacfparam -p none

"Do not suppress RMS activation if an error is detected" was specified

1. Execute the "clsetacfparam(1M)" command.

   Example:

   # /etc/opt/FJSVcluster/bin/clsetacfparam -p none

• For details on the RMS environment variables, see "10 Appendix - Environment Variables" in the "PRIMECLUSTER Reliant Monitor Service (RMS) with Wizard Tools Configuration and Administration Guide."
• For details on the "clsetacfparam(1M)" command, see the manual page for clsetacfparam(1M).