Checking and testing the configuration for clustered gateways

Methods of checking cluster configuration

There are two ways you can check your cluster configuration prior to placing your cluster online:

Running the Cluster Configuration Checker script--The script runs under a UNIX shell or as a Common Gateway Interface (CGI) script from a UNIX Web server. It detects the following configuration problems:
- Mismatched licenses
- Mismatched option settings
- Incorrectly configured network interfaces
- Different versions of Data ONTAP on the local and partner gateways
Manually, through the command line

The Cluster Configuration Checker is available on the Tools & Utilities page on NOW.

The following tables shows the requirements that you need to use the cluster configuration checker

UNIX shell	CGI
Names or addresses of each partner of the NetApp cluster
Login names and passwords on both gFilers
UNIX computer that can access both gFilers (you can use the ping command to confirm this)	Web server that can access both gFilers A place to store the Perl script where the Web server can find it

To run the Cluster Configuration Checker script from a UNIX shell prompt, complete the following steps.

Step	Action
1	Execute the cf-config-check.cgi script from a Unix shell prompt, with the name or IP address of both gFilers on the command line.
2	Add the word verbose to the end of the command line to view more detail.
3	When prompted, enter the rsh login names and passwords of both gFilers.
4	Examine the script output and correct any errors or differences identified in the output.

To run the Cluster Configuration Checker as a CGI script, complete the following steps.

Step	Action
1	Point your browser to the URL corresponding to the location of the Perl script.
2	Enter the names or IP addresses of the two gFilers.
3	Enter the rsh login names and passwords.
4	Click Check.
5	If no results are displayed, click your browser's Stop button.
6	Examine the script output and correct any errors or differences identified in the output.

To check options manually on the local and partner gateways, complete the following steps.

Step	Action
1	View the options set on the local gateway by entering the following command at the gateway's console: options Result The current option settings on the gateway are listed. You should see output similar to the following: autosupport.doit DONT autosupport.enable on . . dns.domainname gfiler1 (might be overwritten in takeover) dns.enable off (might be overwritten in takeover) ftpd.enable off . .
2	Note the values of these options for the local gateway, and repeat Step 1 on the partner gateway.
3	Verify that the options that have comments in parentheses are set to the same value for both gateways. The comments you see are as follows: Value might be overwritten in takeover Same value required in local+partner Same value in local+partner recommended
4	On either gateway console, correct any mismatched options using the following command: options option_name option_value
5	Check the license settings on both gateways using the following command: license
6	Correct mismatched license settings using the following command: license code

To test takeover and giveback, complete the following steps.

Step	Action
1	Check the cabling on the interconnect cables to make sure that they are secure.
2	Verify that you can create and retrieve files on both nodes for each licensed protocol.
3	From the local node, enter the following command and observe the output: cf enable Result Takeover of the partner node is enabled on the local node.
4	Enter the following command from the local node console: cf takeover Result The local node takes over the partner node and gives the following output: takeover completed
5	Make sure that the local node can access both the local and partner disks.
6	Give back the partner node by entering the following command: cf giveback Result The local node releases the partner node, which reboots and resumes normal operation. The following message is displayed on the console when the process is completed: giveback completed
7	If takeover and giveback...	Then...
	Is completed successfully	Repeat Steps 2 through 6 on the partner node.
	Fails	Go to Correcting takeover or giveback failure.

The following table describes tests you can run to determine why takeover or giveback failed and also provides information to correct the problem.

Note
You must reboot the cluster after completing each test.

To correct a takeover or giveback failure, complete the following steps.

Step	Action
1	Check communication between the local and partner nodes by entering the following command and observing the messages: cf status Results If there is a problem with cabling or the adapter, the appropriate message is displayed.
2	If the error message indicates...	Then...
	A cluster adapter error	Check the cluster adapter cabling. Make sure that the cabling is correct and properly seated.
	A cabling error	Check the cabling and reseat and tighten any loose cables.
	A networking error	Check for network connectivity. See the Data ONTAP System Administrator's Guide for more information.
3	Run the cluster configuration checker script or check the cluster configuration manually. Correct any errors or differences displayed in the output.
4	Reboot the cluster and rerun the takeover and giveback tests.
5	If clustering is still not enabled, call NetApp Technical Support or your authorized reseller.