| Netra™ High Availability Suite 3.0 1/08 Foundation Services Troubleshooting Guide | 
   |
| C H A P T E R 8 |
For information about how to recover from errors that occur when you are using the Node Management Agent (NMA), see the following sections:
If an external client cannot communicate with the NMA on a peer node, perform the following procedure.
|
Confirm that an RMI registry has not been instantiated on a designated NMA TCP port.
When a Java Dynamic Management Kit agent enables an RMI connector, the agent instantiates an RMI registry on the port it is using. If an RMI registry uses the same TCP port as the NMA, communication with the NMA is impossible. The RMI registry that initially allocated the port has control of the port.
If the NMA is not accessible through RMI, do the following.
Uncomment the java.rmi.server property in the nma.properties file.
Set the value of the java.rmi.server property to the IP address or host name of the node.
# /etc/opt/SUNWcgha/init.d/nma stop |
If the NMA fails to restart, see NMA Not Restarted After Failure.
If you cannot resolve this problem, contact your customer support center.
The NMA process is monitored by the Daemon Monitor. If the NMA on a peer node fails, the Daemon Monitor restarts it. If the NMA fails to restart, perform the following procedure.
Find the current number of times that the Daemon Monitor has attempted to restart the NMA.
This parameter is called the reset retry count. For information, see Daemon Monitor Statistics in the Netra High Availability Suite 3.0 1/08 Foundation Services NMA Programming Guide.
Confirm that the reset retry count does not exceed 10.
The Daemon Monitor attempts to restart the NMA up to 10 times.
If the reset retry count has not been exceeded, go to Step 5.
If the reset retry count has been exceeded, try one or both of the following:
a. Reset the reset retry counter by using the resetRetryCount method.
For information, see Manipulating Daemon Monitor Retry Settings in the Netra High Availability Suite 3.0 1/08 Foundation Services NMA Programming Guide.
If you cannot resolve this problem, contact your customer support center.
If the NMA does not send SNMP traps to a target, confirm that the following files are configured:
where installDir is the root installation directory. Replace installDir with the root installation directory if the root installation directory is not /.
For information about these files, see their man pages in the Netra High Availability Suite 3.0 1/08 Foundation Services Reference Manual. For information about sending SNMP traps, see “Developing a Remote SNMP Manager” in the Netra High Availability Suite 3.0 1/08 Foundation Services NMA Programming Guide.
The switchOver method is run on the NMA on the master node to initiate a switchover. When you use the switchOver method, do not use the floating external address to access the master node. When the floating external address is used, the switchOver method breaks the connection. Instead, use a fixed IP address on the master node.
Cascading is the transfer of statistics from the NMA on each of the peer nodes into the namespace of the master node. Cascading provides the NMA of the master node with a view of the entire cluster. If the NMA on a peer node cannot be detected by the NMA on the master node, perform the following procedure.
Confirm that cascading is enabled.
Confirm that the following line is commented out:
com.sun.nhas.ma.cascading.enabled=false |
If the line is commented out, go to Step 6.
If the line is not commented out, comment it out and restart the NMA:
# /etc/opt/SUNWcgha/init.d/nma stop |
If the NMA fails to restart, see NMA Not Restarted After Failure.
Confirm that the maximum number of HTTP clients has not been exceeded.
Connect to the NMA of the node that is not cascading correctly, using an Internet browser.
For information, see Accessing the NMA in the Netra High Availability Suite 3.0 1/08 Foundation Services NMA Programming Guide.
Find the number of connected clients.
The number of connected clients is specified in the com.sun.nhas.ma.connectors.http.client property.
If the number of connected clients exceeds the maximum, increase the maximum.
Change the connectors.http.client parameter in the /etc/opt/SUNWcgha/nma.properties file.
# /etc/opt/SUNWcgha/init.d/nma stop |
If the NMA fails to restart, see NMA Not Restarted After Failure.
If you cannot resolve this problem, contact your customer support center.
| Netra™ High Availability Suite 3.0 1/08 Foundation Services Troubleshooting Guide | 819-5248-13 |
|
Copyright © 2008, Sun Microsystems, Inc. All rights reserved.
To Investigate Why an External Client
Cannot Communicate With an NMA on a Peer Node