Sun Cluster 3.2 SNMP Interface

System administrators in data centers usually use their corporate enterprise management system/tools to monitor and manage all computer systems, include both hardware and software. System management tools, such as HP Openview, BMC Patrol, IBM Tivoli and SunMC(Sun Management Center) all provide SNMP support in order to monitor and manage the heterogeneous computer environment in data centers.

Like many other computer system Sun Cluster also provides SNMP management interface. There are two SNMP interfaces provided in Sun Cluster 3.2. One is for the cluster events, another one is for the cluster objects and their status changes.

SNMP Interface for Cluster Event

The cluster event SNMP interface is newly added to Sun Cluster 3.2. SNMP module can be enabled on any cluster nodes without rely on any other SNMP agent infrastructure. Since all cluster nodes get the same cluster events, usually SNMP event module only need to be enabled on one of the cluster nodes. The event SNMP interface includes


  • Store upto 50 most recent cluster events to MIB table when SNMP module is enabled on a node.

  • Send SNMP traps to any hosts that you want for the events with warning or high severities.


The list of cluster event severities from low to high is

  • INFO

  • WARNING

  • ERROR

  • CRITICAL

  • FATAL


The default port number for the event SNMP module is 11161, and the default port for the SNMP traps is 11162. They can be changed by modifying the Common Agent Container property file, which is /etc/cacao/instances/default/private/cacao.properties. The event MIB OID is 1.3.6.1.4.1.42.2.80. The management user interface for the event SNMP is added to both GUI(Sun Cluster Manager) and new CLI. Here are a couple of screen snapshots for the GUI.








The new commands for SNMP interface are

  • clsnmpmib: Administer SNMP MIBs

  • clsnmphost: Administer hosts for SNMP traps

  • clsnmpuser: Administer SNMP users(specific for SNMP V3 protocol).


Here are a couple of examples of using the above commands.

Enable event MIB on the current cluster node
# clsnmpmib enable event

Add a host to receive SNMP traps for the cluster events with warning or higher severity.
# clsnmphost add

SNMP Interface for Cluster objects

SNMP interface for the Cluster objects, such as Nodes, Resource Groups, Quorums and their corresponding status is provided via Sun Cluster-SunMC integration module. The integration module is installed by default on all cluster nodes.

To enable the module on a cluster node, SunMC agent infrastructure has to be installed and running on the node. After SunMC agent starts on the node, Sun Cluster-SunMC integration module can be loaded and enabled via SunMC management console. SunMC server and console can be installed on one of the cluster node or another system that is supported by SunMC(refer to SunMC User Guide for more detail). Similar to the event SNMP module, Sun Cluster-SunMC module only need to be enabled on one of the cluster node if you only want to get Cluster SNMP MIB data and traps. The following left side picture shows SunMC console UI where to load and enable the Sun Cluster-SunMC module. The right side screen snapshot is the SunMC console UI to monitor the configuration and status of cluster objects. Sun Cluster-SunMC SNMP module MIB OID is 1.3.6.1.4.1.42.2.80.1.1.1









SunMC agent running on a cluster node will send out SNMP traps to SunMC server whenever there are any object status changes, such as Cluster Resource Group online/offline, managed/unmanaged, node on/off and etc. The SNMP traps also can be sent directly from SunMC agent to any other hosts by adding the secondary trap destination via SunMC CLI es-trapdest. For example, the following statement is to add myHost to receive SunMC SNMP traps from SunMC agent via port 11162.

# /opt/SUNWsymon/sbin/es-trapdest -c agent -a myHost 11162
(note: SunMC agent has to be restarted to take effect of above setting).

Leland Chen
Solaris Cluster Engineering 

Comments:

I have Sun Cluster 3.2 monitored by SunMC 4.0. I have found an issue where the offline resources have a status of Caution even though they are okay on the other node. Do I just have to acknowledge and ignore or am I missing something?

Posted by Tim Sutton on July 13, 2008 at 11:22 PM PDT #

I think this is the default behavior.

Basically default thresholds are set on most of cluster objects, such as Resources, Resource Groups, nodes and etc. If you want to change the default behavior, you can use the SunMC UI to modify the threshold configuration on any of the cluster objects.

Posted by Leland Chen on July 14, 2008 at 03:18 AM PDT #

umm well I have acknowledged them for now, I guess I don't want to change the threshold for online/offline when the service failsover, was just hoping for a clever way of saying that if it is online on one side and offline on the other all is good :)

Maybe a new features for the SunMC Cluster module :)

Thanks for responding

Tim

Posted by Tim Sutton on July 14, 2008 at 03:29 AM PDT #

Post a Comment:
  • HTML Syntax: NOT allowed
About

Oracle Solaris Cluster Engineering Blog

Search

Archives
« May 2015
SunMonTueWedThuFriSat
     
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
18
19
20
21
22
23
24
25
26
27
28
29
30
31
      
Today