My Oracle Support Banner

Solaris Cluster - Shared Address Load Balancer Is Not Working Properly (Doc ID 2578215.1)

Last updated on AUGUST 14, 2020

Applies to:

Solaris Cluster - Version 3.2 to 4.4 [Release 3.2 to 4.4]
Information in this document applies to any platform.
Solaris Cluster allows for the configuration of a Scalable Resource Group (RG) in which one or more application resources
are allowed to run simulanteously on all members in the RG Nodelist. These are sometimes referred to as Multi-mastered
(Scalable) services. They are partnered with a Failover RG which provides a 'Shared-Address' (SA) resource. A dependency is
configured such that the Scalable RG (and resource(s)) within it depend on this SA RG/Resource. The SA RG/Resource provides
a load-balancer which distributes the client requests to the RG members. There are various properties which effect how this
is implemented. The Apache Web Server is one example of a Scalable Application which would benefit from the use of a Shared
Address resource.

Oracle® Solaris Cluster Data Service for Apache Guide

https://docs.oracle.com/cd/E69294_01/html/E69329/index.html

The primary node for the SA RG is called the Global Interface Node. It will have the SA IP configured on the appropriate
network interface as determined by the values assigned to the NetIfList and HostnameLIst properties. The non-primary nodes
will have the same IP configured on it Local Loopback (lo0) interface.

You can read more on the Scalable Data Services in the Concepts Guide.

https://docs.oracle.com/cd/E69294_01/html/E69310/cacfifia.html#CLCONcacdhegh

Symptoms

In this case study, we have a 4-node cluster providing providing a Scalable service using port 8080

(SCALABLE-RG) Res Group Nodelist: Node1 Node2 Node3 Node4
(SCALABLE-RG) Res Group Maximum_primaries: 4
(SCALABLE-RG) Res Group Desired_primaries: 4
(SCALABLE-RG) Res Group RG_dependencies: Shared-Address-RG

(SCALABLE-RG:Scalable-resource) Res resource type: SUNW.gds:6
(SCALABLE-PACO:paco-service) Res strong dependencies: Shared-Address-Resource
(SCALABLE-RG:Scalable-resource:Load_balancing_policy) Res property value: LB_WEIGHTED
(SCALABLE-RG:Scalable-resource:Network_resources_used) Res property value: Shared-Address-Resource
(SCALABLE-RG:scalable-service:Port_list) Res property value: 8080/tcp
(SCALABLE-RG:Scalable-resource:Start_command) Res property value: /global/scalable/services/start-script

-

(Shared-Address-RG) Res Group mode: Failover
(Shared-Address-RG) Res Group Nodelist: Node1 Node2 Node3 Node4

(Shared-Address-RG:Shared-Address-Resource) Res resource type: SUNW.SharedAddress:2
(Shared-Adress-RG:Shared-Address-Resource:NetIfList) Res property value: net@1 net@2 net@3 net@4
(Shared-Adress-RG:Shared-Address-Resource) Res property value: Shared-Address-Hostname

-

root@Node1 # netstat -an|grep 8080|wc -l
773

root@Node2 # netstat -an|grep 8080|wc -l
2

root@Node3 # netstat -an|grep 8080|wc -l
3

root@Node4 # netstat -an|grep 8080|wc -l
5

Changes

The problem was noticed after rebooting the cluster following installation of a Critical Patch Update (CPU).

Cause

To view full details, sign in with your My Oracle Support account.

Don't have a My Oracle Support account? Click to get started!


In this Document
Symptoms
Changes
Cause
Solution

My Oracle Support provides customers with access to over a million knowledge articles and a vibrant support community of peers and Oracle experts.