[
https://jira.jboss.org/browse/JBAS-8710?page=com.atlassian.jira.plugin.sy...
]
Richard Achmatowicz commented on JBAS-8710:
-------------------------------------------
Two test cases relating to PreferredMasterElectionPolicy have been fixed:
1. PreferredMasterElectionPolicyTestCase was faling on
testUsePreferredMasterViaAddress().
This test case tests what happens when we set up a set of cluster nodes {(localhost,
10000), (localhost, 10001), (localhost,10002)} and then try to set the preferred master
using in IP address literal. localhost translates to either 127.0.0.1 or ::1 depending on
the settings of java.net.preferIPv4Stack and java.net.preferIPv6Addresses. I made things
simple and use 127.0.0.1:10002 if java.net.preferIPv4Stack=true and [::1]:10002 otherwise.
2. HASingletonElectionPolicyTestCase
This test was failng as the testcase deployment ha-electionpolicy-jboss-beans.xml was not
deploying correctly, due to IPv6 addresses for preferredMaster attribute not being
enclosed in square brackets. Fixed the XML descriptor by replacing <property
name="preferredMaster">${node0.bind.address}:1099</property> with
<property
name="preferredMaster">${node0.bind.address.url}:1099</property>
HASingleton PreferredMasterElectionPolicy failing with IPv6
addresses
---------------------------------------------------------------------
Key: JBAS-8710
URL:
https://jira.jboss.org/browse/JBAS-8710
Project: JBoss Application Server
Issue Type: Bug
Security Level: Public(Everyone can see)
Components: Clustering, Test Suite
Affects Versions: 6.0.0.CR1
Reporter: Richard Achmatowicz
Assignee: Richard Achmatowicz
Fix For: 6.0.0.Final
The bean org.jboss.ha.singleton.PreferredMasterElectionPolicy has an attribute
preferredMaster which identified a process in a cluster which should be the master node in
a clustered partition.
The preferred master attriubute is specified in the form host:port.
This format works fine for IPv4 addresses, but breaks down for IPv6 addresses unless the
host part is surrounded with square brackets, according to rfc2732.
There are two places which above class may break:
(i) in setting the preferred master attribute
(ii) in comparing ClusterNode instances using PreferredMasterElectionPolicy.find() to
check if a given preferred master is in the current cluster.
--
This message is automatically generated by JIRA.
-
For more information on JIRA, see:
http://www.atlassian.com/software/jira