Ron Sigal commented on JBREM-1144:
The need for a server identity has arisen in a support case, so this issue will be
If org.jboss.remoting.Remoting.USE_SERVER_CONNECTION_IDENTITY (actual value
"useServerConnectionIdentity") is set to true, then
org.jboss.remoting.SocketInvoker will create for itself a unique serverId. Moreover, if
an org.jboss.remotinig.ConnectionListener is installed, which leads to the creation of an
org.jboss.remoting.ConnectionValidator, then the ConnectionValidator will get the serverId
from the ServerInvoker with its first PING, and it will compare the serverId returned with
each subsequent PING to the original serverId. If the serverId changes, then the
ServerInvoker has been replaced, and the ConnectionValidator will declare the connection
to be broken.
Note that the use of this feature will change the wire format of the response to the PING
if the ConnectionValidator has "tieToLease" set to true. However,
ConnectionValidator and ServerInvoker are written to accomodate older versions that do not
support this feature.
* unit tests: org.jboss.test.remoting.connection.identity.ServerIdentityTestCase
* versioning tests:
The changes have been applied to the 2.2 and 2.x branches. Waiting for results in
Extend connection identity to server side
Project: JBoss Remoting
Issue Type: Feature Request
Security Level: Public(Everyone can see)
Affects Versions: 2.5.3.SP1, 2.2.3.SP3
Reporter: Ron Sigal
Assignee: Ron Sigal
Fix For: 2.5.3.SP2, 2.2.3.SP4
In JBREM-1128 "Introduce connection identity concept" and JBREM-1132
"CLONE [JBREM-1128] - Introduce connection identity concept", the server was
equipped to recognized when a client had been replaced by another client. In other words,
clients were assigned an identity. This concept should be extended to the server side, so
that ConnectionValidator can detect that a server has been replaced.
This message is automatically generated by JIRA.
For more information on JIRA, see: http://www.atlassian.com/software/jira