[JBoss JIRA] Created: (JBMESSAGING-1128) Message not routed from one node to node with only consumer.
by Mike Clark (JIRA)
Message not routed from one node to node with only consumer.
------------------------------------------------------------
Key: JBMESSAGING-1128
URL: http://jira.jboss.com/jira/browse/JBMESSAGING-1128
Project: JBoss Messaging
Issue Type: Bug
Components: JMS Clustering
Affects Versions: 1.4.0.GA
Environment: Sun JDK 1.5.0_12, JBoss EAP 4.2.0, JBoss Messaging 1.4.GA installed against the "all" configuration running on RHEL5 (CSB), MySQL 5 used as backing database.
Reporter: Mike Clark
Assigned To: Tim Fox
In trying to reproduce behavior described in case 18885, I can reproduce the following:
Start two instances of JBoss (ports-01 and ports-02 using the usual binding manager approach). Deployed a simple MDB to ports-02; simply logs messages with INFO priority. Deployed a simple MBean to both ports-01 and ports-02 that allows me to send TextMessage messages to the respective boxes from the JMX console.
Messages sent to ports-02 are consumed by the MDB as expected. Messages sent to ports-01 are not consumed unless ports-02 is shutdown (normal shutdown) and restarted. The MDB on ports-02 will consume the messages queued to ports-01 regardless of whether ports-01 is up or not (ports-01 shutdown cleanly). Messages sent to ports-01 are, of course, seen in the database as soon as they are submitted (I'm using PERSISTENT delivery mode.)
Both instances were created from a clean "all" instance.
--
This message is automatically generated by JIRA.
-
If you think it was sent incorrectly contact one of the administrators: http://jira.jboss.com/jira/secure/Administrators.jspa
-
For more information on JIRA, see: http://www.atlassian.com/software/jira
18 years, 5 months
[JBoss JIRA] Created: (JBMESSAGING-979) Servlet transport for JBoss Messaging
by Daniel Weeks (JIRA)
Servlet transport for JBoss Messaging
-------------------------------------
Key: JBMESSAGING-979
URL: http://jira.jboss.com/jira/browse/JBMESSAGING-979
Project: JBoss Messaging
Issue Type: Feature Request
Components: JMS Remoting
Affects Versions: 1.2.0.SP1
Environment: Windows 2003 with IIS as front-end web server.
Reporter: Daniel Weeks
Assigned To: Tim Fox
Although there is an https transport for JBoss Messaging, it does not support going over the same port that Tomcat is using like the JBoss MQ jbossmq-httpil.sar does. This capability is required for environments in which port 443 is the only port that is open for communication. Although url proxying/forwarding could be used under Apache to work around this issue, there is not a good solution for doing this with IIS (or for applications that don't use a front-end web server and are limited to port 443). Many enterprise PKI solutions (such as DoD) favor the use of IIS over Apache due to cost of COTS products, so this is a show stopper for JBoss Messaging in those environments.
This issue is preventing us from migrating an enterprise application that runs in the DoD environment to to JBoss Messaging. The application makes heavy use of JMS and would benefit greatly from the improved performance that JBoss Messaging provides.
This feature request is originating from a JBoss Customer Support Portal case at the following URL:
https://network.jboss.com/jbossnetwork/restricted/caseDetail.html?caseId=...
--
This message is automatically generated by JIRA.
-
If you think it was sent incorrectly contact one of the administrators: http://jira.jboss.com/jira/secure/Administrators.jspa
-
For more information on JIRA, see: http://www.atlassian.com/software/jira
18 years, 5 months
[JBoss JIRA] Created: (JBMESSAGING-1107) Less noisy client-side failover
by Ovidiu Feodorov (JIRA)
Less noisy client-side failover
-------------------------------
Key: JBMESSAGING-1107
URL: http://jira.jboss.com/jira/browse/JBMESSAGING-1107
Project: JBoss Messaging
Issue Type: Feature Request
Components: JMS Clustering
Affects Versions: 1.4.0.GA
Reporter: Ovidiu Feodorov
Assigned To: Ovidiu Feodorov
Priority: Minor
When a cluster node failure occurs, the client side reports a significant amount of ERROR logging. This is technically correct, those errors are actually occurring, but a clustered connection factory has in-place a fail-over mechanism whose job is to cope with those errors and hopefully make them disappear.
If the fail-over succeeds, the user should not see any ERROR-level logging. At most, there should be a WARN along the lines of "node failure detected, the fail-over mechanism is dealing with it". The user should see ERRORs only if the fail-over itself fails.
Tons of stack traces produce a significant level of discomfort even if the fail-over succeeds, leaving the user in doubt whether actually the process worked.
The logging information should be naturally still available, but at DEBUG level.
--
This message is automatically generated by JIRA.
-
If you think it was sent incorrectly contact one of the administrators: http://jira.jboss.com/jira/secure/Administrators.jspa
-
For more information on JIRA, see: http://www.atlassian.com/software/jira
18 years, 5 months
[JBoss JIRA] Created: (JBAS-4858) SecurityDeployer changes with metadata
by Anil Saldhana (JIRA)
SecurityDeployer changes with metadata
--------------------------------------
Key: JBAS-4858
URL: http://jira.jboss.com/jira/browse/JBAS-4858
Project: JBoss Application Server
Issue Type: Task
Security Level: Public (Everyone can see)
Components: Security
Affects Versions: JBossAS-5.0.0.Beta2
Reporter: Anil Saldhana
Assigned To: Anil Saldhana
Fix For: JBossAS-5.0.0.Beta3
The Security Deployer needs to be changed as per the following rant from Adrian
=============================
To me this is the wrong approach anyway.
There shouldn't be any code that does
if (ejbdeployment)
doThis();
else if (webdeployment)
doThat();
else if (unknowntype) // OOPS (pun intended ;-)
cantDoIt();
The webservice deployers should be working off some generic metadata that triggers them to operate.
Then each type of deployment can populate that metadata to say "I want a webservice endpoint for this".
This is a seperate deployer for each type we know how to map to a webservice it takes the ejb/web/other metadata and maps or bridges it to the webservice metadata.
It shouldn't just be restricted to ejbs and wars.
e.g. An MBean should be capable of being an endpoint if there is a deployer that creates the relevant metadata attachment from say a META-INF/jboss-webservice.xml in the sar.
The security deployers take the same brain dead (non object orientated) approach.
============================================================
--
This message is automatically generated by JIRA.
-
If you think it was sent incorrectly contact one of the administrators: http://jira.jboss.com/jira/secure/Administrators.jspa
-
For more information on JIRA, see: http://www.atlassian.com/software/jira
18 years, 5 months
[JBoss JIRA] Created: (JBRULES-1229) separate compile classpath from class loader's getResource() facility
by Godmar Back (JIRA)
separate compile classpath from class loader's getResource() facility
---------------------------------------------------------------------
Key: JBRULES-1229
URL: http://jira.jboss.com/jira/browse/JBRULES-1229
Project: JBoss Rules
Issue Type: Feature Request
Security Level: Public (Everyone can see)
Components: Drl Parser/Builder
Affects Versions: 4.0.1
Reporter: Godmar Back
Assigned To: Mark Proctor
Priority: Minor
The current compiler relies on being able to read the bytecode for referred-to classes as a resource. For instance, to resolve the name "org.jboss.pkg.Clazz", a resource "org/jboss/pkg/Clazz.class" must be accessible.
This requirement is not part of the public interface of a class loader. Read the Sun Java documentation, which states: "A resource is some data (images, audio, text, etc) that can be accessed by class code in a way that is independent of the location of the code."
Although, this approach works by coincidence since most class loaders in fact return this resource - at least class loaders that don't do anything fancy to bytecode before defining classes, a more principled solution would separate the classpath used by the runtime engine from the classpath used by the compiler. (This is, for instance, what javac or practically all command-line compilers do.)
My suggestion is to either document this dependency on this particular class loader behavior, or, better, introduce a specific type resolver for the compilation process that is separated from the class loader's resource loading mechanism.
--
This message is automatically generated by JIRA.
-
If you think it was sent incorrectly contact one of the administrators: http://jira.jboss.com/jira/secure/Administrators.jspa
-
For more information on JIRA, see: http://www.atlassian.com/software/jira
18 years, 5 months