[JBoss JIRA] Created: (JBAS-5536) UnifiedInvokerHAProxy replicants are not updated when other nodes undeploy beans
11:01 a.m.
UnifiedInvokerHAProxy replicants are not updated when other nodes undeploy beans
--------------------------------------------------------------------------------
Key: JBAS-5536
URL: http://jira.jboss.com/jira/browse/JBAS-5536
Project: JBoss Application Server
Issue Type: Bug
Security Level: Public (Everyone can see)
Components: Clustering, Remoting
Affects Versions: JBossAS-5.0.0.Beta4, JBossAS-4.2.2.GA
Reporter: Galder Zamarreno
Assigned To: Galder Zamarreno
Priority: Critical
We have a big regression at out hands. When we moved to unified
invokers, we lost the ability for home/bean proxies to be updated
when another node undeploys a bean. Why?
ProxyFactoryHA.replicantsChanged looks something like this
(the trace logs were added by me during debugging):
public synchronized void replicantsChanged (String key,
List newReplicants,
int newReplicantsViewId)
{
try
{
if (homeInvoker instanceof InvokerProxyHA)
{
if (log.isTraceEnabled())
{
log.trace("Updating home proxy with new replicants " +
target.getReplicants() + ", view=" + target.getCurrentViewId ());
}
((InvokerProxyHA)homeInvoker).updateClusterInfo (target.getReplicants(),
target.getCurrentViewId ());
}
if (beanInvoker instanceof InvokerProxyHA)
{
if (log.isTraceEnabled())
{
log.trace("Updating bean proxy with new replicants " +
target.getReplicants() + ", view=" + target.getCurrentViewId ());
}
((InvokerProxyHA)beanInvoker).updateClusterInfo (target.getReplicants(),
target.getCurrentViewId ());
}
log.debug ("Rebinding in JNDI... " + key);
rebindHomeProxy ();
}
catch (Exception none)
{
log.debug (none);
}
}
UnifiedInvokerHAProxy *does not* implement InvokerProxyHA, so unified invoker ha
proxies will never have their replicants updated.
--
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
11:16 a.m.
New subject: [JBoss JIRA] Commented: (JBAS-5536) UnifiedInvokerHAProxy replicants are not updated when other nodes undeploy beans
[ http://jira.jboss.com/jira/browse/JBAS-5536?page=comments#action_12413480 ]
Galder Zamarreno commented on JBAS-5536:
----------------------------------------
Funnily enough, UnifiedInvokerHAProxy does have a updateClusterInfo()
method:
private void updateClusterInfo(ArrayList newReplicants, long currentViewId)
{
if(familyClusterInfo != null)
{
familyClusterInfo.updateClusterInfo(newReplicants, currentViewId);
if (trace)
{
log.trace("Updating cluster info. New view id: " + currentViewId);
log.trace("New cluster target list is:");
for(int x = 0; x < newReplicants.size(); x++)
{
log.trace(newReplicants.get(x));
}
}
}
}
But it does not extend InvokerProxyHA. That's good even though the method is
declared as private cos for a 4.x fix, we'd need to try to avoid changing
UnifiedInvokerHAProxy
at all if possible. This could be achieved making ProxyFactoryHA call updateClusterInfo
via reflection.
For JBoss 5, we just have to make that method public and UnifiedInvokerHAProxy
extend UnifiedInvokerHAProxy. If we want to do this properly for JB 5, we have to debate
the
suitability instanceof calls in ProxyFactoryHA which is rather dubious.
UnifiedInvokerHAProxy replicants are not updated when other nodes
undeploy beans
--------------------------------------------------------------------------------
Key: JBAS-5536
URL: http://jira.jboss.com/jira/browse/JBAS-5536
Project: JBoss Application Server
Issue Type: Bug
Security Level: Public(Everyone can see)
Components: Clustering, Remoting
Affects Versions: JBossAS-5.0.0.Beta4, JBossAS-4.2.2.GA
Reporter: Galder Zamarreno
Assigned To: Galder Zamarreno
Priority: Critical
We have a big regression at out hands. When we moved to unified
invokers, we lost the ability for home/bean proxies to be updated
when another node undeploys a bean. Why?
ProxyFactoryHA.replicantsChanged looks something like this
(the trace logs were added by me during debugging):
public synchronized void replicantsChanged (String key,
List newReplicants,
int newReplicantsViewId)
{
try
{
if (homeInvoker instanceof InvokerProxyHA)
{
if (log.isTraceEnabled())
{
log.trace("Updating home proxy with new replicants " +
target.getReplicants() + ", view=" + target.getCurrentViewId ());
}
((InvokerProxyHA)homeInvoker).updateClusterInfo (target.getReplicants(),
target.getCurrentViewId ());
}
if (beanInvoker instanceof InvokerProxyHA)
{
if (log.isTraceEnabled())
{
log.trace("Updating bean proxy with new replicants " +
target.getReplicants() + ", view=" + target.getCurrentViewId ());
}
((InvokerProxyHA)beanInvoker).updateClusterInfo (target.getReplicants(),
target.getCurrentViewId ());
}
log.debug ("Rebinding in JNDI... " + key);
rebindHomeProxy ();
}
catch (Exception none)
{
log.debug (none);
}
}
UnifiedInvokerHAProxy *does not* implement InvokerProxyHA, so unified invoker ha
proxies will never have their replicants updated.
--
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
11:24 a.m.
New subject: [JBoss JIRA] Updated: (JBAS-5536) UnifiedInvokerHAProxy replicants are not updated when other nodes undeploy beans
[ http://jira.jboss.com/jira/browse/JBAS-5536?page=all ]
Galder Zamarreno updated JBAS-5536:
-----------------------------------
Workaround Description:
The obvious workaround is to use JRMP invokers instead of Unified invokers while this
gets sorted as JRMP invoker proxy ha does implement InvokerProxyHA. To do that, take
your EJB and add the following to the beans jboss.xml so that it uses JRMP invoker instead
of Unified:
<jboss>
<enterprise-beans>
<session>
<ejb-name>Boo</ejb-name>
<jndi-name>ejb/BooEjb</jndi-name>
<configuration-name>JRMP Clustered Stateless
SessionBean</configuration-name>
<clustered>true</clustered>
</session>
</enterprise-beans>
<container-configurations>
<container-configuration extends="Clustered Stateless SessionBean">
<container-name>JRMP Clustered Stateless
SessionBean</container-name>
<invoker-proxy-binding-name>clustered-stateless-rmi-invoker</invoker-proxy-binding-name>
</container-configuration>
</container-configurations>
</jboss>
was:
The obvious workaround is to use JRMP invokers instead of Unified invokers while this
gets sorted as JRMP invoker proxy ha does implement InvokerProxyHA. To do that, take
your EJB and add the following so that it uses JRMP invoker instead of Unified:
<jboss>
<enterprise-beans>
<session>
<ejb-name>Boo</ejb-name>
<jndi-name>ejb/BooEjb</jndi-name>
<configuration-name>JRMP Clustered Stateless
SessionBean</configuration-name>
<clustered>true</clustered>
</session>
</enterprise-beans>
<container-configurations>
<container-configuration extends="Clustered Stateless SessionBean">
<container-name>JRMP Clustered Stateless
SessionBean</container-name>
<invoker-proxy-binding-name>clustered-stateless-rmi-invoker</invoker-proxy-binding-name>
</container-configuration>
</container-configurations>
UnifiedInvokerHAProxy replicants are not updated when other nodes
undeploy beans
--------------------------------------------------------------------------------
Key: JBAS-5536
URL: http://jira.jboss.com/jira/browse/JBAS-5536
Project: JBoss Application Server
Issue Type: Bug
Security Level: Public(Everyone can see)
Components: Clustering, Remoting
Affects Versions: JBossAS-5.0.0.Beta4, JBossAS-4.2.2.GA
Reporter: Galder Zamarreno
Assigned To: Galder Zamarreno
Priority: Critical
We have a big regression at out hands. When we moved to unified
invokers, we lost the ability for home/bean proxies to be updated
when another node undeploys a bean. Why?
ProxyFactoryHA.replicantsChanged looks something like this
(the trace logs were added by me during debugging):
public synchronized void replicantsChanged (String key,
List newReplicants,
int newReplicantsViewId)
{
try
{
if (homeInvoker instanceof InvokerProxyHA)
{
if (log.isTraceEnabled())
{
log.trace("Updating home proxy with new replicants " +
target.getReplicants() + ", view=" + target.getCurrentViewId ());
}
((InvokerProxyHA)homeInvoker).updateClusterInfo (target.getReplicants(),
target.getCurrentViewId ());
}
if (beanInvoker instanceof InvokerProxyHA)
{
if (log.isTraceEnabled())
{
log.trace("Updating bean proxy with new replicants " +
target.getReplicants() + ", view=" + target.getCurrentViewId ());
}
((InvokerProxyHA)beanInvoker).updateClusterInfo (target.getReplicants(),
target.getCurrentViewId ());
}
log.debug ("Rebinding in JNDI... " + key);
rebindHomeProxy ();
}
catch (Exception none)
{
log.debug (none);
}
}
UnifiedInvokerHAProxy *does not* implement InvokerProxyHA, so unified invoker ha
proxies will never have their replicants updated.
--
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
12:30 p.m.
New subject: [JBoss JIRA] Closed: (JBAS-5536) UnifiedInvokerHAProxy replicants are not updated when other nodes undeploy beans
[ http://jira.jboss.com/jira/browse/JBAS-5536?page=all ]
Galder Zamarreno closed JBAS-5536.
----------------------------------
Fix Version/s: JBossAS-5.0.0.Beta4
JBossAS-4.2.2.GA
Resolution: Duplicate Issue
UnifiedInvokerHAProxy replicants are not updated when other nodes
undeploy beans
--------------------------------------------------------------------------------
Key: JBAS-5536
URL: http://jira.jboss.com/jira/browse/JBAS-5536
Project: JBoss Application Server
Issue Type: Bug
Security Level: Public(Everyone can see)
Components: Clustering, Remoting
Affects Versions: JBossAS-5.0.0.Beta3, JBossAS-4.2.1.GA
Reporter: Galder Zamarreno
Assigned To: Galder Zamarreno
Priority: Critical
Fix For: JBossAS-5.0.0.Beta4, JBossAS-4.2.2.GA
We have a big regression at out hands. When we moved to unified
invokers, we lost the ability for home/bean proxies to be updated
when another node undeploys a bean. Why?
ProxyFactoryHA.replicantsChanged looks something like this
(the trace logs were added by me during debugging):
public synchronized void replicantsChanged (String key,
List newReplicants,
int newReplicantsViewId)
{
try
{
if (homeInvoker instanceof InvokerProxyHA)
{
if (log.isTraceEnabled())
{
log.trace("Updating home proxy with new replicants " +
target.getReplicants() + ", view=" + target.getCurrentViewId ());
}
((InvokerProxyHA)homeInvoker).updateClusterInfo (target.getReplicants(),
target.getCurrentViewId ());
}
if (beanInvoker instanceof InvokerProxyHA)
{
if (log.isTraceEnabled())
{
log.trace("Updating bean proxy with new replicants " +
target.getReplicants() + ", view=" + target.getCurrentViewId ());
}
((InvokerProxyHA)beanInvoker).updateClusterInfo (target.getReplicants(),
target.getCurrentViewId ());
}
log.debug ("Rebinding in JNDI... " + key);
rebindHomeProxy ();
}
catch (Exception none)
{
log.debug (none);
}
}
UnifiedInvokerHAProxy *does not* implement InvokerProxyHA, so unified invoker ha
proxies will never have their replicants updated.
--
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
12:30 p.m.
New subject: [JBoss JIRA] Updated: (JBAS-5536) UnifiedInvokerHAProxy replicants are not updated when other nodes undeploy beans
[ http://jira.jboss.com/jira/browse/JBAS-5536?page=all ]
Galder Zamarreno updated JBAS-5536:
-----------------------------------
Affects Version/s: JBossAS-5.0.0.Beta3
JBossAS-4.2.1.GA
(was: JBossAS-5.0.0.Beta4)
(was: JBossAS-4.2.2.GA)
UnifiedInvokerHAProxy replicants are not updated when other nodes
undeploy beans
--------------------------------------------------------------------------------
Key: JBAS-5536
URL: http://jira.jboss.com/jira/browse/JBAS-5536
Project: JBoss Application Server
Issue Type: Bug
Security Level: Public(Everyone can see)
Components: Clustering, Remoting
Affects Versions: JBossAS-5.0.0.Beta3, JBossAS-4.2.1.GA
Reporter: Galder Zamarreno
Assigned To: Galder Zamarreno
Priority: Critical
Fix For: JBossAS-5.0.0.Beta4, JBossAS-4.2.2.GA
We have a big regression at out hands. When we moved to unified
invokers, we lost the ability for home/bean proxies to be updated
when another node undeploys a bean. Why?
ProxyFactoryHA.replicantsChanged looks something like this
(the trace logs were added by me during debugging):
public synchronized void replicantsChanged (String key,
List newReplicants,
int newReplicantsViewId)
{
try
{
if (homeInvoker instanceof InvokerProxyHA)
{
if (log.isTraceEnabled())
{
log.trace("Updating home proxy with new replicants " +
target.getReplicants() + ", view=" + target.getCurrentViewId ());
}
((InvokerProxyHA)homeInvoker).updateClusterInfo (target.getReplicants(),
target.getCurrentViewId ());
}
if (beanInvoker instanceof InvokerProxyHA)
{
if (log.isTraceEnabled())
{
log.trace("Updating bean proxy with new replicants " +
target.getReplicants() + ", view=" + target.getCurrentViewId ());
}
((InvokerProxyHA)beanInvoker).updateClusterInfo (target.getReplicants(),
target.getCurrentViewId ());
}
log.debug ("Rebinding in JNDI... " + key);
rebindHomeProxy ();
}
catch (Exception none)
{
log.debug (none);
}
}
UnifiedInvokerHAProxy *does not* implement InvokerProxyHA, so unified invoker ha
proxies will never have their replicants updated.
--
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
4:30 p.m.
New subject: [JBoss JIRA] Commented: (JBAS-5536) UnifiedInvokerHAProxy replicants are not updated when other nodes undeploy beans
[ http://jira.jboss.com/jira/browse/JBAS-5536?page=comments#action_12413686 ]
Galder Zamarreno commented on JBAS-5536:
----------------------------------------
It's also worth noting that this has effects on start up of
nodes:
1.- Let's say node A starts up deploying EJB A, the proxy will
only be pointing to node A.
2.- Then, node B starts and deploys EJB A. Node A should
receive an update indicating that node B has EJB A but because
of this bug, the proxy won't be updated, so the proxy for EJB A
in node A will think it's the only owner.
3.- If you make several calls to EJB A in node A, these should
be load balanced as Round Robin by default, but this because
of this bug, it'd look as if it was First Available.
UnifiedInvokerHAProxy replicants are not updated when other nodes
undeploy beans
--------------------------------------------------------------------------------
Key: JBAS-5536
URL: http://jira.jboss.com/jira/browse/JBAS-5536
Project: JBoss Application Server
Issue Type: Bug
Security Level: Public(Everyone can see)
Components: Clustering, Remoting
Affects Versions: JBossAS-5.0.0.Beta3, JBossAS-4.2.1.GA
Reporter: Galder Zamarreno
Assigned To: Galder Zamarreno
Priority: Critical
Fix For: JBossAS-5.0.0.Beta4, JBossAS-4.2.2.GA
We have a big regression at out hands. When we moved to unified
invokers, we lost the ability for home/bean proxies to be updated
when another node undeploys a bean. Why?
ProxyFactoryHA.replicantsChanged looks something like this
(the trace logs were added by me during debugging):
public synchronized void replicantsChanged (String key,
List newReplicants,
int newReplicantsViewId)
{
try
{
if (homeInvoker instanceof InvokerProxyHA)
{
if (log.isTraceEnabled())
{
log.trace("Updating home proxy with new replicants " +
target.getReplicants() + ", view=" + target.getCurrentViewId ());
}
((InvokerProxyHA)homeInvoker).updateClusterInfo (target.getReplicants(),
target.getCurrentViewId ());
}
if (beanInvoker instanceof InvokerProxyHA)
{
if (log.isTraceEnabled())
{
log.trace("Updating bean proxy with new replicants " +
target.getReplicants() + ", view=" + target.getCurrentViewId ());
}
((InvokerProxyHA)beanInvoker).updateClusterInfo (target.getReplicants(),
target.getCurrentViewId ());
}
log.debug ("Rebinding in JNDI... " + key);
rebindHomeProxy ();
}
catch (Exception none)
{
log.debug (none);
}
}
UnifiedInvokerHAProxy *does not* implement InvokerProxyHA, so unified invoker ha
proxies will never have their replicants updated.
--
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
4:10 p.m.
New subject: [JBoss JIRA] Reopened: (JBAS-5536) UnifiedInvokerHAProxy replicants are not updated when other nodes undeploy beans
[ http://jira.jboss.com/jira/browse/JBAS-5536?page=all ]
Galder Zamarreno reopened JBAS-5536:
------------------------------------
UnifiedInvokerHAProxy replicants are not updated when other nodes
undeploy beans
--------------------------------------------------------------------------------
Key: JBAS-5536
URL: http://jira.jboss.com/jira/browse/JBAS-5536
Project: JBoss Application Server
Issue Type: Bug
Security Level: Public(Everyone can see)
Components: Clustering, Remoting
Affects Versions: JBossAS-5.0.0.Beta3, JBossAS-4.2.1.GA
Reporter: Galder Zamarreno
Assigned To: Galder Zamarreno
Priority: Critical
Fix For: JBossAS-5.0.0.Beta4, JBossAS-4.2.2.GA
We have a big regression at out hands. When we moved to unified
invokers, we lost the ability for home/bean proxies to be updated
when another node undeploys a bean. Why?
ProxyFactoryHA.replicantsChanged looks something like this
(the trace logs were added by me during debugging):
public synchronized void replicantsChanged (String key,
List newReplicants,
int newReplicantsViewId)
{
try
{
if (homeInvoker instanceof InvokerProxyHA)
{
if (log.isTraceEnabled())
{
log.trace("Updating home proxy with new replicants " +
target.getReplicants() + ", view=" + target.getCurrentViewId ());
}
((InvokerProxyHA)homeInvoker).updateClusterInfo (target.getReplicants(),
target.getCurrentViewId ());
}
if (beanInvoker instanceof InvokerProxyHA)
{
if (log.isTraceEnabled())
{
log.trace("Updating bean proxy with new replicants " +
target.getReplicants() + ", view=" + target.getCurrentViewId ());
}
((InvokerProxyHA)beanInvoker).updateClusterInfo (target.getReplicants(),
target.getCurrentViewId ());
}
log.debug ("Rebinding in JNDI... " + key);
rebindHomeProxy ();
}
catch (Exception none)
{
log.debug (none);
}
}
UnifiedInvokerHAProxy *does not* implement InvokerProxyHA, so unified invoker ha
proxies will never have their replicants updated.
--
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
4:12 p.m.
New subject: [JBoss JIRA] Updated: (JBAS-5536) UnifiedInvokerHAProxy replicants are not updated when other nodes undeploy beans
[ http://jira.jboss.com/jira/browse/JBAS-5536?page=all ]
Galder Zamarreno updated JBAS-5536:
-----------------------------------
Workaround Description:
The obvious workaround is to use JRMP invokers instead of Unified invokers while this
gets sorted as JRMP invoker proxy ha does implement InvokerProxyHA. To do that, take
your EJB and add the following to the beans jboss.xml so that it uses JRMP invoker instead
of Unified:
<jboss>
<enterprise-beans>
<session>
<ejb-name>Boo</ejb-name>
<jndi-name>ejb/BooEjb</jndi-name>
<configuration-name>JRMP Clustered Stateless
SessionBean</configuration-name>
<clustered>true</clustered>
</session>
</enterprise-beans>
<container-configurations>
<container-configuration extends="Clustered Stateless SessionBean">
<container-name>JRMP Clustered Stateless
SessionBean</container-name>
<invoker-proxy-binding-name>clustered-stateless-rmi-invoker</invoker-proxy-binding-name>
</container-configuration>
</container-configurations>
</jboss>
If you wanna apply the workaround to all your beans, apply these changes to the
"Clustered Stateless SessionBean" configuration in conf/standardjboss.xml
was:
The obvious workaround is to use JRMP invokers instead of Unified invokers while this
gets sorted as JRMP invoker proxy ha does implement InvokerProxyHA. To do that, take
your EJB and add the following to the beans jboss.xml so that it uses JRMP invoker instead
of Unified:
<jboss>
<enterprise-beans>
<session>
<ejb-name>Boo</ejb-name>
<jndi-name>ejb/BooEjb</jndi-name>
<configuration-name>JRMP Clustered Stateless
SessionBean</configuration-name>
<clustered>true</clustered>
</session>
</enterprise-beans>
<container-configurations>
<container-configuration extends="Clustered Stateless SessionBean">
<container-name>JRMP Clustered Stateless
SessionBean</container-name>
<invoker-proxy-binding-name>clustered-stateless-rmi-invoker</invoker-proxy-binding-name>
</container-configuration>
</container-configurations>
</jboss>
UnifiedInvokerHAProxy replicants are not updated when other nodes
undeploy beans
--------------------------------------------------------------------------------
Key: JBAS-5536
URL: http://jira.jboss.com/jira/browse/JBAS-5536
Project: JBoss Application Server
Issue Type: Bug
Security Level: Public(Everyone can see)
Components: Clustering, Remoting
Affects Versions: JBossAS-5.0.0.Beta3, JBossAS-4.2.1.GA
Reporter: Galder Zamarreno
Assigned To: Galder Zamarreno
Priority: Critical
Fix For: JBossAS-5.0.0.Beta4, JBossAS-4.2.2.GA
We have a big regression at out hands. When we moved to unified
invokers, we lost the ability for home/bean proxies to be updated
when another node undeploys a bean. Why?
ProxyFactoryHA.replicantsChanged looks something like this
(the trace logs were added by me during debugging):
public synchronized void replicantsChanged (String key,
List newReplicants,
int newReplicantsViewId)
{
try
{
if (homeInvoker instanceof InvokerProxyHA)
{
if (log.isTraceEnabled())
{
log.trace("Updating home proxy with new replicants " +
target.getReplicants() + ", view=" + target.getCurrentViewId ());
}
((InvokerProxyHA)homeInvoker).updateClusterInfo (target.getReplicants(),
target.getCurrentViewId ());
}
if (beanInvoker instanceof InvokerProxyHA)
{
if (log.isTraceEnabled())
{
log.trace("Updating bean proxy with new replicants " +
target.getReplicants() + ", view=" + target.getCurrentViewId ());
}
((InvokerProxyHA)beanInvoker).updateClusterInfo (target.getReplicants(),
target.getCurrentViewId ());
}
log.debug ("Rebinding in JNDI... " + key);
rebindHomeProxy ();
}
catch (Exception none)
{
log.debug (none);
}
}
UnifiedInvokerHAProxy *does not* implement InvokerProxyHA, so unified invoker ha
proxies will never have their replicants updated.
--
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
4:12 p.m.
New subject: [JBoss JIRA] Closed: (JBAS-5536) UnifiedInvokerHAProxy replicants are not updated when other nodes undeploy beans
[ http://jira.jboss.com/jira/browse/JBAS-5536?page=all ]
Galder Zamarreno closed JBAS-5536.
----------------------------------
Resolution: Done
UnifiedInvokerHAProxy replicants are not updated when other nodes
undeploy beans
--------------------------------------------------------------------------------
Key: JBAS-5536
URL: http://jira.jboss.com/jira/browse/JBAS-5536
Project: JBoss Application Server
Issue Type: Bug
Security Level: Public(Everyone can see)
Components: Clustering, Remoting
Affects Versions: JBossAS-5.0.0.Beta3, JBossAS-4.2.1.GA
Reporter: Galder Zamarreno
Assigned To: Galder Zamarreno
Priority: Critical
Fix For: JBossAS-5.0.0.Beta4, JBossAS-4.2.2.GA
We have a big regression at out hands. When we moved to unified
invokers, we lost the ability for home/bean proxies to be updated
when another node undeploys a bean. Why?
ProxyFactoryHA.replicantsChanged looks something like this
(the trace logs were added by me during debugging):
public synchronized void replicantsChanged (String key,
List newReplicants,
int newReplicantsViewId)
{
try
{
if (homeInvoker instanceof InvokerProxyHA)
{
if (log.isTraceEnabled())
{
log.trace("Updating home proxy with new replicants " +
target.getReplicants() + ", view=" + target.getCurrentViewId ());
}
((InvokerProxyHA)homeInvoker).updateClusterInfo (target.getReplicants(),
target.getCurrentViewId ());
}
if (beanInvoker instanceof InvokerProxyHA)
{
if (log.isTraceEnabled())
{
log.trace("Updating bean proxy with new replicants " +
target.getReplicants() + ", view=" + target.getCurrentViewId ());
}
((InvokerProxyHA)beanInvoker).updateClusterInfo (target.getReplicants(),
target.getCurrentViewId ());
}
log.debug ("Rebinding in JNDI... " + key);
rebindHomeProxy ();
}
catch (Exception none)
{
log.debug (none);
}
}
UnifiedInvokerHAProxy *does not* implement InvokerProxyHA, so unified invoker ha
proxies will never have their replicants updated.
--
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
6065
days inactive
6093
days old
8 comments
1 participants
participants (1)
-
Galder Zamarreno (JIRA)