[JBoss JIRA] Created: (JBAS-4455) LoadBalancePolicy that tries to pin all requests associated with a tx to one server
12:30 p.m.
LoadBalancePolicy that tries to pin all requests associated with a tx to one server
-----------------------------------------------------------------------------------
Key: JBAS-4455
URL: http://jira.jboss.com/jira/browse/JBAS-4455
Project: JBoss Application Server
Issue Type: Feature Request
Security Level: Public (Everyone can see)
Components: Clustering
Reporter: Brian Stansberry
Assigned To: Galder Zamarreno
Priority: Minor
The HA proxies don't allow failover once a tx has reached a server. This can lead to
this kind of situation:
EJB A deployed on nodes 1 and 2.
EJB B deployed on nodes 1 and 2.
1) Start tx, invoke on A.1.
2) Lookup B.
3) Invoke on B. LB policy picks B.2
4) Node 2 is dead for some reason so call fails.
5) Can't fail over because tx context won't allow a failover after a call has
reached a server.
But, B.1 is fine and is running on the only node the tx has invoked on. :(
Approach to improving this. This is based on JRMPInovkerProxyHA methods, but the idea is
generic:
In invocationHasReachedAServer, instead of storing null as the value, you store the target
(key is the tx).
In invoke, if there's a tx, you add that target (if there is one) to the invocation as
transient metadata
The LB policy gets passed the invocation as an arg when it chooses a target (this is
already in the API)
LB policy checks for the metadata. If there, and that target is in its target list, return
that target
--
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
8:28 a.m.
New subject: [JBoss JIRA] Updated: (JBAS-4455) LoadBalancePolicy that tries to pin all requests associated with a tx to one server
[ http://jira.jboss.com/jira/browse/JBAS-4455?page=all ]
Galder Zamarreno updated JBAS-4455:
-----------------------------------
JBoss Forum Reference:
http://www.jboss.com/index.html?module=bb&op=viewtopic&t=117013
LoadBalancePolicy that tries to pin all requests associated with a tx
to one server
-----------------------------------------------------------------------------------
Key: JBAS-4455
URL: http://jira.jboss.com/jira/browse/JBAS-4455
Project: JBoss Application Server
Issue Type: Feature Request
Security Level: Public(Everyone can see)
Components: Clustering
Reporter: Brian Stansberry
Assigned To: Galder Zamarreno
Priority: Minor
The HA proxies don't allow failover once a tx has reached a server. This can lead to
this kind of situation:
EJB A deployed on nodes 1 and 2.
EJB B deployed on nodes 1 and 2.
1) Start tx, invoke on A.1.
2) Lookup B.
3) Invoke on B. LB policy picks B.2
4) Node 2 is dead for some reason so call fails.
5) Can't fail over because tx context won't allow a failover after a call has
reached a server.
But, B.1 is fine and is running on the only node the tx has invoked on. :(
Approach to improving this. This is based on JRMPInovkerProxyHA methods, but the idea is
generic:
In invocationHasReachedAServer, instead of storing null as the value, you store the
target (key is the tx).
In invoke, if there's a tx, you add that target (if there is one) to the invocation
as transient metadata
The LB policy gets passed the invocation as an arg when it chooses a target (this is
already in the API)
LB policy checks for the metadata. If there, and that target is in its target list,
return that target
--
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
6:08 a.m.
New subject: [JBoss JIRA] Commented: (JBAS-4455) LoadBalancePolicy that tries to pin all requests associated with a tx to one server
[ http://jira.jboss.com/jira/browse/JBAS-4455?page=comments#action_12382040 ]
Galder Zamarreno commented on JBAS-4455:
----------------------------------------
[private]
Another potential customer interested in this feature:
https://na1.salesforce.com/50030000003i3bC
LoadBalancePolicy that tries to pin all requests associated with a tx
to one server
-----------------------------------------------------------------------------------
Key: JBAS-4455
URL: http://jira.jboss.com/jira/browse/JBAS-4455
Project: JBoss Application Server
Issue Type: Feature Request
Security Level: Public(Everyone can see)
Components: Clustering
Reporter: Brian Stansberry
Assigned To: Galder Zamarreno
Priority: Minor
The HA proxies don't allow failover once a tx has reached a server. This can lead to
this kind of situation:
EJB A deployed on nodes 1 and 2.
EJB B deployed on nodes 1 and 2.
1) Start tx, invoke on A.1.
2) Lookup B.
3) Invoke on B. LB policy picks B.2
4) Node 2 is dead for some reason so call fails.
5) Can't fail over because tx context won't allow a failover after a call has
reached a server.
But, B.1 is fine and is running on the only node the tx has invoked on. :(
Approach to improving this. This is based on JRMPInovkerProxyHA methods, but the idea is
generic:
In invocationHasReachedAServer, instead of storing null as the value, you store the
target (key is the tx).
In invoke, if there's a tx, you add that target (if there is one) to the invocation
as transient metadata
The LB policy gets passed the invocation as an arg when it chooses a target (this is
already in the API)
LB policy checks for the metadata. If there, and that target is in its target list,
return that target
--
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
1:18 p.m.
New subject: [JBoss JIRA] Commented: (JBAS-4455) LoadBalancePolicy that tries to pin all requests associated with a tx to one server
[ http://jira.jboss.com/jira/browse/JBAS-4455?page=comments#action_12388021 ]
Galder Zamarreno commented on JBAS-4455:
----------------------------------------
The implementation of this jira can now carry on for EJB2, but for
EJB3, the following forum post needs to be resolved:
http://www.jboss.com/index.html?module=bb&op=viewtopic&t=123798
LoadBalancePolicy that tries to pin all requests associated with a tx
to one server
-----------------------------------------------------------------------------------
Key: JBAS-4455
URL: http://jira.jboss.com/jira/browse/JBAS-4455
Project: JBoss Application Server
Issue Type: Feature Request
Security Level: Public(Everyone can see)
Components: Clustering
Reporter: Brian Stansberry
Assigned To: Galder Zamarreno
Priority: Minor
The HA proxies don't allow failover once a tx has reached a server. This can lead to
this kind of situation:
EJB A deployed on nodes 1 and 2.
EJB B deployed on nodes 1 and 2.
1) Start tx, invoke on A.1.
2) Lookup B.
3) Invoke on B. LB policy picks B.2
4) Node 2 is dead for some reason so call fails.
5) Can't fail over because tx context won't allow a failover after a call has
reached a server.
But, B.1 is fine and is running on the only node the tx has invoked on. :(
Approach to improving this. This is based on JRMPInovkerProxyHA methods, but the idea is
generic:
In invocationHasReachedAServer, instead of storing null as the value, you store the
target (key is the tx).
In invoke, if there's a tx, you add that target (if there is one) to the invocation
as transient metadata
The LB policy gets passed the invocation as an arg when it chooses a target (this is
already in the API)
LB policy checks for the metadata. If there, and that target is in its target list,
return that target
--
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
12:12 p.m.
New subject: [JBoss JIRA] Updated: (JBAS-4455) LoadBalancePolicy that tries to pin all requests associated with a tx to one server
[ http://jira.jboss.com/jira/browse/JBAS-4455?page=all ]
Galder Zamarreno updated JBAS-4455:
-----------------------------------
Affects: [Documentation (Ref Guide, User Guide, etc.)]
LoadBalancePolicy that tries to pin all requests associated with a tx
to one server
-----------------------------------------------------------------------------------
Key: JBAS-4455
URL: http://jira.jboss.com/jira/browse/JBAS-4455
Project: JBoss Application Server
Issue Type: Feature Request
Security Level: Public(Everyone can see)
Components: Clustering
Reporter: Brian Stansberry
Assigned To: Galder Zamarreno
Priority: Minor
The HA proxies don't allow failover once a tx has reached a server. This can lead to
this kind of situation:
EJB A deployed on nodes 1 and 2.
EJB B deployed on nodes 1 and 2.
1) Start tx, invoke on A.1.
2) Lookup B.
3) Invoke on B. LB policy picks B.2
4) Node 2 is dead for some reason so call fails.
5) Can't fail over because tx context won't allow a failover after a call has
reached a server.
But, B.1 is fine and is running on the only node the tx has invoked on. :(
Approach to improving this. This is based on JRMPInovkerProxyHA methods, but the idea is
generic:
In invocationHasReachedAServer, instead of storing null as the value, you store the
target (key is the tx).
In invoke, if there's a tx, you add that target (if there is one) to the invocation
as transient metadata
The LB policy gets passed the invocation as an arg when it chooses a target (this is
already in the API)
LB policy checks for the metadata. If there, and that target is in its target list,
return that target
--
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
1:55 p.m.
New subject: [JBoss JIRA] Updated: (JBAS-4455) LoadBalancePolicy that tries to pin all requests associated with a tx to one server
[ http://jira.jboss.com/jira/browse/JBAS-4455?page=all ]
Galder Zamarreno updated JBAS-4455:
-----------------------------------
Fix Version/s: JBossAS-5.0.0.Beta3
LoadBalancePolicy that tries to pin all requests associated with a tx
to one server
-----------------------------------------------------------------------------------
Key: JBAS-4455
URL: http://jira.jboss.com/jira/browse/JBAS-4455
Project: JBoss Application Server
Issue Type: Feature Request
Security Level: Public(Everyone can see)
Components: Clustering
Reporter: Brian Stansberry
Assigned To: Galder Zamarreno
Priority: Minor
Fix For: JBossAS-5.0.0.Beta3
The HA proxies don't allow failover once a tx has reached a server. This can lead to
this kind of situation:
EJB A deployed on nodes 1 and 2.
EJB B deployed on nodes 1 and 2.
1) Start tx, invoke on A.1.
2) Lookup B.
3) Invoke on B. LB policy picks B.2
4) Node 2 is dead for some reason so call fails.
5) Can't fail over because tx context won't allow a failover after a call has
reached a server.
But, B.1 is fine and is running on the only node the tx has invoked on. :(
Approach to improving this. This is based on JRMPInovkerProxyHA methods, but the idea is
generic:
In invocationHasReachedAServer, instead of storing null as the value, you store the
target (key is the tx).
In invoke, if there's a tx, you add that target (if there is one) to the invocation
as transient metadata
The LB policy gets passed the invocation as an arg when it chooses a target (this is
already in the API)
LB policy checks for the metadata. If there, and that target is in its target list,
return that target
--
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
1:55 p.m.
New subject: [JBoss JIRA] Commented: (JBAS-4455) LoadBalancePolicy that tries to pin all requests associated with a tx to one server
[ http://jira.jboss.com/jira/browse/JBAS-4455?page=comments#action_12389859 ]
Galder Zamarreno commented on JBAS-4455:
----------------------------------------
This feature is completed and has been committed in trunk for
EJB2 with JRMP and Unified invokers.
I'm not gonna be able to touch this for a couple of weeks, so
marking the some todos:
1.- test with legacy pooled invoker, which uses the JRMP invoker
proxy ha, so don't foresee any issues there.
2.- add some tests where the new transaction sticky lbps work as the
default ones when not running within a transaction. Run a quick test
and this works fine.
3.- make further comments to the classes created.
4.- port this feature to 4.2.x.
5.- create subtasks for documentation for 4.2.x and trunk
6.- create an EJB3 jira feature request to track down the EJB3 version
of this feature.
LoadBalancePolicy that tries to pin all requests associated with a tx
to one server
-----------------------------------------------------------------------------------
Key: JBAS-4455
URL: http://jira.jboss.com/jira/browse/JBAS-4455
Project: JBoss Application Server
Issue Type: Feature Request
Security Level: Public(Everyone can see)
Components: Clustering
Reporter: Brian Stansberry
Assigned To: Galder Zamarreno
Priority: Minor
Fix For: JBossAS-5.0.0.Beta3
The HA proxies don't allow failover once a tx has reached a server. This can lead to
this kind of situation:
EJB A deployed on nodes 1 and 2.
EJB B deployed on nodes 1 and 2.
1) Start tx, invoke on A.1.
2) Lookup B.
3) Invoke on B. LB policy picks B.2
4) Node 2 is dead for some reason so call fails.
5) Can't fail over because tx context won't allow a failover after a call has
reached a server.
But, B.1 is fine and is running on the only node the tx has invoked on. :(
Approach to improving this. This is based on JRMPInovkerProxyHA methods, but the idea is
generic:
In invocationHasReachedAServer, instead of storing null as the value, you store the
target (key is the tx).
In invoke, if there's a tx, you add that target (if there is one) to the invocation
as transient metadata
The LB policy gets passed the invocation as an arg when it chooses a target (this is
already in the API)
LB policy checks for the metadata. If there, and that target is in its target list,
return that target
--
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
11:43 a.m.
New subject: [JBoss JIRA] Commented: (JBAS-4455) LoadBalancePolicy that tries to pin all requests associated with a tx to one server
[ http://jira.jboss.com/jira/browse/JBAS-4455?page=comments#action_12390316 ]
Galder Zamarreno commented on JBAS-4455:
----------------------------------------
2, 3 and 6 todos have been done.
Still todo:
1.- test with legacy pooled invoker, which uses the JRMP invoker
proxy ha, so don't foresee any issues there.
4.- port this feature to 4.2.x.
5.- create subtasks for documentation for 4.2.x and trunk
LoadBalancePolicy that tries to pin all requests associated with a tx
to one server
-----------------------------------------------------------------------------------
Key: JBAS-4455
URL: http://jira.jboss.com/jira/browse/JBAS-4455
Project: JBoss Application Server
Issue Type: Feature Request
Security Level: Public(Everyone can see)
Components: Clustering
Reporter: Brian Stansberry
Assigned To: Galder Zamarreno
Priority: Minor
Fix For: JBossAS-5.0.0.Beta3
The HA proxies don't allow failover once a tx has reached a server. This can lead to
this kind of situation:
EJB A deployed on nodes 1 and 2.
EJB B deployed on nodes 1 and 2.
1) Start tx, invoke on A.1.
2) Lookup B.
3) Invoke on B. LB policy picks B.2
4) Node 2 is dead for some reason so call fails.
5) Can't fail over because tx context won't allow a failover after a call has
reached a server.
But, B.1 is fine and is running on the only node the tx has invoked on. :(
Approach to improving this. This is based on JRMPInovkerProxyHA methods, but the idea is
generic:
In invocationHasReachedAServer, instead of storing null as the value, you store the
target (key is the tx).
In invoke, if there's a tx, you add that target (if there is one) to the invocation
as transient metadata
The LB policy gets passed the invocation as an arg when it chooses a target (this is
already in the API)
LB policy checks for the metadata. If there, and that target is in its target list,
return that target
--
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
1 p.m.
New subject: [JBoss JIRA] Commented: (JBAS-4455) LoadBalancePolicy that tries to pin all requests associated with a tx to one server
[ http://jira.jboss.com/jira/browse/JBAS-4455?page=comments#action_12391615 ]
Galder Zamarreno commented on JBAS-4455:
----------------------------------------
Todo 1 was tested last week and works as expected with
ha pooled invoker because it uses the ha JRMP proxy which
delegates to the pooled invoker proxy. No changes needed
there, so won't be adding changes for this.
Still todo:
4.- port this feature to 4.2.x.
5.- create subtasks for documentation for 4.2.x and trunk
LoadBalancePolicy that tries to pin all requests associated with a tx
to one server
-----------------------------------------------------------------------------------
Key: JBAS-4455
URL: http://jira.jboss.com/jira/browse/JBAS-4455
Project: JBoss Application Server
Issue Type: Feature Request
Security Level: Public(Everyone can see)
Components: Clustering
Reporter: Brian Stansberry
Assigned To: Galder Zamarreno
Priority: Minor
Fix For: JBossAS-5.0.0.Beta3
The HA proxies don't allow failover once a tx has reached a server. This can lead to
this kind of situation:
EJB A deployed on nodes 1 and 2.
EJB B deployed on nodes 1 and 2.
1) Start tx, invoke on A.1.
2) Lookup B.
3) Invoke on B. LB policy picks B.2
4) Node 2 is dead for some reason so call fails.
5) Can't fail over because tx context won't allow a failover after a call has
reached a server.
But, B.1 is fine and is running on the only node the tx has invoked on. :(
Approach to improving this. This is based on JRMPInovkerProxyHA methods, but the idea is
generic:
In invocationHasReachedAServer, instead of storing null as the value, you store the
target (key is the tx).
In invoke, if there's a tx, you add that target (if there is one) to the invocation
as transient metadata
The LB policy gets passed the invocation as an arg when it chooses a target (this is
already in the API)
LB policy checks for the metadata. If there, and that target is in its target list,
return that target
--
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
5 a.m.
New subject: [JBoss JIRA] Commented: (JBAS-4455) LoadBalancePolicy that tries to pin all requests associated with a tx to one server
[ http://jira.jboss.com/jira/browse/JBAS-4455?page=comments#action_12391930 ]
Galder Zamarreno commented on JBAS-4455:
----------------------------------------
While porting this feature, I have come across JBAS-5074.
This needs to be fixed first, which will involve slight changes
to the tests in trunk, which will them be ported to 4.2.x.
LoadBalancePolicy that tries to pin all requests associated with a tx
to one server
-----------------------------------------------------------------------------------
Key: JBAS-4455
URL: http://jira.jboss.com/jira/browse/JBAS-4455
Project: JBoss Application Server
Issue Type: Feature Request
Security Level: Public(Everyone can see)
Components: Clustering
Reporter: Brian Stansberry
Assigned To: Galder Zamarreno
Priority: Minor
Fix For: JBossAS-5.0.0.Beta3
The HA proxies don't allow failover once a tx has reached a server. This can lead to
this kind of situation:
EJB A deployed on nodes 1 and 2.
EJB B deployed on nodes 1 and 2.
1) Start tx, invoke on A.1.
2) Lookup B.
3) Invoke on B. LB policy picks B.2
4) Node 2 is dead for some reason so call fails.
5) Can't fail over because tx context won't allow a failover after a call has
reached a server.
But, B.1 is fine and is running on the only node the tx has invoked on. :(
Approach to improving this. This is based on JRMPInovkerProxyHA methods, but the idea is
generic:
In invocationHasReachedAServer, instead of storing null as the value, you store the
target (key is the tx).
In invoke, if there's a tx, you add that target (if there is one) to the invocation
as transient metadata
The LB policy gets passed the invocation as an arg when it chooses a target (this is
already in the API)
LB policy checks for the metadata. If there, and that target is in its target list,
return that target
--
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
11:18 a.m.
New subject: [JBoss JIRA] Updated: (JBAS-4455) LoadBalancePolicy that tries to pin all requests associated with a tx to one server
[ http://jira.jboss.com/jira/browse/JBAS-4455?page=all ]
Dimitris Andreadis updated JBAS-4455:
-------------------------------------
Fix Version/s: JBossAS-5.0.0.CR1
(was: JBossAS-5.0.0.Beta3)
Moving to CR1. If there is a problem with this, please contact me.
LoadBalancePolicy that tries to pin all requests associated with a tx
to one server
-----------------------------------------------------------------------------------
Key: JBAS-4455
URL: http://jira.jboss.com/jira/browse/JBAS-4455
Project: JBoss Application Server
Issue Type: Feature Request
Security Level: Public(Everyone can see)
Components: Clustering
Reporter: Brian Stansberry
Assigned To: Galder Zamarreno
Priority: Minor
Fix For: JBossAS-5.0.0.CR1
The HA proxies don't allow failover once a tx has reached a server. This can lead to
this kind of situation:
EJB A deployed on nodes 1 and 2.
EJB B deployed on nodes 1 and 2.
1) Start tx, invoke on A.1.
2) Lookup B.
3) Invoke on B. LB policy picks B.2
4) Node 2 is dead for some reason so call fails.
5) Can't fail over because tx context won't allow a failover after a call has
reached a server.
But, B.1 is fine and is running on the only node the tx has invoked on. :(
Approach to improving this. This is based on JRMPInovkerProxyHA methods, but the idea is
generic:
In invocationHasReachedAServer, instead of storing null as the value, you store the
target (key is the tx).
In invoke, if there's a tx, you add that target (if there is one) to the invocation
as transient metadata
The LB policy gets passed the invocation as an arg when it chooses a target (this is
already in the API)
LB policy checks for the metadata. If there, and that target is in its target list,
return that target
--
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
4:33 p.m.
New subject: [JBoss JIRA] Commented: (JBAS-4455) LoadBalancePolicy that tries to pin all requests associated with a tx to one server
[ http://jira.jboss.com/jira/browse/JBAS-4455?page=comments#action_12394855 ]
Galder Zamarreno commented on JBAS-4455:
----------------------------------------
Feature has been implemented in AS 4.2 community
branch. Future inclusion in EAP 4.x needs further discussion.
A wiki has also been created explaining the problems this feature
tries to resolve, how to use these new load balance policies
and some examples:
http://wiki.jboss.org/wiki/Wiki.jsp?page=TransactionStickyLoadBalancePoli...
LoadBalancePolicy that tries to pin all requests associated with a tx
to one server
-----------------------------------------------------------------------------------
Key: JBAS-4455
URL: http://jira.jboss.com/jira/browse/JBAS-4455
Project: JBoss Application Server
Issue Type: Feature Request
Security Level: Public(Everyone can see)
Components: Clustering
Reporter: Brian Stansberry
Assigned To: Galder Zamarreno
Priority: Minor
Fix For: JBossAS-5.0.0.Beta4
The HA proxies don't allow failover once a tx has reached a server. This can lead to
this kind of situation:
EJB A deployed on nodes 1 and 2.
EJB B deployed on nodes 1 and 2.
1) Start tx, invoke on A.1.
2) Lookup B.
3) Invoke on B. LB policy picks B.2
4) Node 2 is dead for some reason so call fails.
5) Can't fail over because tx context won't allow a failover after a call has
reached a server.
But, B.1 is fine and is running on the only node the tx has invoked on. :(
Approach to improving this. This is based on JRMPInovkerProxyHA methods, but the idea is
generic:
In invocationHasReachedAServer, instead of storing null as the value, you store the
target (key is the tx).
In invoke, if there's a tx, you add that target (if there is one) to the invocation
as transient metadata
The LB policy gets passed the invocation as an arg when it chooses a target (this is
already in the API)
LB policy checks for the metadata. If there, and that target is in its target list,
return that target
--
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
4:35 p.m.
New subject: [JBoss JIRA] Resolved: (JBAS-4455) LoadBalancePolicy that tries to pin all requests associated with a tx to one server
[ http://jira.jboss.com/jira/browse/JBAS-4455?page=all ]
Galder Zamarreno resolved JBAS-4455.
------------------------------------
Fix Version/s: JBossAS-4.2.3.GA
Resolution: Done
LoadBalancePolicy that tries to pin all requests associated with a tx
to one server
-----------------------------------------------------------------------------------
Key: JBAS-4455
URL: http://jira.jboss.com/jira/browse/JBAS-4455
Project: JBoss Application Server
Issue Type: Feature Request
Security Level: Public(Everyone can see)
Components: Clustering
Reporter: Brian Stansberry
Assigned To: Galder Zamarreno
Priority: Minor
Fix For: JBossAS-5.0.0.Beta4, JBossAS-4.2.3.GA
The HA proxies don't allow failover once a tx has reached a server. This can lead to
this kind of situation:
EJB A deployed on nodes 1 and 2.
EJB B deployed on nodes 1 and 2.
1) Start tx, invoke on A.1.
2) Lookup B.
3) Invoke on B. LB policy picks B.2
4) Node 2 is dead for some reason so call fails.
5) Can't fail over because tx context won't allow a failover after a call has
reached a server.
But, B.1 is fine and is running on the only node the tx has invoked on. :(
Approach to improving this. This is based on JRMPInovkerProxyHA methods, but the idea is
generic:
In invocationHasReachedAServer, instead of storing null as the value, you store the
target (key is the tx).
In invoke, if there's a tx, you add that target (if there is one) to the invocation
as transient metadata
The LB policy gets passed the invocation as an arg when it chooses a target (this is
already in the API)
LB policy checks for the metadata. If there, and that target is in its target list,
return that target
--
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
6110
days inactive
6337
days old
12 comments
3 participants
participants (3)
-
Brian Stansberry (JIRA) -
Dimitris Andreadis (JIRA)
-
Galder Zamarreno (JIRA)