I have been reading and fiddling with this for several days, now, and I think I have tried everything suggested on the lists, but I am still unable to get session replication working.
Issue: Both running clusterjsp and one of my own apps, if I shutdown all but one instance, session information is persisted and consistent with cookie versions properly updating and all. When I have more nodes on the cluster running, I get the following messages, but only the session persistence messages in one of the three instances at any time:
:
1: MemberId: server, MemberType: SPECTATOR, Address: urn:jxta:uuid-488C5EA5BD9449A58AADB23E2FDBDF002DE658F932AB436995B78E0CB3E080DA03
2: MemberId: JWS2, MemberType: CORE, Address: urn:jxta:uuid-488C5EA5BD9449A58AADB23E2FDBDF00673301B13D144E5B9038D74ED37F44D003
3: MemberId: JWS3, MemberType: CORE, Address: urn:jxta:uuid-488C5EA5BD9449A58AADB23E2FDBDF0070B24CA6C3CA406C9341F7FFA3B4C99903
4: MemberId: JWS4, MemberType: CORE, Address: urn:jxta:uuid-488C5EA5BD9449A58AADB23E2FDBDF00F8E1AA3BEF4C4D61AB35BD2DACDE511A03
|#]
Clustering appears to be working as 228.8.7.9 with my DAS and all cluster members showing up.
VLAN VLAN144-Tools (144) Snooping=Enabled
Group Sender Age
228.8.7.9 172.18.144.178 0
228.8.7.9 172.18.144.179 2
228.8.7.9 172.18.144.180 24
228.8.7.9 172.18.144.181 9
4 Incoming multicast streams
I'm not sure where to go to even see a problem in the round-robin memory replication that is supposed to be happening with the session data.
Applications are marked with <distributable />
Applications were deployed with Availability enabled.
I am not sure about how the Availability service should be configured in the cluster configuration and whether or not it should be configured in the DAS stand-alone configuration.
DAS configuration is default.
Cluster is as follows:
::Availability Service::
Availability Service: Enabled
MQ Store Pool Name: <blank>
HA Store Name: <blank>
HA Agent Hosts: <blank>
HA Agent Port: <blank>
HA Agent Password: <blank>
Auto Manage Store: Enabled
Store Health Check: Enabled
HA Store health Check: 5
No additional properties
::Web Container Availability::
Availability Service: Enabled
Persistence Type: Replicated
Persistence Frequency: web-method
Persistence Scope: modified-session
Single-Sign-On State: Enabled
HTTP Session Store: jdbc/hastore (default, but it doesn't exist in my jndi dir)
Persistence Store Health Check: Enabled
::EJB Container Availability::
Availability Service: Enabled
HA Persistence Type: replicated
SFSB Persistence Type: replicated
SFSB Store Pool Name: jdbc/hastore
::JMS Availability:: disabled
Hardware: 4 X4100s running Solaris 10
Deployment: DAS on JWS5, NA and single instance on JWS2, JWS3, JWS4
LB Technique: Apache 2.2 with mod_jk for the moment (hosted on JWS2)
Profile: cluster
JDK: 1.6 update 10
Glassfish: v2ur2
Installation followed the HA guide and
http://blogs.sun.com/technical/entry/2_machine_cluster_setup_cli version with the note on node-agent only install to make sure there is only one DAS instance. I can provide those steps if needed.
I'm not seeing any messages that look like errors even though I have turned all logging levels I can find up to high. Any help would be greatly appreciated.
[Message sent by forum member 'jrobey' (jrobey)]
http://forums.java.net/jive/thread.jspa?messageID=318732