[jboss-jira] [JBoss JIRA] Commented: (JBPORTAL-1459) portal:service=TreeCache,type=hibernate should define ClusterConfig
[
https://jira.jboss.org/jira/browse/JBPORTAL-1459?page=com.atlassian.jira....
]
Peter Hufnagel commented on JBPORTAL-1459:
------------------------------------------
Actually, I would think that you would use ${jboss.partition.udpGroup:224.0.0.36} (same
multicast as is passed on the cmd line for jbossAS), so you are only using one multicast
for both jboss & portal (and different ports should prevent the multicast receive
issues). However, yes, I'm in the process of spinning up several more portal
instances on our environment, and this is *really* annoying - and in fact this was
reported a year ago, this was a problem in portal 2.2.1 (had to fix this outside the
"clustering" notes to get multiple 2.2.1 instances running, we skipped 2.4 but
its a problem there too looking at the code...).
Yet another instance of short-sighted "not thinking about anyone running more than
one instance on a box" thinking - especially ugly in linux since even on seperate
bind IP's for each JVM, all your instances will get cache broadcasts from every other
one...).
portal:service=TreeCache,type=hibernate should define ClusterConfig
-------------------------------------------------------------------
Key: JBPORTAL-1459
URL: https://jira.jboss.org/jira/browse/JBPORTAL-1459
Project: JBoss Portal
Issue Type: Task
Security Level: Public(Everyone can see)
Components: Portal Core
Affects Versions: 2.4.1 Final, 2.4.1 SP1, 2.6.CR2, 2.6.CR1
Reporter: Galder Zamarreno
Assignee: Julien Viet
portal:service=TreeCache,type=hibernate omits ClusterConfig which can lead to confusion
when
trying to separate different AS + Portal clusters.
Here's, bearing in mind the status of the JBC/HB integration, a good starting point:
<attribute name="ClusterConfig">
<config>
<UDP mcast_addr="${portal.hibernate.cache.udpGroup:224.0.0.36}"
mcast_port="${portal.hibernate.cache.udpPort:55566}"
ip_ttl="8" ip_mcast="true"
mcast_send_buf_size="150000" mcast_recv_buf_size="80000"
ucast_send_buf_size="150000" ucast_recv_buf_size="80000"
loopback="false"
up_thread="false" down_thread="false" />
<PING timeout="2000" num_initial_members="3"
up_thread="false" down_thread="false" />
<MERGE2 min_interval="10000" max_interval="20000"
up_thread="false" down_thread="false" />
<FD_SOCK up_thread="false" down_thread="false" />
<FD shun="true" timeout="20000" max_tries="5"
up_thread="false" down_thread="false" />
<VERIFY_SUSPECT timeout="1500" up_thread="false"
down_thread="false"/>
<pbcast.NAKACK gc_lag="50"
retransmit_timeout="600,1200,2400,4800" max_xmit_size="8192"
discard_delivered_msgs="true" up_thread="false"
down_thread="false"/>
<UNICAST timeout="600,1200,2400" window_size="100"
min_threshold="10" up_thread="false"
down_thread="false"/>
<pbcast.STABLE desired_avg_gossip="20000" up_thread="false"
down_thread="false"/>
<pbcast.GMS join_timeout="5000" join_retry_timeout="2000"
shun="true" print_local_addr="true" up_thread="false"
down_thread="false"/>
<FRAG2 frag_size="8192" up_thread="false"
down_thread="false"/>
<pbcast.STATE_TRANSFER up_thread="false"
down_thread="false"/>
</config>
</attribute>
It'd be interesting to change the ClusterName as well:
<attribute
name="ClusterName">${portal.hibernate.cache.partition.name:portal.hibernate}</attribute>
This configuration works and should be much better than the default one provided by
JBossCache.
--
This message is automatically generated by JIRA.
-
If you think it was sent incorrectly contact one of the administrators:
https://jira.jboss.org/jira/secure/Administrators.jspa
-
For more information on JIRA, see: http://www.atlassian.com/software/jira