If you want to use HA-JNDI autodiscovery, you must have a JBoss server as the
JNDI store since HA-JNDI is not a separate application. You can write your own
remote client to use autodiscovery or use the one provided by JBoss. See the
clustering doc for further information.
If you want remote
The Naming Service must run on a JBoss server as it's not an independent module.
A remote client can use the auto discovery feature. This is described in the
HA-JNDI section of the Clustering Guide. The server and client must use the
same multicast address and port for this to work (they do by
I think the preferred approach is to deploy the application (e.g., the ear) to
the farm directory of a node and then let JBoss distribute it throughout the
cluster. There are some open issues with the farming feature so you should
verify that it works properly for you in you choose this option.
I'm not sure what the problem is but you can't use "the same IP address and
same port number (8009)." If you're using a binding address at JBoss server
startup (-b parameter), the IP's will differ. If you're using JBoss service
bindings, the ports will differ.
View the original post :
http:/
Are you actually using the "default" server? If so, your cluster instances
need to be copies of the "all" server.
View the original post :
http://www.jboss.com/index.html?module=bb&op=viewtopic&p=3953426#3953426
Reply to the post :
http://www.jboss.com/index.html?module=bb&op=posting&mode=rep
Clustering is useful if you have a need for any of the following when running
your application on multiple servers: failover support, replication, load
balancing of artifacts other than sessions (e.g., EJB's), true singletons,
notifications, etc.
View the original post :
http://www.jboss.com/i
You need to differentiate your clusters. They're distinguished by name,
multicast address and multicast port in cluster-service.xml.
The HA-JNDI provider url is used to locate an HA-JNDI server. After that,
HA-JNDI lookups use round-robin load balancing throughout the cluster so your
requests
If you're using JNDI to perform lookups on multiple nodes in a cluster or
clusters from a remote client, you'll need to know how to access each node. If
your cluster is dynamic, you'll need to know how to detect nodes entering and
exiting the cluster if you want to access each one remotely.
As
How did you change the stack to use TCP? Disabling AutoDiscovery shouldn't
have the effect you note. I disabled AutoDiscovery and used a TCP stack for
ClusterPartition in both 4.0.2 and 4.0.4 (I don't have 4.0.3 installed) - the
servers started fine in both cases.
View the original post :
ht
mod_jk instructions -
http://wiki.jboss.org/wiki/Wiki.jsp?page=UsingMod_jk1.2WithJBoss
View the original post :
http://www.jboss.com/index.html?module=bb&op=viewtopic&p=3950120#3950120
Reply to the post :
http://www.jboss.com/index.html?module=bb&op=posting&mode=reply&p=3950120
_
I use the sample-bindings file from 5.0 and the host attribute has been removed
there. In either case, the bindings should work. Are you getting a "port in
use" error for 1100 and 1101?
View the original post :
http://www.jboss.com/index.html?module=bb&op=viewtopic&p=3944200#3944200
Reply to
Why do you have the host attributes defined? The following works fine for me.
Note that if you have multiple IP addresses on your machine, you can
alternatively bind your nodes at runtime using the run -b myIP parameter and
avoid having to use the bindings file.
1
there is no dependency on clustering in the "default" configuration. If you
started with this configuration, it seems likely that you added something which
does have a clustering dependency.
You note that you removed cluster-service.xml. This configuration file isn't
included in the default c
Use your server addresses. For example -
java.naming.provider.url=localhost:1100,localhost:1200
-or-
java.naming.provider.url=192.168.1.100:1100,192.168.1.101:1100
View the original post :
http://www.jboss.com/index.html?module=bb&op=viewtopic&p=3941863#3941863
Reply to the post :
http://ww
Statistics are generated by cache interceptors. See
org.jboss.cache.interceptors.*. Generally, each interceptor maintains its own
statistics (e.g., CacheLoaderInterceptor maintains cache loading statistics).
CacheMgmtInterceptor maintains statistics for the cache itself (e.g., hits,
misses).
The initial implementation of statistics isn't region based. I don't know if
there's a JIRA issue for this yet but it's certainly a feasible enhancement.
View the original post :
http://www.jboss.com/index.html?module=bb&op=viewtopic&p=3941548#3941548
Reply to the post :
http://www.jboss.com/
Code like the folloiwng should be sufficient to direct your JNDI lookups to a
specific server. Change the provide url as needed.
You can use HA-JNDI by using the HA-JNDI port in the url. By default, this is
port 1100. Note that HA-JNDI requests may not go to the specified server as
HA-JNDI u
The first step in debugging would be to confirm that JGroups works properly in
your environment. You can see
http://wiki.jboss.org/wiki/Wiki.jsp?page=BasicClusterTest for further details.
FWIW - the "usual" way to cluster JBoss nodes on a single machine is to copy
the "all" server directory an
You can use HA-JNDI instead of a local JNDI address if you're running in a
cluster. If you know which server to use for various lookups, you might also
try creating initial contexts for each server and use the one appropriate for
each lookup.
View the original post :
http://www.jboss.com/inde
Here's the doc for clustering.
http://docs.jboss.org/jbossas/clustering/JBossClustering7.pdf
The JBoss clustering wiki also has useful information.
http://wiki.jboss.org/wiki/Wiki.jsp?page=JBossHA
Not sure what your second question is. JBoss itself uses JGroups for cluster
node communication; t
HA-JNDI definitely works in a single machine cluster. I use it in this mode
frequently and all of the HA-JNDI tests in the test suite do so as well.
You can differentiate the HA-JNDI servers either through service-bindings or by
using multihome addresses. In either case, HA-JNDI should work pr
It's a timeout value for remote method calls. You can change it via the
ClusterPartition MBean in the JMX Console. See the mbean named
jboss:service=DefaultPartition.
I don't know if changing the value dynamically in the console will have any
effect. You may need to restart the service; I su
The three clusters are different logical groups. The Tomcat cluster grouping
is handling session replication. The DefaultPartition cluster handles core
clustering including HA-JNDI. Any TreeCache related clusters are handling
replicated TreeCache instances.
These clusters all use JGroups int
Clustering is independent of EJB's.
View the original post :
http://www.jboss.com/index.html?module=bb&op=viewtopic&p=3940124#3940124
Reply to the post :
http://www.jboss.com/index.html?module=bb&op=posting&mode=reply&p=3940124
---
Using Tom
I'm just noting that if you're running mod_jk and node1 on the same machine,
you need to direct your application requests to mod_jk, not to node1. So if
mod_jk is running on port 80 and node1 on port 8080, you need to direct your
application requests to port 80, not 8080.
View the original pos
Are you directing your application requests to the load balancer? If not, you
need to do this.
View the original post :
http://www.jboss.com/index.html?module=bb&op=viewtopic&p=3939533#3939533
Reply to the post :
http://www.jboss.com/index.html?module=bb&op=posting&mode=reply&p=3939533
"All" is a superset of "default", containing additional services including
clustering support.
The Farm directory will be in your "all" server, not the "default" server as
the latter doesn't support clustering.
View the original post :
http://www.jboss.com/index.html?module=bb&op=viewtopic&p=3
Have you tried without the jnp prefix and the trailing slash? I use the
following successfully on my local machine. You should be able to access
HA-JNDI services by specifying providers in this format.
java.naming.provider.url=localhost:1100,localhost:1200
View the original post :
http://www
This has been fixed in JBossCache 1.3.0 GA release. See JBCACHE-523 for
further details.
View the original post :
http://www.jboss.com/index.html?module=bb&op=viewtopic&p=3934072#3934072
Reply to the post :
http://www.jboss.com/index.html?module=bb&op=posting&mode=reply&p=3934072
--
This has been fixed in JBossCache 1.3.0 GA release. See JBCACHE-523 for
further details.
View the original post :
http://www.jboss.com/index.html?module=bb&op=viewtopic&p=3934071#3934071
Reply to the post :
http://www.jboss.com/index.html?module=bb&op=posting&mode=reply&p=3934071
--
A "multihomed" machine would have multiple network cards. If you have a laptop
with a wireless and ethernet card, then you should be able to run a cluster
using the two IP addresses for your nodes.
You can use this in the clustering test suite by modifying local.properties in
the testsuite fol
I've recently run the tests-clustering target for both 4.0.4 and 5.0 on Windows
XP Pro SP2, using a multihome machine. Most of the tests execute successfully.
View the original post :
http://www.jboss.com/index.html?module=bb&op=viewtopic&p=3932388#3932388
Reply to the post :
http://www.jbos
1) You shouldn't have to create a local.properties file in the testsuite
folder. It should already be there as it's under cvs control.
2) There was a problem late last week with running cluster tests in 4.0.4. It
was fixed over the weekend. You should refresh your 4.0.4 code if you haven't
Have you checked your local.properties file and configured the node0= and
node1= entries?
When you run the clustering tests (e.g., build tests-clustering), the suite
should add node0 and node1 to your server and then run tests using them.
View the original post :
http://www.jboss.com/index.htm
So your cache should be replicated and if the first server node goes down, the
application should still be able to retrieve cached entries from the cache
instance on the second node.
If the cache isn't replicated successfully, then it's likely that your nodes
aren't communicating properly. You
Where is your cache configuration? Did you specify CacheMode as REPL_SYNC or
REPL_ASYNC? This is necessary to indicate that you want a replicated cache.
View the original post :
http://www.jboss.com/index.html?module=bb&op=viewtopic&p=3930653#3930653
Reply to the post :
http://www.jboss.com/
HA-JNDI bindings are replicated so any HA-JNDI bindings on one node would be
copied to the other node. Global JNDI bindings would not be replicated as
they're only bound locally. In either case, using HA-JNDI to perform a lookup
should locate all local and ha bindings on nodes in the cluster.
Does your PROVIDER_URL property specify the HA-JNDI port? For example,
jnp://localhost:1100
View the original post :
http://www.jboss.com/index.html?module=bb&op=viewtopic&p=3928762#3928762
Reply to the post :
http://www.jboss.com/index.html?module=bb&op=posting&mode=reply&p=3928762
---
I don't have any information on your problem, just an observation about the
Global Namespace in the JNDIView mbean.
The Global Namespace is unrelated to HA-JNDI; it's a store for local JNDI
bindings. So if you bind something in HA-JNDI, it won't appear there. Of
course, HA-JNDI should locate
See section on HTTP Session Clustering in
http://docs.jboss.org/jbossas/clustering/JBossClustering7.pdf
See wiki pages on similar topic.
http://wiki.jboss.org/wiki/Wiki.jsp?page=TomcatClustering
View the original post :
http://www.jboss.com/index.html?module=bb&op=viewtopic&p=3923582#3923582
R
Seems like cache loader default persistence should be consistent, or at least
better documented so that it's more clear.
I can look at the documentation on this subject and clean it where appropriate
so that this behavior is better described.
View the original post :
http://www.jboss.com/index
I didn't have a "shared" setting; I guess the default value is false. I added
a shared=true setting to my configuration and the problem disappeared.
Thanks.
View the original post :
http://www.jboss.com/index.html?module=bb&op=viewtopic&p=3921481#3921481
Reply to the post :
http://www.jboss.
Using latest JBossCache HEAD code -
If I configure a replicated cache (two nodes) with a JDBC cache loader as
backing store, my transactions hang and never complete.
This problem only occurs if the cache instances share a common JDBC data
source; it doesn't occur if they use separate data sourc
If I use FileCacheLoader, cache entries are persisted and consequently
available on cache restart. If I use JDBCCacheLoader, its default behavior is
to drop the cache table during shutdown so that entries aren't available on
restart. Is this difference in behavior intentional?
View the origin
I've just tried using two clustered cache instances on a single server and it
seemed to run fine. I created the second cache configuration by copying the
first and then modifying the service name. The instances were registered
properly and replication worked as expected.
View the original pos
The fix seems to be fine. At least my own sample code having transactions in a
replicated cache now works. :)
re: the multiple instance issue - I recently modified the interceptor mbean
registration code so that the mbean will only be registered if not already
registered. So the server shoul
I'm trying to cache several nodes within a transaction in a replicated cache
environment. I'm using the latest JBossCache HEAD code along with a fairly
recent build of JBossAS HEAD. The transaction is being executed from a servlet
running on one of the cluster nodes.
If I use a non-replicated
If you require an app server independent solution, JBossCache can be used with
other app servers. HA-JNDI and DistributedState have dependencies on JBossAS.
The JBossAS Clustering manual contains a brief description of DistributedState.
See the index for its location.
http://docs.jboss.org/jb
If you want to use IP addresses, you should only need to modify the command you
use to start the node. The -b parameter is used for the IP address. For
example,
run.bat -c node1 -b 192.168.1.101
View the original post :
http://www.jboss.com/index.html?module=bb&op=viewtopic&p=3917114#391711
JBossCache running in replication mode is an example of a service that ''s
exposed via an mbean on each node.
If you want to use the singleton mbean approach, you can store your state in
one of the replicated JBoss services (JBossCache, HA-JNDI or DistributedState).
When you need to nominate a
As noted, you can store your data in a common database so that it's accessible
from each node.
You can also store your data in a replicated store in a JBoss cluster.
Possibilities include JBossCache (using replication), HAJNDI and
DistributedState.
View the original post :
http://www.jboss.c
I don't think it's currently feasible as HA services and distributed cache
instances are independent groups. Currently you need to distinguish these
groups via group name, bind address, and bind port. I think that future plans
for JGroups include supporting a common configuration; perhaps some
HAJNDI should work out of the box. in the default configuration for the "all"
server.
There's no cross-dependency between DefaultPartition and HAJNDI as the latter
depends on the former, not vice-versa.
You can try relocating HAJNDI to a different port to see if the problem
persists on the new
What are you trying to achieve with your cluster? Clustering will work without
a load balancer but you won't have server load balancing or session failover.
The nodes will still be in a cluster and some cluster functionality will work
(e.g., HA-JNDI, DistributedState, TreeCache replication etc
If you specify port 1100, your lookups will be directed to HA-JNDI. While this
will return local JNDI bindings, it won't be as efficient as if you use local
jndi lookups. Of course, local jndi lookups won't work if the target isn't on
the same cluster node.
View the original post :
http://ww
What version of JBoss are you using? Sourceforge issue 1037726 corresponds to
JBoss JIRA issue JBAS-1103. According to the JIRA issue, this problem was
fixed in JBoss 4.0.1.
A similar issue is reported and open per JBCLUSTER-73. It's possible that the
thisis a duplicate and the issue has bee
Do you have a reference to the reported problem or fix (e.g., a JBoss JIRA
issue)?
View the original post :
http://www.jboss.com/index.html?module=bb&op=viewtopic&p=3911555#3911555
Reply to the post :
http://www.jboss.com/index.html?module=bb&op=posting&mode=reply&p=3911555
-
Load balancing and failover should normally work if configured properly. Have
you reviewed the instructions provided on the wiki page at
http://wiki.jboss.org/wiki/Wiki.jsp?page=UsingMod_jk1.2WithJBoss
View the original post :
http://www.jboss.com/index.html?module=bb&op=viewtopic&p=3911495#39
If you want load balancing and/or failover, you'll need Apache mod_jk or some
other means (e.g., hardware) to accomplish this. The JBoss servers will run in
a cluster without this (i.e., they'll communicate and share clustered
artifacts) but you won't get load balancing or failover.
Jerry
Vie
1. If you're running copies of the "all" configuration on the same network,
the copies will be clustered (along with "all" if you're using it). The
software and configuration files included in the "all" configuration are
cluster enabled by default. You don't need to do anything to "enable" th
I can't answer your question about the discarded messages but I can advise
about the DefaultPartition.
DefaultPartition is the name of the primary cluster group in JBoss.
If you're using a clustered JBoss server (i.e., "all" configuration), JBoss
itself uses two cluster groups. One is named De
Retrieve the initial context (or any other context) and add your subcontext to
it.
Context ctx = new InitialContext();
Context newCtx = ctx.createSubcontext(newCtxName);
View the original post :
http://www.jboss.com/index.html?module=bb&op=viewtopic&p=3903781#3903781
Reply to the post :
http:
I think you can't configure HA-JNDI to direct all lookups to a single node.
I'm not familiar with 3.2.7 but in later code, HA-JNDI uses the smart proxy
with a round-robin policy to service HA-JNDI lookups. Perhaps I'm
misinterpreting something but that's my understanding of the current
implem
Are you directing your requests to the Apache server (i.e., port 80)?
Are the application requests being load balanced to the two servers or is
everything being forwarded to one of the servers? If you use the configuration
provided on the wiki page, you should have failover and load balancing e
Are you sure you've configured Apache correctly? If Apache is running on port
8080, how can one of your nodes be running on 8080 on the same machine? Are
you using multiple IP addresses?
View the original post :
http://www.jboss.com/index.html?module=bb&op=viewtopic&p=3899659#3899659
Reply t
Are you sure you've configured mod_jk correctly with JBoss? You can check the
wiki page for guidance at
http://wiki.jboss.org/wiki/Wiki.jsp?page=UsingMod_jk1.2WithJBoss
If you've configured everything correctly, your application will be accessible
on whatever port you've specified for Apache.
You'l need to run multiple copies of your server, using different node names
such as node1 and node2. Copy the 'all' server configuration to to create the
new servers. You can then start a server using "run.bat -c node1"
You'll also need to vary the port addresses of the second server if you'r
There are several ways to configure multiple nodes on a single Windows machine.
1) If you have multiple IP addresses (e.g., wired and wireless connections),
you can start each node using a separate IP address. For example -
run.bat -c node1 -b 168.1.68.100
2) You can use the service-bindings
You should be able to invoke list() or listBindings() to see what's stored in
HA-JNDI. These calls don't access the local JNDI tree.
One observation I have after working with HA_JNDI in JBoss 5 Alpha is that even
though you specify a provider url such as "jnp://nodeA:1100", the "smart stub"
is
"this has been posted more than a month back.still i didn't get a single reply.
this is very dissapointing. this is big problem with open source."
Why is this a problem with open source? You've posted to a free list that
developers review and respond to as time/circumstances permit. If you want
1. You can use a hardware load balancer for your http sessions. I don't know
if there are other software solutions besides apache/mod_jk.
2. The JBoss Clustering Guide recommends that you deploy beans to all nodes
for performance. If you only deploy to a single node, you can probably locate
If you have a conflict with another application, you can change the JBoss
configuration for a service's ports. In the case of port 1098, this port
binding is configured in the conf/jboss-service.xml file of your server.
Change the port number in the configuration file and restart the server.
Are you directing the requests to the port on which Apache is running? If node
1 is running on port 8080 and Apache is running on port 80, the requests need
to be directed to port 80 on the Apache server.
View the original post :
http://www.jboss.org/index.html?module=bb&op=viewtopic&p=3891804
I think your web requests need to be directed to the Apache server where you
have mod_jk configured. The requests will then be distributed per your
configuration.
e.g., http://:8080/xlWebApp
View the original post :
http://www.jboss.org/index.html?module=bb&op=viewtopic&p=3891795#3891795
Re
Not sure why you're having problems with the OIL & UIL2 ports but the bindings
are probably the ones defined in ..\deploy-hasingleton\jms\oil-service.xml and
..\deploy-hasingleton\jms\uil2-service.xml.
View the original post :
http://www.jboss.org/index.html?module=bb&op=viewtopic&p=3891714#389
Have you looked at JBoss Services Binding Management? It's described at
http://docs.jboss.com/jbossas/admindevel326/html/ch10.html#d0e20806.
This should allow you to specify all of the incremented ports in a single file.
View the original post :
http://www.jboss.org/index.html?module=bb&op=vie
There are several ways to run multiple nodes from a single machine.
1) As noted in an earlier reply, you can configure multiple IP addresses on a
machine and then start the JBoss server with an associated IP address (e.g.,
run.bat -c mynode1 -b 192.168.1.100)
2) You can use a single IP addres
77 matches
Mail list logo