Hi Tim,
For NFS, I will create a cluster for that later, or I will use another
alternative is add a new LUN on ESXi and create a logical NFS share for all
three brokers to mount for KahaDB.
Would you mind check out my activemq.xml file and help me point out possible
error(s) I have?
Thank you,
So... You have three ActiveMQ nodes that are capable of failing over in the
case of a problem with the current master broker. That's great. But your
NFS server is a single point of failure, and if it goes down, all three
ActiveMQ servers are useless. You might want to figure out how to make your
NF
Hi Tim,
I think I got it works. Here is my note, in case any newbie like me have
some idea where to start from :-)
Here is what I did for 3 ActiveMQ servers, and 1 NFS file server:
- Set up an NFS share server on windows 2018, and create a shared folder
name "activemqdata". Instruction on how to
Phu,
You can use any shared filesystem for storing the data, but to act as the
shared lock, the filesystem must support specific lock operations. NFSv4
supports those options. NFSv3 and before does not. Other filesystems may or
may not, though your experiment makes it sound like ESXi may not.
How
Hi Tim,
Thank you for your reply.
Yes, I have read this document before, and master/slave concept is clear to
me. My big question is on NFS shared file system where ActiveMQ is writing
data on.
I have only one option for this setup is to use NFS shared file system for
the data directory. I don't k
Hi François,
Yes, shared NFS file store will be a place for KahaBD directory between the
brokers. According to ActiveMQ, it supports NFS version 4.x; which implies
that I need to configure an additional server running NFS so the three VM
brokers can share a data directory over network.
If I ch
Have you read the official documentation at
https://activemq.apache.org/shared-file-system-master-slave? Does it
answer your questions?
If not, if you post specific questions about specific things you don't
understand to this list, hopefully we can help you work through anything
that's not clear.
Hi,
Are you using NFS to share the KahaBD directory between the brokers?
regards,
François
fpa...@apache.org
Le 08/08/2019 à 09:02, Arshkr a écrit :
> Hello,
>
> I'm a newbie to ActiveMQ; I try to configure 2 or 3 VM ActiveMQ nodes
> working as Master/slave using NFSv4. I'm reading lots of docu
Thanks, much clear now.
For the benefit to other users who may be attempting to do the same.
Following files will need to be amended
\server\default\deploy\reform-jms-ds.xml
\server\default\deploy\activemq-ra.rar\broker-config.xml
\server\default\deploy\activemq-ra.rar\META-INF\ra.xml
http://rab
Activemq doesn't do "shared store"
The first machine to attach will lock the db and the second will wait in
failover mode.
On Jun 20, 2013, at 12:50, deepak_a wrote:
> Thanks!
> I can confirm that my brokers are able to Discover each other.
>
>
> But from your earlier statements
> 'you'll won
Thanks!
I can confirm that my brokers are able to Discover each other.
But from your earlier statements
'you'll won't be able to share databases'
'they would all have their own persistence stores'
I see no reason why this won't work - so is this something not-recommended
by activeMQ? becuase of
is for client connections, is for
peer broker connections
On Thu, Jun 20, 2013 at 10:58 AM, deepak_a wrote:
> Thanks,
>
> I am getting the following exception when I start Jboss (in node 1)
>
> 2013-06-20 15:23:21,775 ERROR [org.apache.activemq.broker.BrokerService]
> (Starting ActiveMQ Broke
I so see the artefact static in the following location:
activemq-ra.rar\activemq-core-5.6.0.jar\*META-INF\services\org\apache\activemq\transport\/discoveryagent/\*
which is different from what I see in the exception
Caused by: java.io.IOException: Could not find factory class for resource:
META-
Thanks,
I am getting the following exception when I start Jboss (in node 1)
2013-06-20 15:23:21,775 ERROR [org.apache.activemq.broker.BrokerService]
(Starting ActiveMQ Broker) Failed to start ActiveMQ JMS Message Broker
(jboss-activemq-broker1, null). Reason: java.io.IOException: Transport
Connec
Yep, that's correct. The broker will know about each other via the
config
On Thu, Jun 20, 2013 at 10:19 AM, deepak_a wrote:
> Thanks for the reply.
>
> I am reading the explanation under link(s)
> http://activemq.apache.org/clustering.html
> and
> http://activemq.apache.org/static-transport-re
Thanks for the reply.
I am reading the explanation under link(s)
http://activemq.apache.org/clustering.html
and
http://activemq.apache.org/static-transport-reference.html
Am I right in understanding that static-transport-reference is to be set at
the Broker level? to enable one broker understand
For active deployments (all brokers active in your cluster), you'll won't
be able to share databases. You could try out network of brokers for
clustering where all brokers are active, connected to each other, etc, but
they would all have their own persistence stores.
On Thu, Jun 20, 2013 at 6:29
As per this link
http://activemq.apache.org/can-two-brokers-share-the-same-database.html
It appears the recommendation is not to share database.
This link recommends another strategy
http://activemq.apache.org/jdbc-master-slave.html
But this will work if I were to follow "Clustered Singleton Ser
On Wed, 2012-08-22 at 23:49 -0700, joesan wrote:
> This is how my MessageListener looks like:
>
> *public class Listener implements MessageListener {
>
> public void onMessage(Message message) {
> try {
> MapMessage map = (MapMessage)message;
>
This is how my MessageListener looks like:
*public class Listener implements MessageListener {
public void onMessage(Message message) {
try {
MapMessage map = (MapMessage)message;
Long msgCount = map.getLong("counter");
Below is how my Consumer looks like:
*public class Consumer {
private static String brokerURL =
"failover:(tcp://localhost:61616,tcp://localhost:61617)?maxReconnectAttempts=-1";
private static transient ConnectionFactory factory;
private transient Connection connection;
private tr
Well as per the docs, it is so!
http://activemq.apache.org/failover-transport-reference.html
maxReconnectAttempts=-1 will try indefinitely! I will try that tomorrow on
my cluster configuration and let know the results here!
Thanks for all the help!
--
View this message in context:
http://act
I would try increasing the maxReconnectAttempts to a bigger value. Is there a
rule of thumb as to how big this should be? Is there an indefinite try? what
happens if I specify -1? Will it try indefinitely until it gets a
connection? I would love to have that!
--
View this message in context:
ht
At least from what I'm seeing in *ActiveMQ in Action*, that looks right.
Could it be that your maxRecoveryAttempts=1 is preventing it from
recovering the second time? (I'm fairly new to this myself)
(*Chris*)
On Wed, Aug 22, 2012 at 10:32 AM, joesan wrote:
> I'm happy to see some help at las
I'm happy to see some help at last.
Here is how the fail-over url looks like from my clients (both producer and
consumer)
failover:(tcp://localhost:61616,tcp://localhost:61617)
--
View this message in context:
http://activemq.2283324.n4.nabble.com/ActiveMQ-Clustering-Issue-tp4655306p4655440.
What is the connection URL you are using on your clients?
(*Chris*)
On Wed, Aug 22, 2012 at 12:16 AM, joesan wrote:
> Any help friends?
>
>
>
> --
> View this message in context:
> http://activemq.2283324.n4.nabble.com/ActiveMQ-Clustering-Issue-tp4655306p4655410.html
> Sent from the ActiveMQ -
What is more troubling me is that the messages sent by the Producer to a
Topic is lost when the Master goes down. When I restart my consumer, all
those messages that the Producer had sent to the Topic was lost. How do I
handle this? Where is the high availability in picture in this AMQ
Clustering?
ed to restart my consumer app (tomcat
based app) to make it working. I am using activemq 5.2
Thanks,
Chaitanya
-Original Message-
From: Matt Pavlovich [mailto:mattr...@gmail.com]
Sent: Wednesday, January 25, 2012 9:15 PM
To: users@activemq.apache.org
Subject: Re: Activemq Clustering and fai
I'm guessing you still have client-side code problems, which is very
common. I suggest using the out-of-the-box test harnesses available
from ActiveMQ to help eliminate the broker configuration. Try running
100, 1,000 and then ramp up to 100,000 messages through your Master
broker. If your c
Hi Matt
First, thank you.
Second , unfortunely it does not work.
I did both changes in the configuration you had suggested, but it does not
work.
In the first 2 attempts, I was really excited, because it had worked . But
when I did another attempt, it did not.
To give you a scenario : I 'm runnin
Abimael-
A couple of things:
You don't use a networkConnector for master/slave. This is definitely
causing a problem. Network Connectors are to connect to *other*
master/slave pairs to provide scaling of traffic load. Remember:
Master/Slave *only* provides high availability for the messag
Hi Matt
I shared both (master and slave ) there :
http://www.4shared.com/folder/uSPBPhA4/activemq.html
The message that appears on console for the Producer console when the
master shut down :
.7585 [ActiveMQ Transport: tcp://sim1648/127.0.1.1:61616] WARN
org.apache.activemq.transport.failov
Abimael-
It might be helpful to understand that the client-side fail-over is
handled by the ActiveMQ client-side library. The end-user just needs to
code the consumer program as if it was talking to one broker.
I suggest adding DEBUG, or TRACE level logging on the client-side. You
may find
Oh no! No it does not work.. I did another test to check and it does not
work even with Durable SUbscriber .
Scenario :
2 brokers >: master and slave
1 producer with failover uri on the code to connect and provide messages;
1 consumer with failover uri on the code to connect and consume messa
I just forgot to mention that it works properly IF you use durable topic
subscriber .
If you use "dynamic topics" , created by code , I noticed that it does not
work (as I had explained in my previous comments and I assume is very
similar your scenario ) .
Regards
--
View this message in context
I had the same issue and unfortunely it seems that it is back .
My test applications are not equals of you, as I am doing some tests about
failover and H.A. using 2 simple applications.
But, the bottom line is : when master goes down, you (and me) want that the
slave start working properly AND ,b
9:15 PM
To: users@activemq.apache.org
Subject: Re: Activemq Clustering and failover
Chaitanya-
There are several solutions for this, and I suggest starting with just
two brokers connected via network connectors. On the client side, you
configure the clients to point to both brokers and they will
Chaitanya-
There are several solutions for this, and I suggest starting with just
two brokers connected via network connectors. On the client side, you
configure the clients to point to both brokers and they will
automatically fail-over to the "live" broker if you are doing maintenance.
Ser
2008/9/23 kumarchill <[EMAIL PROTECTED]>:
>
> Hey,
> I'm new to ActiveMQ.
>
> I have a situation where there are many publishers publishing messages and
> many subscribers that subscribes these messages.
> Assume
> ->For every subscriber there is one ActiveMQ instance that it listens.
> ->ActiveMQ
Also, I want to recieve messages only from one of the instance in the
cluster.
kumarchill wrote:
>
> Hey,
> I'm new to ActiveMQ.
>
> I have a situation where there are many publishers publishing messages and
> many subscribers that subscribes these messages.
> Assume
> ->For every subscriber
40 matches
Mail list logo
