Apache Artemis Cluster and High Availability questions

[cobba <fa...@yahoo.de>]

Hi,

i'm new to Apache ActiveMQ and Artemis. So far running a standalone server
is working great. Now i tried to setup an Artemis Master/Slave Cluster. Now
i have some general questions. I tried looking through the manual and the
mailing list, but i couldn't find the answers i'm looking for.

1. Is it usual, that all my server start with the same NodeID?

2. If i have configured a Master/Slave Cluster. Should i see both servers in
the Web-Console under Nodes?
(If i'm configuring a colocated cluster i can see all nodes there, but not
with master/slave)

3. Is it necessary to use the "start" operation on the slave?
(it always starts up with the status "started: false")

Here are some infos about my setup:

- 2 nodes - Apache Artemis 2.8.0 (also tried 2.7.0 and 2.6.4)
- each node in it's own virtual machine
- created the brokers with the following commands
-- ./artemis create --clustered --shared-store --data /mnt/nfs/share/artemis
/var/opt/apache-artemis-2.8.0/broker/server_sharedstorage_1
-- ./artemis create --clustered --shared-store --data /mnt/nfs/share/artemis
--slave /var/opt/apache-artemis-2.8.0/broker/server_sharedstorage_2
- the nodes are configured using the same shared nfs storage

Kind regards
Fabian



--
Sent from: http://activemq.2283324.n4.nabble.com/ActiveMQ-User-f2341805.html

Re: Apache Artemis Cluster and High Availability questions

[Justin Bertram <jb...@apache.org>]

> Is it usual, that all my server start with the same NodeID?

A master and its slave(s) will have the same node ID. However, only one
node with the same ID should ever be active at the same time. Otherwise
it's highly likely that your data integrity will be compromised (since
there will be two brokers running with the same journal data).

> If i have configured a Master/Slave Cluster. Should i see both servers in
the Web-Console under Nodes?

Please clarify where the "Nodes" indicator is. Looking at the web console
of the master broker of a master/slave pair I don't see anything labelled
as "Nodes."

> Is it necessary to use the "start" operation on the slave?

When you "start" a slave it runs in a special, in-between state while it
waits for its master to fail. Once it detects that its master has failed it
will automatically start the rest of the way so that clients can access it.


Justin

On Mon, May 13, 2019 at 5:57 AM cobba <fa...@yahoo.de> wrote:

> Hi,
>
> i'm new to Apache ActiveMQ and Artemis. So far running a standalone server
> is working great. Now i tried to setup an Artemis Master/Slave Cluster. Now
> i have some general questions. I tried looking through the manual and the
> mailing list, but i couldn't find the answers i'm looking for.
>
> 1. Is it usual, that all my server start with the same NodeID?
>
> 2. If i have configured a Master/Slave Cluster. Should i see both servers
> in
> the Web-Console under Nodes?
> (If i'm configuring a colocated cluster i can see all nodes there, but not
> with master/slave)
>
> 3. Is it necessary to use the "start" operation on the slave?
> (it always starts up with the status "started: false")
>
> Here are some infos about my setup:
>
> - 2 nodes - Apache Artemis 2.8.0 (also tried 2.7.0 and 2.6.4)
> - each node in it's own virtual machine
> - created the brokers with the following commands
> -- ./artemis create --clustered --shared-store --data
> /mnt/nfs/share/artemis
> /var/opt/apache-artemis-2.8.0/broker/server_sharedstorage_1
> -- ./artemis create --clustered --shared-store --data
> /mnt/nfs/share/artemis
> --slave /var/opt/apache-artemis-2.8.0/broker/server_sharedstorage_2
> - the nodes are configured using the same shared nfs storage
>
> Kind regards
> Fabian
>
>
>
> --
> Sent from:
> http://activemq.2283324.n4.nabble.com/ActiveMQ-User-f2341805.html
>