[prev in list] [next in list] [prev in thread] [next in thread] 

List:       activemq-users
Subject:    Re: Dual Mirroring and Core Bridges
From:       Stephen Baker <stephen.baker () rmssoftwareinc ! com>
Date:       2022-09-23 18:05:45
Message-ID: DS7PR01MB7855BCC60F7CD3B952202D3692519 () DS7PR01MB7855 ! prod ! exchangelabs ! com
[Download RAW message or body]


> I wouldn't mirror and send to the same broker!

That’s the problem I’m trying to solve.

I have disabled the core bridges on the mirror side, but that means there is an extra \
step in failover – stop the bridge on the old active side, start on the new active \
side. Dual mirroring was an easy sell to our IT department specifically because it \
meant they didn’t need to reconfigure artemis during failover.

In theory if I exclude the source queues for the bridge from the mirror, then the \
bridge on the mirror side can keep running but will never get any messages. When we \
failover that bridge is already running to take the messages that are now coming in \
to it.

I don’t know if you have thoughts on any simpler ways.

From: Clebert Suconic <clebert.suconic@gmail.com>
Date: Thursday, September 22, 2022 at 11:07 AM
To: users@activemq.apache.org <users@activemq.apache.org>
Subject: Re: Dual Mirroring and Core Bridges
I wouldn't mirror and send to the same broker!


a Mirror is a mirror.. .if you send to that mirror, the send will then
update the mirror back.. that's a bit of a mess! Make it simple!

On Mon, Sep 19, 2022 at 8:46 AM Stephen Baker
<stephen.baker@rmssoftwareinc.com> wrote:
> 
> Or would you suggest swapping the core bridges for additional broker connections
> 
> <broker-connections>
> <amqp-connection ... name=”B”>
> <sender queue-name=”fowardBfoo” />
> <receiver queue-name=”foo” />
> </amqp-connection>
> <amqp-connection ... name=”A_DC2” />
> <mirror address-filter=”!fowardBfoo” />
> </amqp-connection>
> </broker-connections>
> 
> 
> From: Stephen Baker <stephen.baker@rmssoftwareinc.com>
> Date: Monday, September 19, 2022 at 8:38 AM
> To: users@activemq.apache.org <users@activemq.apache.org>
> Subject: Re: Dual Mirroring and Core Bridges
> To isolate different sets of services they run on separate artemis instances, and \
> to avoid concern spilling into the application layer a limited number of queues \
> have core bridges to other artemis clusters in the same datacenter for cross \
> communication. 
> The dual mirroring on the other hand is for disaster recovery.
> 
> I’m proposing something like:
> 
> <mirror address-filter=”!forwardAfoo”/>
> ...
> <bridge ...>
> <queue-name>fowardAfoo</queue-name>
> <forwarding-address>foo</queue-name>
> 
> Should that be ok?
> 
> From: Clebert Suconic <clebert.suconic@gmail.com>
> Date: Friday, September 16, 2022 at 1:01 PM
> To: users@activemq.apache.org <users@activemq.apache.org>
> Subject: Re: Dual Mirroring and Core Bridges
> I wouldn't use both (bridge and mirroring between the servers) I would
> simplify your setup with either one or the other.
> 
> On Fri, Sep 16, 2022 at 8:19 AM Stephen Baker
> <stephen.baker@rmssoftwareinc.com> wrote:
> > 
> > We are running artemis multiple artemis clusters between hot and cold datacenters \
> > in a dual mirroring setup, so: 
> > A mirrors with A’
> > B mirrors with B’
> > 
> > We also have core bridges between A and B so:
> > 
> > A.forwardBfoo goes to B.foo
> > A’.forwardBfoo goes to B’.foo (by virtue of symmetric configuration)
> > B.forwardAfoo goes to A.foo
> > A’.forwardAfoo goes to A’.foo
> > 
> > My question, is does this potentially amplify/duplicate messages as the core \
> > bridge picks up the forward on both sides of the mirror? Does the answer depend \
> > on whether we’re running earlier/later than 2.24.0? 
> > If so is the recommended workaround to exclude the forward* queues in the mirror \
> > configuration? 
> > Thanks,
> > Stephen E. Baker
> 
> 
> 
> --
> Clebert Suconic
> [EXTERNAL]: This email originated from outside of Rave Mobile Safety. Do not click \
> links or open attachments unless you recognize the sender and know the content is \
> safe. [EXTERNAL]: This email originated from outside of Rave Mobile Safety. Do not \
> click links or open attachments unless you recognize the sender and know the \
> content is safe.



--
Clebert Suconic
[EXTERNAL]: This email originated from outside of Rave Mobile Safety. Do not click \
links or open attachments unless you recognize the sender and know the content is \
safe.



[prev in list] [next in list] [prev in thread] [next in thread]