Perhaps federation would work instead of clustering? On Fri, Mar 15, 2019 at 1:54 PM Justin Bertram <[email protected]> wrote:
> #5 is expected. If a broker goes down then all the messages on the broker > are no longer available. You'd need to configure HA to mitigate this. > > #6 may be expected depending on your configuration. Can you provide your > broker.xml? > > > Justin > > On Fri, Mar 15, 2019 at 12:49 PM artemisn00b <[email protected]> > wrote: > > > So, I was testing my clustered setup (I have 3 nodes in a symmetric > cluster > > and I want all messages to be delivered to all nodes which have > consumers), > > and I was testing out large messages with a slow consumer. The situation > I > > ran into was - > > > > 1. I send lot of messages fast, and large to broker 1. > > 2. My client is listening on broker 3, receiving the messages, but slow. > I > > added a sleep there. > > 3. Now, a lot of messages get pages on broker 1, and I ungracefully > > shutdown > > broker1. > > 4. My producer reconnects, and moves to broker2. > > 5. The client keeps receiving messages on broker3, but only the new ones. > > The ones that were pages on broker1, won't be received. > > 6. Now, even after I bring broker1 up live, the paged messages won't be > > delivered to the consumer. > > > > I had to tear down the client and make it connect to broker1, to receive > > all > > the paged messages. This is not something I can do in a production > > environment. > > > > Do you guys have any solution to this as well? Would federated address > > handle this in a different way? > > > > Thank you so much! > > > > > > > > -- > > Sent from: > > http://activemq.2283324.n4.nabble.com/ActiveMQ-User-f2341805.html > > > -- Clebert Suconic
