Re: [Quickfixj-users] High availability and sequence numbers?

[Robert N. <rob...@gm...>]

What I mean is that when I failover to the other machine that the other machine has the correct nextTarget and nextSender sequence numbers.

So how do you persist these efficiently across machines when you want to support failover of your process to another machine?

For example I’ve seen cases before that use IMDG or JDBC Store but the later takes a hit with any out of the box implementation.

> On Apr 18, 2020, at 9:03 AM, Colin DuPlantis <co...@ma...> wrote:
> 
> QuickFIX/J Documentation: http://www.quickfixj.org/documentation/
> QuickFIX/J Support: http://www.quickfixj.org/support/
> 
> 
> What exactly do you mean "sequence numbers are available"? Do you mean making sure that the new master node uses the correct sequence number?
> 
> If so, when we fail over, before starting a session, we look back and figure out the last sequence number we processed. We set the session next sequence number to that value on the new node before starting the session.
> 
> On 4/18/20 5:18 AM, Robert Nicholson wrote:
>> QuickFIX/J Documentation: http://www.quickfixj.org/documentation/
>> QuickFIX/J Support: http://www.quickfixj.org/support/
>> 
>> 
>> So how are folks ensuring that sequence numbers are available when you have to fail over to another machine. JDBC doesn’t seem scalable as it’s a synchronous write when the sequence number is updated isn’t it? What asynchronous methods are folks using instead?
>> 
>> _______________________________________________
>> Quickfixj-users mailing list
>> Qui...@li...
>> https://lists.sourceforge.net/lists/listinfo/quickfixj-users
> 
> -- 
> Colin DuPlantis
> Chief Architect, Marketcetera
> Download, Run, Trade
> 888.868.4884
> https://www.marketcetera.com
> 
> 
> 
> _______________________________________________
> Quickfixj-users mailing list
> Qui...@li...
> https://lists.sourceforge.net/lists/listinfo/quickfixj-users