Re: [xmlblaster] Clarification on Master/Slave clustering when Master is down...

[Marcel Ruff <mr at marcelruff.info>]

David R Robison wrote:
> I just wanted to clarify a point on how the current version of 
> XMLBlaster works in the Master/Slave Clustering. I have three nodes, 
> A/B/C. A is the master for domain DATA, B is a slave of A, and C is a 
> slave of B. In normal operations, messages for the domain DATA are 
> sent to node A and then available at node C through node B. If, 
> however, node A is down and I start sending the messages for domain 
> DATA to node B, I understand that the messages will be queued for A 
> and sent to A (the master for the domain) when it returns back 
> on-line. While the messages are being sent to node B, will they be 
> available to node C as well or will they not be available until A 
> comes back online and the messages resync'ed with the master (node A)?
>
> David
I think what you are after is dirtyRead which is supported.
See
http://www.xmlblaster.org/xmlBlaster/doc/requirements/cluster.dirtyRead.html

regards
Marcel