Hi Adi, Let me be the next to pile on and say this is a great idea. What you are describing is session consistency, which seems like a very promising load balancing model. We are currently working on the same thing for our SQL Router JDBC driver. (See http://forge.continuent.org/jira/browse/TSR-8)
To make SCNs work over failovers you will probably need to add something like an epoch number-at least that's how we do it on our replicator; there are no doubt other ways. Each transaction has an ID that looks like the following: <epoch #><SCN> When each master increments the epoch number each time it comes online, which gives you monotonically increasing numbers albeit with gaps between them each time a master starts replicating. Incidentally, another way to pass the SCN back would be to have a function similar to last_insert_id() that gives you the last SCN on a connection. This seems more flexible because it would allow drivers to get this information without having to make calls that break standard interfaces like JDBC, DBI, and ODBC. Cheers, Robert On 4/30/09 4:46 PM PDT, "MARK CALLAGHAN" <[email protected]> wrote: On Thu, Apr 30, 2009 at 4:06 PM, Adi Alurkar <[email protected]> wrote: > Greetings, > It was great to meet everybody at Developer day. I have been catching up > with all the docs on the protocol and discussing them with our CTO and in > the last brainstorm with him we realized that although the proposed protocol > offers great enhancements in terms of exposing shard id / QUERY / QUERY_RO > in the protocol itself, the following idea would make it even better. > To simplify and explain the problem consider the following architecture: > (also note that this would apply to sharded DBD scenario as well) At a functional level this is a great problem to solve. We frequently remind users to direct queries to slaves when possible. That is not a good use of my time or the users time. I want to automate this in software without a significant overhead and without putting (and then managing) an additional set of processes between clients and servers. -- Mark Callaghan [email protected] _______________________________________________ Mailing list: https://launchpad.net/~drizzle-discuss Post to : [email protected] Unsubscribe : https://launchpad.net/~drizzle-discuss More help : https://help.launchpad.net/ListHelp -- Robert Hodges, CTO, Continuent, Inc. Email: [email protected] Mobile: +1-510-501-3728 Skype: hodgesrm
_______________________________________________ Mailing list: https://launchpad.net/~drizzle-discuss Post to : [email protected] Unsubscribe : https://launchpad.net/~drizzle-discuss More help : https://help.launchpad.net/ListHelp
