As Dan points out, currently geode does no support distributed transaction across the multiple nodes.
>From gemfireXD stand point, DAP is used for partitioned data. To make it under transaction, you need to define triggers to achieve this. You may need to contact pivotal for further comments regarding to gemfireXD. On Thu, May 5, 2016 at 10:58 AM, Dan Smith <[email protected]> wrote: > You're looking at the docs for gemfireXD, which is a sql database built on > top of gemfire. Check out the docs for geode here > http://geode.docs.pivotal.io/docs/ > > With geode functions (the equivalent of data aware procedures) > 1. FunctionService.onRegion will make sure all of your data is covered > exactly once. So with a replicated region, that means it's executed on a > single node. With a partitioned region, it will run on as many nodes as it > takes to cover all of the buckets. You can choose to execute the function > on only the primaries by setting optimizeForWrite to be true. Replicated > regions don't have primaries; it's a multi-master system. > 2. You can execute a function within a transaction - but be aware that > geode does not support distributed transactions so if updates need to span > more than one member you can't do that in a single transaction. > > -Dan > > On Thu, May 5, 2016 at 6:52 AM, Amit Shah <[email protected]> wrote: > >> I have a couple of queries on how DAP >> <http://gemfirexd.docs.pivotal.io/latest/developers_guide/topics/server-side/data-aware-procedures.html#data-aware-procedures> >> and SYS listeners >> <http://gemfirexd.docs.pivotal.io/latest/reference/system_procedures/add_listener.html> >> work. >> >> >> 1. Event Listeners - Can the event listener be configured to be >> executed *only* on the primary node in the case when the table is >> replicated on multiple nodes? >> 2. Data Aware Procedure - I wanted to understand the behavior of a >> data aware procedure code execution when >> 1. It is called ON a table that is replicated on multiple nodes? >> Would the procedure execute on all the replicated nodes or only on the >> primary node? If it executes on all nodes, is there a way to configure >> it >> to execute on only on the primary node? >> 2. I would like the DAP to join the transaction started by the >> jdbc client that invoked it. This could also lead to recursive >> invocation >> of the procedure to perform DML queries on multiple tables. Is that >> theoretically possible? >> >> >> I would be helpful if someone could share their insights on this. >> >> Thanks, >> Amit >> > >
