Teiid 7.x and cluster. Status

[vplanat]

Hi,

the recent 7.2 alpha2 announce mentions effort of the community into the direction of "replication of result sets and internal materialization in clustered mode". On the other side, the Client developper doc (7.1) mentions that JDBC drivers can load-balance request accross different node of the same cluster. Is their a small doc/article describing the current status on this topic ?

Thanks very much for this great job

Regards

[rareddy]

For Load-Balancing and Fail-Over see the Client Developer's guide in "Using Multiple Hosts" section. Please note that the documents are updated but with the Alpha releases the documentation links on the website are not updated. However, if you download the Teiid runtime, it includes the document in the PDF format those are  up to date with the software.

For the "materilization" replication, check out the "Caching Guide" as to what are internal materialization tables, how you can use to boost the performence of the system by caching certain tables.

Hope this helps.

Ramesh..

[vplanat]

Thanks Ramesh for this quick and accurate answer.

The Client's dev guide is enhanced in alpha-2 version obviously. This clarify much more my understanding for both load-balancing and post-failover behavior.

For data replication in case of fail over I understand that Materialization view has to be used with the constraint that a cache replication may introduce in the cluster.

By the way when is the final 7.2 version planed ? Things are progressing so quick on teiid ...

Regards

vplanat

[rareddy]

Vincent,

In a cluster, using materialization works like this. Say on Host-1, you issued a query on a table that uses a implicit materialized table. This operation will load (and cache) this materialization table and provide you with results. Now, on Host-2, you issue another query that uses the same materialization table, then instead of re-loading the contents of this table again, it will fetch the contents from Host-1 and cache them locally and provide you with the required results.

7.2 we are planning to release towards the end of this month. If you are planning to use 7.2, I would suggest you get the alpha2 release, then go through your usecase, if you see any issues or features you need open JIRA or discussions, so that we can try to get them fixed before the release.

Thanks

Ramesh..

[vplanat]

OK. That very clear. Yes we are going to use alpha-2 for our preliminary tests.

Regards

vplanat