Quantcast
Channel: VMware Communities : All Content - All Communities
Viewing all articles
Browse latest Browse all 179681

SQLFire as cache to Oracle database

$
0
0

Hi,

 

In case if I wanted to utilize SQLFire as a cache to backend oracle database with all tables in SQLFire defined as "overflow to disk" and with redundancy of 1 and without disk persistence and AsynEventListener of each table is defined "with persistence to disk store"

 

1. What will happen if the primary node of a table dies, the secondary node will it be primary now ?, will SQLFire creates another redundant copy in another node automatically ?

 

2. What will when both primary and secondary nodes dies, when both are re-activated, will it load the data again? if not, how to do hot load from oracle to SQLFire for partitioned tables in case of node failures in SQLFire?

 

3. If SQLFire is in vSphere cloud and back-end oracle database is in enterprise, what is the best hardware connectivity required between them for e.g. will a dedicated VPN connectivity of 1 Gigabit with High Availability between cloud and enterprise is sufficient or do I need to provide some more hardware

 

4. Suppose a batch program adds data to Oracle database, what is the best way to push that data to SQLFire, can the same batch program push to SQLFire in cloud ? or do we need a MOM like RabbitMQ and have a separate batch program in cloud to push data to SQLFire reading from MOM, any examples does VMWare has and can it be shared to community ?


5. I know Row Loaders are invoked only for select statement with where condition for primary key. Not all queries will involve primary keys, in such a scenario hot loading all the data to SQLFire becomes mandatory, so if we don't have disk persistence and only overflow to disk, as data grows in each node, will least used consistently be overflowed to disk  and read from disk in case if it is queried ?


6. Do we have any good practices for finding entity groups for collocation of partitioned tables. How do we design tables (am not asking for DDL syntax but ways to find entity groups), any good practices for applying parent-child relationships, do we need to design using Fact-Dimension model ?

 

Appreciate if someone can provide answers to above questions

 

Cheers,

Mani


Viewing all articles
Browse latest Browse all 179681

Trending Articles



<script src="https://jsc.adskeeper.com/r/s/rssing.com.1596347.js" async> </script>