Big Data Technology route choice for small and medium-sized enterprises (ii)-CASSANDRA+PRESTO programme
I have written before: small and medium-sized enterprise's big data technology route choice and low-key, luxurious, has the connotation agile Big Data solution: Flume+cassandra+presto+spagobi.
It took the last two months to finally pass the Cassandra+presto+spagobi program validation. The JDBC Driver, Prestogres Gateway, Shib Three methods of Presto are verified.
One, Presto JDBC Drive scheme
Presto the JDBC driver scheme, Java uses the client, such as Dbeaver, to execute SQL statements using Presto JDBC driver access. One of the issues to note is that you need to add the "User.timezone=shanghai/asia" parameter.
When you write your program to invoke the Presto JDBC driver, be aware that the current version is a standalone version and no longer requires invoking the reference jar package. Also, you need to set the "User.timezone=shanghai/asia": I have been struggling with this problem for a long time.
Also, be aware that SpagoBI cannot use the Presto JDBC driver and can only use the Prestogres gateway scenario.
Second, Prestogres Gateway scheme
When using the Prestogres gateway scenario, be aware of protocolversion=2. This step, I also groped for a long time to find the connection has been unable to test the success of the reasons. After the SpagoBI test is successful, the PostgreSQL client can also access the Presto.
Iii. Shib Programme
Shib Solution installation configuration is simple and easy to succeed. However, after using Shib, it is found that the interface is rough, the function is too simple and difficult to use in the actual production.
Iv. Summary
In the above scenarios, the Prestogres gateway scheme can use PostgreSQL client and JDBC driver to support BI tools such as SpagoBI. However, this approach has a complex architecture and may have performance bottlenecks that do not give you the benefit of this scenario. Shib now has a small amount of code, and members with report development experience in the team can get started. If time and manpower allow, it is also possible to develop a lightweight bigdata bi. I recommend using the shib+ custom development model.
I am already ready to use in the actual project, I hope interested friends can study together.
Big Data Technology route choice for small and medium-sized enterprises (ii)-CASSANDRA+PRESTO programme