PostgreSQL 8.4, the latest version of the high-end open source SQL database system, is now available. This new release has several cool features, including: * Recursive Queries & Common Table Expressions * Windowing Functions * Parallel database restore * Column-level privileges * New monitoring tools This release not only implements some unique new PostgreSQL features, it makes administration of the database easier in several ways. Download it and give Postgres 8.4 a try.

Great progress! Btw: does anyone use a reliable multi-master cluster solution for postgresql?

Great progress!

Btw: does anyone use a reliable multi-master cluster solution for postgresql?

Not really. Having tried sequoia and ha-jdbc but never felt comfortable with neither of them due to some problems found. Currently i am using the PITR way to warm-standby replicate my DB. Although a real cluster would be nice. Looking forward to see if anyone knows any other reliable, production ready multimaster replication; of course, open source ;-) Regards

How many are you planning to replace over prized oracle with postgress ........here is migration suite http://www.enterprisedb.com/solutions/business_case.do

How many are you planning to replace over prized oracle with postgress ........here is migration suite http://www.enterprisedb.com/solutions/business_case.do

And enterpriseDB is the commercial vendor - so ofcourse they had to sell it! :)

At least one advantage PostgreSQL has over 'Commercial Open Source' databases: Nobody can buy PostgreSQL.

I am not a database expert, so I would like to ask: How does Postgres stand in comparison to the market in terms of performance and features? Specilly incomparison to oracle and mysql.

I am not a database expert, so I would like to ask:

How does Postgres stand in comparison to the market in terms of performance and features? Specilly incomparison to oracle and mysql.

PostgreSQL is way more stable than MySQL. Also, PostgreSQL has good support for stored procedures and custom functions while SPs in MySQL are a joke. Query language in PostgreSQL is also much more rich. MySQL might be faster for simple operations, but PostgreSQL completely demolishes it in complex queries. And Oracle... Well, it's hard to find anything which has more features than Oracle :) The only major missing piece in PostgreSQL is synchronous replication...

The only major missing piece in PostgreSQL is synchronous replication...

Some other major pieces are missing, like true table partitioning and the ability to execute a single query in parallel using multiple cpus and core. Nevertheless, PG is truly an awesome database and the relatively low market share is something that continues to surprise me.

MySQL is great for simple read only and/or simple query transaction processing, but it chokes in high throughput environment when complex transactions are involved. We build a high transaction application on top of MySQL and had to really wonder why when it actually went in production and the throughput increased. We've previously worked with Postgres in a relatively complex application and it performed without a hickup. Any performance issues we ever found had to do with application CPU or IO bottlenecks, Postgres executed hundreds of queries per transaction in about .003 percent of the transaction time. This of course doesn't say much, since it's application specific. It also doesn't discount MySQL as a viable alternative, since it can be tuned for high scalability requirements and is used by some of the most high traffic sites out there. I due wonder if majority of mysql transactions in those environments are read only. For highly scalable read/write scenarios, I think relational databases in general (at least their current implementations) are not scalable without a huge hardware investment and even then you're at the mercy of being able to scale up. That's where partitioning/sharding and even different data models are used. You can't possibly scale a highly read/write database up and the only current way to really scale horizontally is through partitioning. I wish these features where more available through some database features, vs. having to roll your own each time. Also, in memory stores, like Terracota and GigaSpaces that provide data partitioning and replication functionality can be used in conjunction with the ACID data storage, which can be used for eventual consistency. Ilya

looking forward to test the new version out in the next prod environment I setup. Excellent that postgresql works well together with jopr.

I am really looking forward to trying out the common table expressions (CTEs, "WITH" clause), and windowing functions. These are features I rely on frequently in Oracle, and it's great to see PostgreSQL support for them. I am surprised to see that "WITH" is not explained in many SQL books. I use it every day, and find that it is a very important feature for many non-trivial queries. John Hurst Wellington, New Zealand