Here is a small feature that has showed up during the [first commit fest] (https://commitfest.postgresql.org/action/commitfest_view?id=22) of Postgres 9.5 allowing to add complementary information in the process names displayed by a server:

The [jsonb datatype] (https://www.postgresql.org/docs/devel/static/datatype-json.html) newly introduced in Postgres 9.4 has a property that makes its manipulation rather facilitate when doing operations on it with embedded functions: it does not allow duplicated object keys. On the contrary, json allows that as you can see here:

Postgres 9.4 is adding a new in-core GiST operator class for the [inet and cidr] (https://www.postgresql.org/docs/9.4/static/datatype-net-types.html#DATATYPE-INET) datatypes. It has been introduced by this commit:

Working with logical replication and a cluster of nodes heavily linked among each other can result in particularly disturbing results. If a logical change is received on a node that itself regenerates additional changes, it is really easy to finish with an uncontrolable chain reaction.

These last couple of days I have been developing for studies a background worker able to fetch changes from a logical decoder already developed and presented on this blog called [decoder_raw] (https://github.com/michaelpq/pg_plugins/tree/master/decoder_raw) able to generate raw queries based on the logical changes decoded on server from its WAL. This receiver, called [receiver_raw] (https://github.com/michaelpq/pg_plugins/tree/master/receiver_raw) runs as a background worker that connects to a node running decoder_raw using a replication connection and then fetches decoded logical changes from it, which are in this case ready-to-be-applied raw queries. Those changes are then applied on the database one by one using the [SPI] (https://www.postgresql.org/docs/devel/static/spi.html) through a loop process that sleeps during a customizable amount of nap time. Note that receiver_raw is actually designed to be performant as it applies one single batch of changes using only one transaction in a single loop process. That’s the best doable particularly for long transactions.

When developing a logical change receiver with the new [logical decoding] (https://www.postgresql.org/docs/devel/static/logicaldecoding.html) facility of Postgres 9.4, there are a couple of new commands and a certain libpq protocol to be aware of before beginning any development for a logical replication receiver (on top of knowing some [basics] (/postgresql-2/postgres-9-4-feature-highlight-basics-logical-decoding/) and what is an [output decoder plugin] (/postgresql-2/postgres-9-4-feature-highlight-output-plugin-logical-replication/)). Here is an exhaustive list of the commands to know.

Today here is a highlight of a new Postgres 9.4 feature interesting for developers and companies doing packaging of Postgres on Windows as it makes possible the installation of client-only binaries and libraries using MSVC. It has been introduced by this commit:

Continuing on the series of posts about logical replication and after looking at [some basics] (/postgresql-2/postgres-9-4-feature-highlight-basics-logical-decoding/) and the concept of [REPLICA IDENTITY] (/postgresql-2/postgres-9-4-feature-highlight-replica-identity-logical-replication/) for a table, now is time to begin more serious things by having a look at how to develop an output plugin for the logical decoding facility.

Among the many things to say about logical replication features added in PostgreSQL 9.4, REPLICA IDENTITY is a new table-level parameter that can be used to control the information written to WAL to identify tuple data that is being deleted or updated (an update being a succession of an insert and a delete in MVCC).

The second huge feature coming in PostgreSQL 9.4 with [jsonb] (postgresql-2/postgres-9-4-feature-highlight-indexing-jsonb/) is called [logical decoding] (https://www.postgresql.org/docs/devel/static/logicaldecoding.html). In short, it is a new plugin facility that can be used to decode changes that happen on a database and stream them to external sources. It can be used for many things like replication, auditing or even online upgrade solutions.