One of the problems with Postgres-XL 9.2 is that it assigns a global transaction identifier (GXID) for every transaction that is started in the cluster. Since its hard for the system to know if a transaction is read-only or not (remember, even SELECT queries can do write activities), Postgres-XL would always assign a GXID and send […]
PgBouncer is a lightweight connection pooler for PostgreSQL. PgBouncer 1.7 was announced on the 18th of December 2015. In this blog post we’ll talk about the major new improvements in PgBouncer.
In the previous blog post, I briefly explained how we got the performance numbers published in the pglogical announcement. In this blog post I’d like to discuss the performance limits of logical replication solutions in general, and also how they apply to pglogical. physical replication Firstly, let’s see how physical replication (built into PostgreSQL since […]
A few days ago we released pglogical, a fully open-source logical replication solution for PostgreSQL, that’ll hopefully get included into the PostgreSQL tree in a not-too-distant future. I’m not going to discuss about all the things enabled by logical replication – the pglogical release announcement presents a quite good overview, and Simon also briefly explained […]
In Postgres-XL, sequences are maintained at the Global Transaction Manager (GTM) to ensure that they are assigned non-conflicting values when they are incremented from multiple nodes. This adds significant overhead for a query doing thousands of INSERTs in a table with a serial column, incrementing sequence one at a time and making a network roundtrip […]
This another simple test in continuation from last time. We will start with the same lineitem table as in the previous example. We will measure the time it takes to load the same 7.2GB text file repeatedly until the table size grows to about 1TB. We create a baseline with a table that has no indexes built on […]
I have been looking at the new Block Range Indexes (BRIN) being developed for PostgreSQL 9.5. BRIN indexes are designed to provide similar benefits to partitioning, especially for large tables, just without the need to declare partitions. That sounds pretty good but let’s look in greater detail to see if it lives up to the hype. How large? Here’s […]
The profiling utility perf that ships with the Linux kernel is extremely useful for examining system-wide and multi-process behaviour – but it does a lot more than the CPU profiling it’s often used for. You’ve probably looked at perf top -az or perf top -u postgres output, but that’s only the littlest bit of what […]