#291 — February 6, 2019 |
Postgres Weekly |
▶ PostgreSQL Goes to Eleven — Joe Conway’s talk from FOSDEM not only highlights all the great new features included within Postgres 11 but digs into some of the history and looks forward to Postgres 12 too. 48 minutes. Joe Conway |
PostgreSQL at Scale: Database Schema Changes Without Downtime — Braintree Payments, a division of PayPal, heavily uses Postgres and due to offering payment processing services, zero downtime is a must. Here’s a deep dive into all the things involved in making schema updates without downtime. James Coleman |
Get a Machine Learning Job, Guaranteed — Take advantage of the first online machine learning course with 1:1 mentorship from machine learning experts and a job guarantee. Get a machine learning job or your money back. Springboard sponsor |
MailChimp's Mandrill Service Experiences Outage Due to Transaction ID Wraparound — This is something you really need to be aware of if using Postgres at scale and aren't fully in tune with your autovacuuming setup. Sentry experienced a similar outage in 2015 and did a full writeup. Hacker News |
Using Postgres with Passphrase-Protected SSL Keys Under Peter Eisentraut |
Implementing “AS OF”-queries in Postgres — Many databases such as Oracle provide “AS OF” queries (a.k.a. ‘time travel’). This blog explains how to mimic a similar thing in Postgres. Hans-Jürgen Schönig |
Now Available: Gartner's 'Definitive Guide for Database Platforms as a Service' — Get your copy of the report. EnterpriseDB sponsor |
How Long Would a 64 Bit Transaction-ID Last in Postgres? — Postgres's MVCC implementation means it can keep multiple versions of data at a time allowing for concurrent transactions to take place. This requires maintaining a set of active transaction IDs. But means TXID wrap-around can happen (see Mandrill story above), so what if Postgres used a 64-bit id instead of 32-bit? Spoiler: it's not likely to change though. Andreas 'ads' Scherbaum |
The Life of a SQL Query — What happens when you run a SQL statement? We follow a Postgres query transformation by transformation as a query is processed and results are returned. Eric Feng |
Using Avinash Vallarapu |
|