CYBERTEC Blog
cybertec.at · Company blog · on HN since 2011
CYBERTEC is an Austrian PostgreSQL consultancy whose blog covers Postgres features, performance, parallel queries, monitoring, and tuning.
What HN saysHN commenters link it in PostgreSQL threads on SKIP LOCKED, parallel aggregates, running Postgres on ARM, and query planner cost settings.
27HN people shared it
7linked it in comments
27stories submitted
342story points
9 years agolast shared
2011people per year2026
Most-shared posts
- 1Beating Uber with a PostgreSQL prototype4 people · 16 points · first shared 2016 · discussion on HN
- 2Why favor PostgreSQL over MariaDB/MySQL?2 people · 230 points · first shared 2017 · 193 comments on HN
- 3Agg: Parallel aggregations for PostgreSQL2 people · 47 points · first shared 2015 · 11 comments on HN
- 4PostgreSQL 9.3: Add a slave with just one line2 people · 6 points · first shared 2013 · discussion on HN
- 5Writing your own aggregation functions in Postgres2 people · 3 points · first shared 2017 · discussion on HN
- 6Pgwatch2 PostgreSQL monitoring tool1 people · 7 points · first shared 2017 · discussion on HN
- 7Two simple Postgres tricks to kick-start year 20171 people · 4 points · first shared 2017 · discussion on HN
- 8Experimenting with scaling and full parallelism in PostgreSQL1 people · 3 points · first shared 2017 · discussion on HN
- 9PostgreSQL connection pooling intro – PgBouncer and pgpool-II1 people · 3 points · first shared 2016 · discussion on HN
- 10Walbouncer – a proxy for selective PostgreSQL replication1 people · 3 points · first shared 2016 · discussion on HN
- 11PostgreSQL: int4 vs. int8 vs. uuid vs. numeric performance on bigger joins1 people · 3 points · first shared 2017 · discussion on HN
- 12Tracking changes in postgresql1 people · 2 points · first shared 2013 · discussion on HN
- 13Shrinking the storage footprint of data (PostgreSQL)1 people · 2 points · first shared 2014 · discussion on HN
- 14Time in PostgreSQL: The simple way › Cybertec - The PostgreSQL Database Company1 people · 2 points · first shared 2013 · discussion on HN
- 15Running production PostgreSQL systems on ARM architecture? [pdf]1 people · 2 points · first shared 2017 · 1 comments on HN
What HN people wrote when they linked it
evadne · 2017-07-09
I noticed that this article was written in 2015 and I wonder whether SKIP LOCKED in Postgres 9.5 could have changed the entire situation,
https://blog.2ndquadrant.com/what-is-select-skip-locked-for-in-postgre…
http://www.cybertec.at/skip-locked-one-of-my-favorite-9-5-features
Also, duplicate of: https://news.ycombinator.com/item?id=9576864
pritambarhate · 2017-01-27
Linked to the PDF because it gives more detailed information. Here is the introduction to it: http://www.cybertec.at/running-production-postgresql-systems-on-arm-ar…
idorosen · 2016-09-26
BDR is a nice building block to multi-master PostgreSQL. I'm looking forward to parallel aggregates in 9.6 being added to core. Using the agg[0] extension for something as core as using more than one core per (aggregate function) query felt strange. (I wonder if the time has come to decouple connections from processes/threads in postgres, as well...)
[0]: http://www.cybertec.at/en/products/agg-parallel-aggregations-postgresq…
dsp1234 · 2016-03-07
I'm honestly not sure how much of that would be related to the query planner vs linked server network code.
Speaking of which, linked servers (cross OS, cross SQL type database queries) are native in MSSQL, but require a third party odbclink in postgres, which does not appear to have been updated since 2010[0].
[0] - http://www.cybertec.at/en/media/downloads/
atomic77 · 2016-02-26
AFAICT pgsql controls this with the random_page_cost and seq_page_cost variables [1], which are relative and seem to default to 4 and 1 respectively. It doesn't look like these are automatically adjusted though based on the hardware.
Some searching and I found some blog posts suggesting lowering them both to closer to one on SSDs [2]. Indeed, the documentation says you can set them both to 1 if you are sure that the database will be fully cached in RAM.
[1] http://www.postgresql.org/docs/current/static/runtime-config-query.htm…
[2] http://www.cybertec.at/2013/01/better-postgresql-performance-on-ssds/
spaznode · 2014-04-03
I was under impression from reading another article that setting the correct kernel io scheduler to use helps with a related concept, maybe possibly: http://www.cybertec.at/postgresql-linux-kernel-io-tuning/
Lennie · 2013-11-13
Setting up a new PostgreSQL slave is just a one-liner with pg_basebackup in 9.3 (no outside tools needed):
http://www.cybertec.at/postgresql-9-3-one-line-one-slave/