A Performance Cheat Sheet for PostgreSQL
severalnines.com
A Performance Cheat Sheet for PostgreSQL
1–10 of 15 posts
Re: A Performance Cheat Sheet for PostgreSQL
#2quick & dirty - PG tunning: https://pgtune.leopard.in.ua
Re: A Performance Cheat Sheet for PostgreSQL
#3Hard to take perf advice from a website that never stops loading.
Re: A Performance Cheat Sheet for PostgreSQL
#4For visualizing EXPLAIN there is a cool tool: http://tatiyants.com/pev/
Re: A Performance Cheat Sheet for PostgreSQL
#5Does anyone have a direct link without signing up? Just put your logo in the pdf, brand it, but I'm not gonna sign up just to get spammed with advertising.
Re: A Performance Cheat Sheet for PostgreSQL
#6Does anyone have a direct link without signing up? Just put your logo in the pdf, brand it, but I'm not gonna sign up just to get spammed with advertising.
There's nothing to download. Everything is right there on the page.
Re: A Performance Cheat Sheet for PostgreSQL
#7Nice reference!
Does anyone have similar references for doing 0-downtime migrations? For example, building your indexes with CONCURRENTLY to prevent table locking.
Re: A Performance Cheat Sheet for PostgreSQL
#8Nice reference! Does anyone have similar references for doing 0-downtime migrations? For example, building your indexes with CONCURRENTLY to prevent table locking.
Not exhaustive but this has come in handy: https://www.braintreepayments.com/blog/safe-operations-for-h...
Re: A Performance Cheat Sheet for PostgreSQL
#9Nice reference! Does anyone have similar references for doing 0-downtime migrations? For example, building your indexes with CONCURRENTLY to prevent table locking.
If you are using Rails, this is a nice guardrail: https://github.com/ankane/strong_migrations
Even if you aren't on Rails or don't want to use the gem, it is still nice a checklist.
Re: A Performance Cheat Sheet for PostgreSQL
#10> It is not recommended to use RAID5 since the performance of this type of RAID for databases is not good.
This strikes me as a generalization that is, at best, questionable. (Also, these days, with spinning disks, I'd more likely do RAID6, but I expect it's comparable enough).
I'd tend to agree that, historically, it's been true, and is demonstrably true for certain situations, such as a read-heavy load or recovering from a failed disk.
However, is it really still true for modern RAID cards?
What about with SSDs? I'd expect that cost is still a significant consideration for anyone specifying hardware, making the difference between RAID5 and RAID10 important.