You know what's funny about this article.... Back in 2012 I had the same problems with Windows Update and that is what forced me to go to a mac. I've never looked back.

That's fine , their repmgr postgres repository was a joke.

"Discontinuing this library of 120 things for everyone is fine, *I* didn't like *one of them* I won't say why"

After setting up numerous clusters with repmgr and patroni along with running them in zero down time production... This is the very last plugin i would ever install. I like to sleep at night.

Lots of changes have to be made to your github action workflow to actually make this work properly.

What things have you needed to change? Can't say I've ever needed to do that but of course I've only used it on a few projects.

In GitHub we use an OIDC token to access some AWS resources. Locally I need to populate tokens etc and so I have an `if: ${{ACT}}` and a not condition to populate it.

it's what I'm deploying everything on right now... it's still early and the update restarts everything... eeek.

Crypto has this covered with 4090's available on clore and several other platforms.

This is great, setting the same up with truenas wasn't as easy as I had hoped.

I just implemented master with read replica with the bitnami postgres-repmgr image. It's not perfect, but it works and running my own instances in aws instead of rds is going to save me close to 80% when i also add in purchasing a savings plan. By setting this up I've learned more about postgres than i ever ever wanted too. lol

Just a heads up, we were using the bitnami charts in our first attempt at an HA postgres instance in our k8s cluster, and we're currently figuring out how to move away from them because they had so many issues.

Granted, we might've messed something up and there are lots of factors, but you should manually check your individual nodes every so often to make sure none of them are going out of sync. Need to connect to each one directly, can't rely on what the bouncer/pgpool/loadbalancer is showing you to check the data is the same. It happened repeatedly to us, and wasn't obvious from any of our monitoring. In the end we had to scale down to 1 node while we sort out moving to a different operator.

interesting thanks for the heads up, i really appreciate it! I do need to find some sort of monitoring to ensure the db's are staying insync.

"They" == "those who can do that", that is, developers under their employers direction. Not every sentence containing "they" is necessarily a flat earth like conspiracy.