In my experience (4 Postgresql cluster 2 pg12 in test and 2 pg10 in production, 3 of them in Kubernetes, one on discrete VM's), Patroni is severely biased towards AWS-based containers, so it was not an option for my standalone K8s clusters managed by Rancher and based on OCI and DigitalOcean VM's.
In my tries, Patroni was continuously creating and destroying the first replica, because of a failed connection to itself, on the very same endpoint you found successful some log lines above.
Investigating the issue further, I've noticed that was not an issue in AWS, but nobody had been testing it outside of AWS, at least at that moment (October 2020).
So, to make a long story short, if you're on AWS you can try Patroni, otherways I would recommend maybe bitnami-repmgr, which in my tries was very close to working in a real scenario. (I'm still living without an auto-failover solution, at the moment).
Adalberto
Hi,
You can investigate patroni in order to manage failover,replication and switchover. The tool mentioned is an open-source tool.
Regards.
Hi, I have to configure a postgresql in high availability.
I want to ask you what tool you recommend to manage replication and failover or switchover.
Thanks.
Regards.-
--
Hüseyin DEMİRDatabase Engineer
