Geo: Investigate support for multiple databases in a single secondary cluster
UPDATE: We've consolidated the work necessary to support high-availability PostgreSQL using Patroni in &2536 (closed). I am closing this to remove duplication.
A customer was attempting to set up multiple databases within a single Geo secondary cluster with Consul. I'm not quite sure how they had their configuration set up, but I think it was inetnded to look something like:
graph LR;
A(Geo secondary DB A) -- replication slot --> PRIMARY(Geo primary DB)
B(Geo secondary DB B) --> PRIMARY
C(Geo secondary DB C) --> PRIMARY
They had Consul and pgbouncer running, so when we shut down A
, B
came up but thought it was a primary (i.e. SELECT pg_is_in_recovery()
was f
).
Is this mode supported, and if so what is the proper way of doing this?
Edited by Fabian Zimmer