We are thinking about this approach to consolidate some small applications of individual clients and to perform administration (data/log/pgbackrest backups etc.) separately which is not possible on a single cluster.
We have tried to create and start the cluster with different port 5433 on separate data directory but getting some errors.
- I replaced the default port with 5433 in the postgresql.conf file.
- We are not using Ubuntu , hence, I think we can’t use the pg_cluster
./initdb -D /opt/PostgreSQL-10/whr_data1_5433 -U postgres -W
./pg_ctl -D /opt/PostgreSQL-10/whr_data1_5433 -l /opt/PostgreSQL-10/whr_dlog/start.log start
Error while creating Cluster:
[postgres@ip-172-31-32-99 bin]$ cat /opt/PostgreSQL-10/whr_dlog/start.log
2022-07-18 13:11:14.404 UTC [13418] LOG: could not bind IPv4 address "127.0.0.1": Address already in use
2022-07-18 13:11:14.404 UTC [13418] HINT: Is another postmaster already running on port 5432? If not, wait a few seconds and retry.
2022-07-18 13:11:14.404 UTC [13418] WARNING: could not create listen socket for "localhost"
2022-07-18 13:11:14.404 UTC [13418] FATAL: could not create any TCP/IP sockets
2022-07-18 13:11:14.404 UTC [13418] LOG: database system is shut down
[postgres@ip-172-31-32-99 bin]$
On 7/18/22 04:46, Guillaume Lelarge wrote:
[snip]Hi,
It would make things easier for you to use a single cluster, and as many databases as you want in this cluster.
It would make PITR restoration of a single database (or small set of databases) impossible (since such backups -- and therefore restores -- are always binary "whole cluster").
--
Angular momentum makes the world go 'round.
