Performance issue when using pgpool for fail-over and recovery of postgresql DB
I am working on a POC of setting up PGPOOL with postgresql for failover and recovery.
Everything is working fine but I am observing huge performance delay with transaction.
Following is my setup..
PGPOOL : Running on separate Amazon Instance (Used for fail-over and recovery[No loa balancing and streaming replication])
POSTGRESQL DB: Setup on two different Amazon Instance with streaming replication setup .
Application Server: Apache running on separate Amazon Instance and connect to pgpool which intern connected to postgresql servers.
In normal scenario where Apache is directly connected to postgresql it takes 5 -7 seconds to complete the transaction...
But if I am connecting through pgpool it is taking more than 40 seconds to complete the transaction.
Note: How can I change the log level (PGPOOL logs all the statements tried to change the log level but it does not work)
Please help me with the possibilities how can i improve the performance
I am not using load balancing of pgpool...
In my Application there is high volume of wite operation...
Is there any alternative of pgpool?? Can repmgr can do all what pgpool can do?