<!DOCTYPE html PUBLIC "-//W3C//DTD HTML 4.0 Transitional//EN" "http://www.w3.org/TR/REC-html40/loose.dtd">
<html><body><p>Hi PGPOOL TEAM,<br><br>I have a 2x pgpool(v4.1.1) cluster 2x pg (v12.1, stream replication) - over 20 databases and currently about 1800 connections (about 900 per backend) - num init children 300, max pool 4, max conn 1100 with each month the number of clients for databases increases.<br><br>I have few problems and questions:<br><br>1. What causes the error in pgpool - "ERROR: Sorry, too many clients already" and "ERROR: unable to read data" and "ERROR: unable to read data from frontend"? On Postgresql log these errors do not occur...<br><br>2. I plan to increase the number of connections to around 3000-4000. Can pgpool manage and apply strategies - increase max_pool or increase num_init_children?<br><br>3. What are pgpool's capabilities / limits for handling a large number of simultaneous connections and a large number of connections in general?<br><br>4. How would be pgpool failover behavior after exceeding the single database connection limit - modify "alter database limit connections" on postgres database?<br data-mce-bogus="1"><style type="text/css">#icselection13 {width: 100% !important; height: 100% !important; padding-top: 0px !important; overflow: visible !important;}</style><br><br>--<br>Pozdrawiam
<br>Radosław Szczygieł</p></body></html>