FATAL: terminating autovacuum process due to administrator command [closed]

Question

I am running postgresql 8.4 on Ubuntu 10.0.4. I was in the process of creating schemas/importing data when I came across some error messages. Here is a snippet of the output at the console:

CREATE AGGREGATE
CREATE AGGREGATE
CREATE DATABASE
FATAL:  terminating autovacuum process due to administrator command
CREATE DATABASE
CREATE DATABASE
CREATE DATABASE
LOG:  unexpected EOF on client connection
LOG:  unexpected EOF on client connection
LOG:  unexpected EOF on client connection
LOG:  unexpected EOF on client connection
LOG:  unexpected EOF on client connection
LOG:  unexpected EOF on client connection
LOG:  unexpected EOF on client connection
LOG:  unexpected EOF on client connection

BTW, I have checked and the postgres processes seem to be still running:

username@localhost:~/work/scripts/init$ ps aux | grep postgres
postgres 13196  0.1  0.1 102240  6264 pts/1    S    00:23   0:00 /usr/lib/postgresql/8.4/bin/postgres -D /mydata/pgdbdata
postgres 13198  0.0  0.2 102456  8612 ?        Ss   00:23   0:00 postgres: writer process                                    
postgres 13199  0.0  0.0 102240  1584 ?        Ss   00:23   0:00 postgres: wal writer process                                
postgres 13200  0.0  0.0 102568  1888 ?        Ss   00:23   0:00 postgres: autovacuum launcher process                       
postgres 13201  0.0  0.0  73956  1576 ?        Ss   00:23   0:00 postgres: stats collector process                           
postgres 13263  0.0  0.1 104276  6484 ?        Ss   00:25   0:00 postgres: postgres postgres 127.0.0.1(42338) idle 

What is likely to be causing this and how may I resolve the problem?.

Answer 1

the output suggests that PostgreSQL has been started from the console and that multiple client connections are fighting for resources. My recommendations:

1. Start PostgreSQL in a way that properly logs, ideally via the init scripts

2. Try reducing your maximum settings allowed. This often increases performance. Double your number of cores is a good place to start.

3. If these happen and it continues please post both PostgreSQL and Syslog snippets.