On Sat, Feb 27, 2010 at 3:38 PM, Craig Ringer <craig@xxxxxxxxxxxxxxxxxxxxx> wrote: > Can you boil this down to a simple PHP test-case that connects to a dummy > database and repeats something that causes the backend to grow in memory > usage? Trying to do this - by progressively cutting things out of your test > until it stops growing - will help you track down what, exactly, is causing > the growth. Thank you for your suggestion. I have done this, and in doing so I have also discovered why this problem is occurring. My application uses a class that abstracts away the db interaction, so I do not normally use the pg_* functions directly. Any time any statement was executed, it created a new "named" prepared statement. I wrongly assumed that calling pg_free_result() on the statement resource would free this prepared statement inside of postgres. I will simply modify the class to use an empty statement name if there is no need for it to be named (which I actually need very infrequently anyway). I have attached the script I created to test with, for those who are interested. The first line of the script has the connection string. I used a db called testdb. run from the command line with: php -f test3.php Note my comment in the php file <<<<<< UNCOMMENT THIS LINE AND MEMORY ISSUE IS FIXED Thanks for the help everyone. Chris
Attachment:
test3.php
Description: Binary data
-- Sent via pgsql-performance mailing list (pgsql-performance@xxxxxxxxxxxxxx) To make changes to your subscription: http://www.postgresql.org/mailpref/pgsql-performance
