Is it a memory leak in PostgreSQL 7.4beta? - Mailing list pgsql-hackers

From Hans-Jürgen Schönig
Subject Is it a memory leak in PostgreSQL 7.4beta?
Date
Msg-id 3F5094A5.3080500@cybertec.at
Whole thread Raw
Responses Re: Is it a memory leak in PostgreSQL 7.4beta?
List pgsql-hackers
I have done some beta testing with PostgreSQL 7.4beta2.
I have run a simple set of SQL statements 1 million times:

-- START TRANSACTION ISOLATION LEVEL READ COMMITTED;
INSERT INTO t_data (data) VALUES ('2500');
UPDATE t_data SET data = '2500' WHERE data = '2500';
DELETE FROM t_data WHERE data = '2500';
-- COMMIT;

The interesting thing was that my postmaster needed around 4mb of RAM 
when I started running my test script using ...

psql test < script.sql

After about 2 1/2 hours the backend process already needed 11mb of ram. 
looking at the output of top you can see that it seems to be in the 
shared memory area:

PID   USER     PRI  NI    SIZE  RSS   SHARE STAT    %CPU    %MEM   TIME     COMMAND
28899 hs        39  19   11456  11M   10620    R  N 89.8      2.9 
150:23 postmaster

this seems very surprising to me because I have no explanation why 
PostgreSQL should consume so much more memory than at the beginning of 
the test.
There are no trigger or something like that around.

The table I am working on consist of two columns (one timestamp, one int4).


In addition to that I have made a test with a different set of SQL 
statements. I have tried 1500 concurrent transaction on my good old AMD 
Athlon 500 box running RedHat Linux. It worked out pretty fine. The 
thing I came across was that my memory consumption raised during the 
first two hours of my test (from about 1 1/2 gigs to 1.7 gigs ram). 
pretty surprising as well.

does anybody have an explanation for this behaviour?
Regards,
    Hans




I have run 1500 concurrent transactions on an AMD Athlon box (RedHat 9).


-- 
Cybertec Geschwinde u Schoenig
Ludo-Hartmannplatz 1/14, A-1160 Vienna, Austria
Tel: +43/2952/30706; +43/664/233 90 75
www.cybertec.at, www.postgresql.at, kernel.cybertec.at




pgsql-hackers by date:

Previous
From: Thomas Swan
Date:
Subject: Re: ALTER TABLE ... TO ... to change related names
Next
From: Joe Conway
Date:
Subject: Re: database corruption