Обсуждение: question with beta5

Hey, I decided to download and run beta5 through a few paces, and just ran 
into something that I wouldn't think it good, but maybe it's expected.

I loaded a decent size database that we support.  From within psql, I 
performed a "select * from quesresp" which has 4.6 million rows.  After doing 
that, psql just started comsuming more and more memory until both my RAM, and 
all my swap were full. I killed the query before I had any big problems.   
From top I could see that psql was using close to 500M.

Is this expected behaviour?

Additional possibly relevant information.

testlang=# \d quesresp                                     Table "quesresp"   Column     |          Type          |
               Modifiers
 
---------------+------------------------+---------------------------------------------------id            | integer
          | not null default 
 
nextval('quesresp_id_seq'::text)applicationid | integer                | not nullquestionid    | integer
|not nullanswerid      | integer                | not nullfreeformresp  | character varying(255) |skillorder    |
integer               |
 
Unique keys: quesresp_app_ques_ans_sk_ndx,            quesresp_pkey

testlang=# select version();                          version
-------------------------------------------------------------PostgreSQL 7.2b5 on i686-pc-linux-gnu, compiled by GCC
2.96
(1 row)

Stock Redhat 7.2 with all patches installed (2.4.9 kernel etc...)

Thanks much,

Matt O'Connor

"Matthew T. O'Connor" <matthew@zeut.net> writes:
> I loaded a decent size database that we support.  From within psql, I
> performed a "select * from quesresp" which has 4.6 million rows.
> After doing that, psql just started comsuming more and more memory
> until both my RAM, and all my swap were full.

It's always done that...
        regards, tom lane