Re: performance question

"Devinder K Rajput" <Devinder(dot)Rajput(at)ipaper(dot)com> writes:
> I am running a query and getting the following output in the log:

> 2002-12-23 10:25:20 DEBUG: ProcessQuery
> 2002-12-23 10:25:20 DEBUG: CommitTransactionCommand
> 2002-12-23 10:25:20 DEBUG: StartTransactionCommand
> 2002-12-23 10:25:20 DEBUG: query: fetch 1000 from my_cursor
> 2002-12-23 10:25:20 DEBUG: ProcessUtility: fetch 1000 from my_cursor
> 2002-12-23 10:30:10 DEBUG: proc_exit(0)
> 2002-12-23 10:30:10 DEBUG: shmem_exit(0)
> 2002-12-23 10:30:10 DEBUG: exit(0)
> /usr/bin/postmaster: reaping dead processes...
> /usr/bin/postmaster: CleanupProc: pid 16940 exited with status 0
> 2002-12-23 10:35:10 DEBUG: proc_exit(0)
> 2002-12-23 10:35:10 DEBUG: shmem_exit(0)
> 2002-12-23 10:35:10 DEBUG: exit(0)
> /usr/bin/postmaster: reaping dead processes...
> /usr/bin/postmaster: CleanupProc: pid 16947 exited with status 0
> 2002-12-23 10:40:10 DEBUG: proc_exit(0)
> 2002-12-23 10:40:10 DEBUG: shmem_exit(0)
> 2002-12-23 10:40:10 DEBUG: exit(0)
> /usr/bin/postmaster: reaping dead processes...
> /usr/bin/postmaster: CleanupProc: pid 16967 exited with status 0

> The last five lines repeat again and again exactly every five minutes.

Presumably you've got a cron-driven task that connects to the database
once every five minutes? Anyway, those entries just show server
processes exiting normally. They've got nothing to do with your cursor
query.

> postmaster is using up to 90% of the CPU at times and is staying above 70%
> for most of the time.

When you haven't shown us the query that underlies the cursor, it's hard
to say anything useful, but I'd wonder about unconstrained joins and
similar mistakes. Or maybe you need to do some query optimization work.
What is the query, and what does EXPLAIN show as its query plan?

regards, tom lane