From: | Ed Loehr <pggeneral(at)bluepolka(dot)net> |
---|---|
To: | Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us> |
Cc: | pgsql-general(at)postgresql(dot)org |
Subject: | Re: Hash Join vs Nested Loops in 7.2.1 ... |
Date: | 2002-04-09 17:44:08 |
Message-ID: | 3CB32868.1000508@bluepolka.net |
Views: | Raw Message | Whole Thread | Download mbox | Resend email |
Thread: | |
Lists: | pgsql-general |
Tom Lane wrote:
>
>>Second, the same query sometimes takes 10-50 seconds shortly after
>>possibly a dump or other high-data-volume queries are executed, after
>>which it then returns to 1 second execution time. Getting crowded out
>>of shared memory?
>>
>>>Sounds like it. What shared-buffers setting are you using? How much
>>>RAM in the box?
>
>>shared_buffers = 256
>
> That's not a lot --- 256*8K = 2MB. You might try something in the low
> thousands.
>
>>RAM: 2.4GB, maybe? Not that familiar with HPUX mem setup...
SAM indicates 512MB of RAM. I upped the shared buffers from 256 to 4096, and
the hashjoin query came down from ~90 seconds to 10, still 10x slower than
the 1-sec nested loops. Is that a performance difference you'd expect
between hash and nested loops on this query because of EXISTS?
Ed
From | Date | Subject | |
---|---|---|---|
Next Message | Jean-Luc Lachance | 2002-04-09 17:49:46 | Re: Getting statistics from tables |
Previous Message | Matthew Stanfield | 2002-04-09 17:38:33 | Re: PostgreSQL and C# |