Mailing-List: contact derby-user-help@db.apache.org; run by ezmlm
Precedence: bulk
Reply-To: "Derby Discussion" <derby-user@db.apache.org>
Received-SPF: pass (herse.apache.org: local policy)
Date: Fri, 03 Nov 2006 10:40:21 +0100
From: John Embretsen <John.Embretsen@Sun.COM>
Subject: Re: Heap and stack size for Derby server
In-reply-to: <454AFFFA.1010107@intland.com>
Sender: John.Embretsen@Sun.COM
To: Derby Discussion <derby-user@db.apache.org>
Reply-to: derby-user@db.apache.org
Message-id: <454B0E85.7020600@Sun.COM>
MIME-version: 1.0
Content-type: text/plain; format=flowed; charset=ISO-8859-1
Content-transfer-encoding: 7BIT
References: <4548B82D.2030301@intland.com> <454AFFFA.1010107@intland.com>
User-Agent: Thunderbird 1.5.0.5 (X11/20060730)

Robert Enyedi wrote:
> Does the lack of answers mean that there are no best practices in 
> setting the stack and heap limits for Derby?

Well, not specifically, that I know of. It all depends on how you (and clients 
connecting to your server) are using Derby.

Default stack sizes work very well in most cases. I have little experience with 
StackOverflowErrors myself, but there are several Jira issues you can check out 
for comments on adjusting _heap_ size, for example DERBY-1713 
(http://issues.apache.org/jira/browse/DERBY-1713) and DERBY-1564.

I guess someone with more knowledge about Derby's internals may be able to tell 
you why your specific query requires a larger stack size.

As for heap size, I would not run the network server with any less than 32 MB of 
max heap space, regardless of the client code. Most of the tests I've run 
against Derby work fine with default heap size (usually minimum 64 MB max size, 
depending on JVM and hardware). Your mileage may vary...


-- 
John