db-derby-user mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Craig L Russell <Craig.Russ...@Sun.COM>
Subject Re: OutOfMemoryErrors when testing Derby with DOTS
Date Tue, 31 Jan 2006 19:31:17 GMT
Hi John,

On Jan 31, 2006, at 9:14 AM, John Embretsen wrote:

> Tuesday, January 31, 2006, 3:03:38 AM CET, Craig L Russell wrote:
>
>> Hi John,
>
>> On Jan 30, 2006, at 12:11 PM, John Embretsen wrote:
>
> [snip old quotes]
>
>>>> Any reason why itemID is bound to the preparedStatement and not
>>>> itself passed as a parameter?
>>>
>>>> Generally, preparing statements like this (those that include
>>>> parameters) just before use has little benefit. The big win is to
>>>> prepare the statement and then reuse it everywhere. So if you  
>>>> change
>>>> the updateBidItemSQL to include the itemID as a ? parameter, then
>>>> pstmt.setInt(3, itemID) you will get the benefit of prepared
>>>> statements.
>>>
>>>> Craig
>>>
>>> I am fully aware of this rather strange and inefficient use of
>>> PreparedStatements in the DOTS test case. As I stated in my first E-
>>> mail
>>> in this thread:
>>>
>>>  "I don't know the rationale behind doing it this way. It may be  
>>> just
>>>  "sloppy code", or it may be intentional."
>
>> I was trying to be nice. I'll reiterate my point. If your use of
>> prepared statements is limited to the following pattern:
>
>> PreparedStatement ps = conn.prepare(savedSQLStatement + "" + itemID +
>> "");
>> ps.setInt(...);
>> ps.execute();
>
>> then you are not taking advantage of prepared statements. Your
>> prepared statement cache is not doing any good at all.
>
> I hope I have not given anyone the impression that I am an advocate  
> for
> such use of PreparedStatements. I am not. I am testing Derby, and the
> purpose of testing Derby is to find faults in Derby, so that they  
> can be
> fixed.

I'm afraid I missed the opening shots of this discussion, and in  
particular where you said that you were using the referenced code as  
tests for the robustness of the various implementations. I would have  
put a slightly different slant on my comments.

> Sure, changing or not running these tests will make the symptoms
> go away, but is that a good long term strategy for Derby? Do we  
> have any
> guarantees that other users of Derby (directly or indirectly) will not
> stumble across this or similar memory leaks in the future, when the
> costs of experiencing and fixing such bugs may be a lot higher?

If this test code is representative of the actual application, then  
the application is in trouble and should be reimplemented in the jdbc  
area. It is a very well-understood requirement of well-behaved  
applications that result sets and prepared statements and connections  
be closed when they are no longer needed. And testing what happens to  
ill-behaved applications in various configurations doesn't seem to be  
a good use of anyone's energy.

Long term, I don't believe that Derby or any other implementation  
should try to optimize for applications written without regard for  
good programming patterns.

That said, OutOfMemoryError is unfortunate, but perhaps unavoidable.  
Does the test succeed, given enough memory? Does closing the result  
sets and prepared statements change the behavior? How can Derby know  
whether you intend to use the result set and prepared statement  
again, and you actually want to keep them open? Do you want Derby to  
internally close result sets and prepared statements that it guesses  
you no longer want? In a large system, wouldn't it be a bug in Derby  
if Derby closed result sets and prepared statements that the  
application still wanted?
>
>> This is the hammer that is making your head hurt. Before you can see
>> if aspirin helps, put down the hammer.
>
> Personally, I would prefer a database with a strong enough helmet to
> withstand such hammer hits. Someone else may find that hammer one
> day, and hit you again ;)

These particular hammers are in an area clearly marked "Danger: Use  
of these hammers may be injurious to your sanity". ;-)

Seriously, if you are trying to evaluate different databases'  
behavior under a simulated application scenario, you should try to  
duplicate in your tests how the application actually behaves. And if  
the test is shown to have programming anomalies, and fixing the test  
informs better behavior in the application, then I'd say the test  
succeeded. After removing the anomalies in the test, you can see the  
underlying behavior of the system in various configurations, and have  
a better way to evaluate alternatives.

Regards,

Craig
>
> -- 
> John
>

Craig Russell
Architect, Sun Java Enterprise System http://java.sun.com/products/jdo
408 276-5638 mailto:Craig.Russell@sun.com
P.S. A good JDO? O, Gasp!


Mime
View raw message