Thanks Bill and Andrew,

In my case, I am with a wonder based command line application. However, it
is mixed with a manual creation of a JDBC connection (in a separate thread
absolutely needed for utmost speed) and that is were I am getting the
exception. I am not really sure why as the biggest headache is that in my
dev environment, I do not get this exceptions but only in the production
env. Plus, why do I only get this exception on the manually created JDBC
connection and not from WO itself? (Strange really)

Reading Bill's insight, probably there is also a "this won't happen
exception" thingy or might even be possible it is running out of memory?
Need to do more investigation and a ton of caffeine.

Amiel

On Wed, Sep 2, 2009 at 11:17 PM, William Hatch <[email protected]> wrote:

> Thanks Andrew,
>
> I don't have problems with it not reconnecting, but in tomcat, it seems
> that exceptions blow up the memory eventually, even if they're caught and
> handled. That's my theory anyway; I could be completely wrong. But I've
> noticed a few other places where I may get loads of caught exceptions (you
> know, "this won't happen that often, so just catch the occasional
> exception..." type implementation, and then it turns out to happen a whole
> lot more than what you'd originally thought), and it crushed the server
> within minutes and it started to swap and thrash. After I'd refactored to
> handle that particular situation differently, the problems vanished. So I'm
> just relating this current issue with that experience. Also, memory settings
> are critical for GC purposes; on a 64 bit os, using Sun's java, I've found
> you really can't go much above 3GB max or GC will end up gradually consuming
> more and more time and the app eventually gets ridiculously slow.
>
> Thanks Andrew.
>
> Bill
>
>
>
> On Sep 2, 2009, at 12:44 AM, Andrew Lindesay wrote:
>
>  Hello Amiel & Bill;
>>
>> Is it possible that there could be a bug in the data-source in the
>> container you are using -- assume you are servletting?  I think MySQL server
>> does close the connection "overnight" if there is no traffic on the
>> connection, but I haven't experienced it doing this during operational
>> hours.  I seem to remember the "autoReconnect" does work.  Maybe you could
>> use a data source which runs a test query each time a connection is taken
>> from the pool?
>>
>> Regards;
>>
>>  Did you find a solution to this? I'm afraid I am also being killed by
>>> this.
>>> Could it be a timeout problem and MySQL is closing the connection? There
>>> is a MySQL option "autoReconnect" but I don't feel comfortable with it.
>>>
>>
>> ___
>> Andrew Lindesay
>> www.lindesay.co.nz
>>
>>
>


-- 
socket error: unable to connect to 127.0.0.1
 _______________________________________________
Do not post admin requests to the list. They will be ignored.
Webobjects-dev mailing list      ([email protected])
Help/Unsubscribe/Update your Subscription:
http://lists.apple.com/mailman/options/webobjects-dev/archive%40mail-archive.com

This email sent to [email protected]

Reply via email to