Michael,
It looks like something is really broken. Are you sure that each
Apache child has it's own connection to the database (netstat will
tell you)? It looks like the Inline::Java messages are all messed up.
Patrick
On 9/15/06, [EMAIL PROTECTED] <[EMAIL PROTECTED]> wrote:
Sorry for the delay in getting back. Train ride, client with limited net
accesss, etc...
>> I'm a little unclear as to how the JVM is handling the connections. If 2
>> apache
>> children are trying to access a static Java object, will they be using
>> the same
>> one?
>
> Yes, they should be using the same one. But if you use the PRIVATE
> option, it should use different class loaders, therefore allow for
> each connection to have it's own static objects. However. I must admit
> that I have no done a lot of tests with PRIVATE mode.
Ok, I'll explore this option.
>> What concerns me is that everything works fine for a while, but during
>> busier
>> periods stuff will start to go "wonky". Connections will be lost or
>> timeout and
>> I see stuff like this in the logs:
>>
>> (in cleanup) In method DESTROY of class
>>
UFT::PlunkIT::JavaService::com::ibm::as400::access::AS400JDBCConnectionHandle:
>> Malformed response from server:
>>
scalar:86.50.10.99.108.97.115.115.32.99.111.109.46.105.98.109.46.97.115.52.48.48.46.97.99.99.101.115.115.46.65.83.52.48.48.74.68.66.67.67.111.110.110.101.99.116.105.111.110.72.97.110.100.108.101.32.106.97.118.97.46.108.97.110.103.46.79.98.106.101.99.116.10.109.101.116.104.111.100.32.105.110.115.116.97.110.99.101.32.99.111.109.46.105.98.109.46.97.115.52.48.48.46.97.99.99.101.115.115.46.65.83.52.48.48.74.68.66.67.67.111.110.110.101.99.116.105.111.110.72.97.110.100.108.101.32.116.111.83.116.114.105.110.103.40.41.10.109.101.116.104.111.100.32.105.110.115.116.97.110.99.101.32.99.111.109.46.105.98.109.46.97.115.52.48.48.46.97.99.99.101.115.115.46.65.83.52.48.48.74.68.66.67.67.111.110.110.101.99.116.105.111.110.72.97.110.100.108.101.32.115.101.116.82.101.97.100.79.110.108.121.40.98.111.111.108.101.97.110.41.10.109.101.116.104.111.10!
>
> If the exclamation point is part of the message then that odd indeed.
> It means that the perl
> part received that data from the java part but it's malformed. That
> should never happen.
I copied that straight from the error log.
> Keep in mind that each connection to the JVM spawns a thread on the
> Java side. Maybe there is a race condition or something lingering
> around (either in the Inline::Java code or in the the rest of the Java
> code). What version are you running?
When I first posted I was running 0.50, but I upgraded to 0.51 and it
seemed to be doing much better, but I could still occassionally get
something:
Use of uninitialized value in split at
/home/mpeters/repos/UFT/trunk/code/application-server/lib/i686-linux/Inline/Java.pm
line 620, <GEN0> line 17.
(in cleanup) In method DESTROY of class
UFT::PlunkIT::JavaService::com::ibm::as400::access::AS400JDBCConnectionHandle:
Malformed response from server:
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!
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!
bm5lY3Rpb25IYW5kbGUgaXNDbG9zZWQoKQptZXRob2QgaW5zdGFuY2UgY29tLm!
libS5hcz
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
at
/home/mpeters/repos/UFT/trunk/code/application-server/lib/i686-linux/Inline/Java/Object.pm
line 361
Thanks for the help so far. Any other ideas I could try?
--
Michael Peters
Developer
Plus Three, LP
--
=====================
Patrick LeBoutillier
Laval, Québec, Canada
