On 03/23/2012 07:15 PM, Bergquist, Brett wrote:
> It is almost as if the second instance was "starved". Through jvisualvm
> I did thread dumps and it seemed the second instance was processing but very slowly.
I'm not really sure what you mean by "instance" here. Is this a Network Server
configuration? An embedded configuration? How many JVMs are running the Derby
engine, against how many databases?
And what was the overall system performance like during this time?
That is, can you frame these observations in a higher-level description
- what kind of system was this (operating system, hardware, etc.)
- How CPU-busy was the overall system
- How was the overall memory management? Was all memory in use? How
much memory did the individual processes have? Were you swapping?
- How busy were the disks? How much read I/O, how much write I/O, etc.?
My first guess is that you increased the load on Derby without increasing
the Derby configuration parameters (specifically, the amount of memory
that Derby has to work with), causing Derby to be memory-constrained.
But it would be good to see that in a broader description of your experiment.