On 28-2-2014 19:08, Steffen Heil (Mailinglisten) wrote:

> Hi
>
>> Does your code by any chance 'restart' or 'shutdown' and then 'start' the pool occasionally? It looks like that has a higher chance of breaking it.
>
> I was tended to answer 'yes', as we can restart the pool.
> However, I looked into that and whenever we stop the pool we call 'dataSource.shutdown()' and we drop that reference completely. Lateron, we create a new FBWrappingDataSource.
> Also - while I cannot be absolutely sure - I suspect that this did not happen during the lasts test.
> I will run our tests again on Monday and see if the pool is restarted.

I have been able to 'sort of' reproduce the problem (at least the pool
occasionally 'loses' connections in my tests). The same test shows that
my changes fixes the problem, but I need to do some more tests and
scrutinizing the code to be sure.

Setting the min and max size of the pool to the same value seems to
reduce the possibility of the problem occurring.

Mark
--
Mark Rotteveel