Hi Roman,

Thanks for your prompt response.

>I assume you mean metadata cache.
>
>

Yes. Sorry I was in a hurry before.

>
>So, if this issue is still there, it should be enough to close the
>connection. If you use connection pool, this will not help, since connection
>will not be closed, only returned to the pool. Currently pool has a
>shutdown() method, which will close all connection. However, after shutdown
>you cannot obtain connections anymore (at least you should not). So, in this
>case you need to add a method called restart() that would walk through all
>connection queues and close all idle connections. Also it should mark all
>working connections to be physically closed when they are returned back to
>the queue.
>
>

I am using the Classic Server and connection pooling. It appears there
is still the problem you mentioned above. I will give the restart
method a try tonight. Thanks for the help and sorry for the short email
before.

Steve