Hi All,
Do you have a ping strategy in CA Riskfort (ODBC) in order to keep TCP connections available? Do Riskfort has a feature to force reset iddle connections?
The idea is keep the connections alive, to avoid TCP connection deads.
For example our linux kills TCP connection after 30min without use and pool connections doesn't know about this event, when Riskfort tries to use the connection is verified that isn't alive and it will be marked as "bad", however this connection isn't deleted of the pool and Riskfort will try uses in the future (but never will work). After some time, all connections are "bad" and server can't process requests.
.....
ArDBPoolManagerImpl::getLockedDBConnection: Pool marked as active [primary] is not alive *now*. Will check another pool
ArDBPoolManagerImpl::recoverPool: [primary] pool [arcotdsn] is marked bad. Will try to clean and connect...
ArDBPool::clean Pool [(primary)]: [1] number of connections are still held. Waiting for [100] milli secs.
...
Best Regards,
Porcel