--- John Martin <[EMAIL PROTECTED]> wrote: <snip> > It seems to me that the insert/delete work around is > just a bandaid and in a > production environment that is not being pounded on > 24x7 is going to run > into the dwindling connection pool issue. To fix > this problem, I think that > the connection pool manager might need a thread that > just cycles throught > the connections and just reconnects them whenever > they fail. I haven't > confirmed this but would imaging that ofBiz would > need to be restarted if I > were to recycle mySQL since all the connections in > the pool would go stale. > Am I off on my assumption here?
You caught me. It does appear to be just a bandaid. The only places where I've used this are deployments where there aren't too many concurrent connections. So it has worked so far. It's those medium sized sites this would be a problem for (too large and you're using all of the connections inside of the 8 hours, too small and the one gimpy connection is good enough :-) ). If you can think of a better way to handle it, I'd be more than happy to help test it out. OFBiz handling the database stuff is one of the reasons that originally drew me to it. So, I've always been apprehensive about messing with it.
