Hi David, do you have test code that you use to run this? Ideally if we could get this into a junit test, I'd love to get some automated tests setup. (it's kind of hard to run junit tests against James since with Avalon it's hard to split code into individual pieces, but still would like to work towards that).
I use the inet driver and M$ SQL by defacto, so I should be able to recreate your situation. The exception you see is when the pool has tried 100 times to grab a connection unsuccessfully. It might be worth exposing some of these pooling settings (or just tweaking them), or I think Leif has added a usable DB pooler to Avalon, so maybe we can drop my pooler and try that one. Anyway, if you can share some code to run your tests, I'll try it and see what I can do. Serge Knystautas Loki Technologies - Unstoppable Websites http://www.lokitech.com/ ----- Original Message ----- From: "David Shpil" <[EMAIL PROTECTED]> To: "James User" <[EMAIL PROTECTED]> Sent: Monday, February 25, 2002 8:56 AM Subject: Spool manager on M$ SQL > Hi, all > > Lately I have started to evaluate James v2.0a2 with M$ SQL sever. > In a production env, I need a mail server to handle large amount of > small ( < 700b) messages. > To test James' behavior under such king of load, I run two threads, > which concurently sending and receiving mail through the server. > > So, here is my problem: > 1. Setting number of threads in spool manager to 1, makes James work > fine with 3-4 messages per second. If I increase message rate the > spool manager becomes overloaded and it takes a lot of time to get > mail delivered to local inbox. > 2. Setting number of threads in spool manager to 2 and more, causes > the following exception to be thrown (Full stack trace is attached) > > java.io.IOException: java.sql.SQLException: Giving up... no connections > available. > > I made the same test with different JDBC drivers: inet, ThinWeb, jnetdirect; > and I got almost identical results. > > Did any of you ever consider this problem? > Are there any test results available for a comparison? > > > Regards, > > David. -- To unsubscribe, e-mail: <mailto:[EMAIL PROTECTED]> For additional commands, e-mail: <mailto:[EMAIL PROTECTED]>