Re: Tracking open connections

[David Gibbs <david@xxxxxxxxxxxx>]

Donald Whittaker wrote:
> Perhaps if you journalled the table (including opens
> and closes) you might be able to prove (or disprove)
> this theory.
> If the lock is encountered from the JT400 side, and if
> the code is still in development, maybe a CommandCall
> for wrkobjlck could be inserted.

The presence of a lock isn't really in question ... that's the only 
thing that could cause the object in use message.

The thing I want to determine is where the open connection is ... and 
perhaps some way to identify it.

This code is in QA at the moment ... so I can't do command alterations. 
 I can, however, make modifications to the JDBC url.

I had a similar problem a while ago and IBM helped me identify that 
there were indeed open connections ... and they had a tool that did some 
trace analysis, but I don't know what tool they used or how to generate 
the trace information.

david