hi folks,

i'm suspecting that idle timeouts on sql Connection objects may not be timing out as i 
expect them to, and i'm trying to trace the behaviour to code in IdleRemover and 
InternalManagedConnectionPool.

what i think i saw in my application was that even though some connections were idle 
for many many hours, they were not being reclaimed by the server --- even though i 
left the default timeout to 15 minutes.

looking in the source code this morning (jboss-3.2 Branch_3_2 head), i see that 
InternalManagedConnectionPool.removeTimedOut() just looks at the oldest connection and 
checks if it timed out. if its not timed out, it does nothing more. 

what i'm thinking of is the situation where the first connection is continuously used, 
and more recently made connections time out. thus, based on my understanding of the 
source code logic (which may be flawed since i just started looking at it!), those 
later connections would never get removed due to time out. is that so? and if it is, 
is it the desired behaviour? i'm guessing this may be the cause of the connection 
resource leak i saw (aside from my bad programming practice where i forgot to execute 
Connection.close() in some cases!)

thanks,
mike

<a 
href="http://www.jboss.org/index.html?module=bb&op=viewtopic&p=3825086#3825086";>View 
the original post</a>

<a 
href="http://www.jboss.org/index.html?module=bb&op=posting&mode=reply&p=3825086>Reply 
to the post</a>


-------------------------------------------------------
This SF.Net email is sponsored by: IBM Linux Tutorials
Free Linux tutorial presented by Daniel Robbins, President and CEO of
GenToo technologies. Learn everything from fundamentals to system
administration.http://ads.osdn.com/?ad_id=1470&alloc_id=3638&op=click
_______________________________________________
JBoss-Development mailing list
[EMAIL PROTECTED]
https://lists.sourceforge.net/lists/listinfo/jboss-development

Reply via email to