Critical: Memory Leak in None, Request and Session Scoped caching

2 views
Skip to first unread message

Mark Mandel

unread,
Jul 18, 2008, 5:50:38 AM7/18/08
to transf...@googlegroups.com
All,

I've managed to track down a pretty serious memory leak in None,
Request and Session scope caching with Transfer.

Details can be found in the ticket:
http://tracker.transfer-orm.com/issue.cfm?p=89977683-A728-9CD3-ABD9545A91734422&i=2D0D297C-0BAF-8587-4401EBE7041F7641

Those of you are using these caches, or are experiencing memory leaks,
please grab the latest copy of Transfer from SVN, and test it out, you
should find the leaks no longer exist.

I've put Transfer through some fairly rigourous memory tests, and
subsequent heap introspections, and it is handling everything very
well. Let me know how your tests go.

Apologies for the leaks in the first place.

Mark

--
E: mark....@gmail.com
W: www.compoundtheory.com

Stephen Moretti

unread,
Jul 18, 2008, 11:07:40 AM7/18/08
to transf...@googlegroups.com


2008/7/18 Mark Mandel <mark....@gmail.com>:


All,

I've managed to track down a pretty serious memory leak in None,
Request and Session scope caching with Transfer.

Details can be found in the ticket:
http://tracker.transfer-orm.com/issue.cfm?p=89977683-A728-9CD3-ABD9545A91734422&i=2D0D297C-0BAF-8587-4401EBE7041F7641

Those of you are using these caches, or are experiencing memory leaks,
please grab the latest copy of Transfer from SVN, and test it out, you
should find the leaks no longer exist.

I've put Transfer through some fairly rigourous memory tests, and
subsequent heap introspections, and it is handling everything very
well.  Let me know how your tests go.


I've just tried to deploy the latest version of Transfer from the SVN trunk and I'm getting the following error.

Bean creation exception during init() of transfer.TransferFactory:Error while attempting to autowire object of type transfer.com.cache.CacheConfigManager:
[Line: 93 :: D:\Lib\transfer\com\util\javaloader\JavaLoader.cfc :: com.compoundtheory.objectcache.Config :: ]

Any thoughts?

Regards

Stephen

Stephen Moretti

unread,
Jul 18, 2008, 11:38:44 AM7/18/08
to transf...@googlegroups.com


2008/7/18 Stephen Moretti <stephen...@gmail.com>:

~groan~  I've been doing too much SVN merging/tagging/branching/exporting today.... 
I put all the wrong changes in the wrong place and exported all the wrong stuff wrongly.   It was all just wrong.

Sorry... :|

Stephen

Mark Mandel

unread,
Jul 18, 2008, 7:00:48 PM7/18/08
to transf...@googlegroups.com
That's okay ;o)

Better you got it solved.

Mark

--
E: mark....@gmail.com
W: www.compoundtheory.com

Bob Silverberg

unread,
Jul 20, 2008, 8:16:45 PM7/20/08
to transf...@googlegroups.com
Mark,

Boy was I glad to see this note. My app that went into production
last Wednesday did appear to have a memory leak. The last place I was
going to point the finger at was Transfer, so I spent quite a bit of
time monitoring and trying to find other reasons for the leak. I
downloaded and installed the latest from SVN as soon as I saw this
note and things are now greatly improved. Memory still seems to be
climbing slowly, but that could just be the cache filling up, as I'm
not removing anything from the cache currently.

When do you think you'll have the cache introspection stuff available?
I could use that to help me understand whether my increases in memory
are due to a growing cache.

Thanks,
Bob

--
Bob Silverberg
www.silverwareconsulting.com

Mark Mandel

unread,
Jul 21, 2008, 12:04:05 AM7/21/08
to transf...@googlegroups.com
I was going to get a bunch of it finished by today, but my new video
card has packed up (got replaced last thurs for failing on my totally
last week). Hopefully I should have a new one shortly.

In the short term, I would experiment with setting some global
maxminutestimeout / accessedminutetimout values, so your cache has
some discarding happening, without the use of the JVM.

Mark

Reply all
Reply to author
Forward
0 new messages