Suspected Memory Leak.

[Gavin Baumanis]

Hi Robin / Phil,

Just thought that I would give you an update.

"About" 5 weeks ago, I reported that on a fortnightly basis, the CF
Server crashed and that the error log reported that the server had run
out of memory.

After a meeting over coffee with Phil, We decided that we would we
stop using Galaxy and use our login service CFC, as a "normal"
singleton object.

Since we stopped using Galaxy a little over 5 weeks now, we have not
run out of memory and the server has not crashed / needed to be
restarted.

What does itmean?
Well I suppose it is simple enough to deduce that the cause was / is
Galaxy...
But that is a very simplistic viewpoint- and thus this email.

is there an appropriate way that we can test for certain that Galaxy
is to blame, as opposed to it being an issue with our CFC - but it is
only an issue when we use it with Galaxy?

We still use ColdSpring and it's bean factory for instantiating the
LoginService CFC.
The only difference is that we call it directly now.

As always - Thanks!

Gavin.

[Robin Hilliard]

On 18/06/2011, at 8:35 PM, Gavin Baumanis wrote:

> is there an appropriate way that we can test for certain that Galaxy
> is to blame, as opposed to it being an issue with our CFC - but it is
> only an issue when we use it with Galaxy?

Have you run any stress/load tests yet, or is it just normal uncontrolled production use?

The only appropriate way to determine this for sure is to run stress tests under reproducible,
controlled conditions. I did not see any memory leaks when I did my early tests on Galaxy,
but I would not be surprised if they existed.

Robin

ROBIN HILLIARD
Chief Technology Officer
ro...@rocketboots.com.au

RocketBoots Pty Ltd
Level 11
189 Kent Street
Sydney NSW 2001
Australia
Phone +61 2 9323 2507
Facsimile +61 2 9323 2501
Mobile +61 418 414 341
www.rocketboots.com.au