JRun OutOfMemory Restarts

Question

I've got a big async process that runs nightly on my CF8
server to rebuild my product database (should take 4-5 hours when
running properly), and it frequently (3-4 times per week) causes CF
to restart with a JRun OutOfMemory error:
# An unexpected error has been detected by Java Runtime
Environment:
# java.lang.OutOfMemoryError: requested 16384000 bytes for
GrET in
C:\BUILD_AREA\jdk6_04\hotspot\src\share\vm\utilities\growableArray.cpp.
Out of swap space?
# Internal Error (allocation.inline.hpp:42), pid=7024,
tid=3756
# Error: GrET in
C:\BUILD_AREA\jdk6_04\hotspot\src\share\vm\utilities\growableArray.cpp
The basic flow of the process is to do a few database queries
to gather about 25,000 items, and then process each one in a
grouped CFOUTPUT (need to group the query by product family,
otherwise I'd normally use CFLOOP). Within each grouped iteration,
I recycle a set of scalar variables for each family, so I'm not
creating any persistent variables. I invoke a few CFCs via CFINVOKE
to assemble some data, one of which does use Session variables, but
my understanding is that all this would be considered one session.
Then the data for that family is written to the database, and the
loop continues to the next family.
What I don't understand is why my JRun memory usage
continually creeps up and up over the course of the process. If I'm
not creating persistent variables as the loop proceeds, what's
accumulating there? I'm not caching any queries, I'm not building
any accumulative variables, but the memory goes up and up till it
causes a restart.
I'm running version 8,0,1,195765, with Java 1.6.0_04 using
the following args:
-server -Dsun.io.useCanonCaches=false -XX:MaxNewSize=16m
-XX:NewSize=16m -XX:+UseConcMarkSweepGC -XX:+UseParNewGC
-Dcoldfusion.rootDir={application.home}/../
-Dcoldfusion.libPath={application.home}/../lib
Can anyone suggest some steps I could take to get this thing
running better? I've used the Server Monitor to watch the progress,
although I can't use Memory Tracking (it eats the server up within
minutes). Am I missing something really obvious in the Java
configuration? Is the use of Session variables in the CFCs I invoke
a bad idea?

Answer 1

jarviswabi wrote:
> I've got a big async process that runs nightly on my CF8
server to rebuild my
> product database (should take 4-5 hours when running
properly), and it
> frequently (3-4 times per week) causes CF to restart
with a JRun OutOfMemory
> error:
>
> # An unexpected error has been detected by Java Runtime
Environment:
> #
> # java.lang.OutOfMemoryError: requested 16384000 bytes
for GrET in
>
C:\BUILD_AREA\jdk6_04\hotspot\src\share\vm\utilities\growableArray.cpp.
Out of
> swap space?
> #
> # Internal Error (allocation.inline.hpp:42), pid=7024,
tid=3756
> # Error: GrET in
>
C:\BUILD_AREA\jdk6_04\hotspot\src\share\vm\utilities\growableArray.cpp
That is not good. Try upgrading your JDK to 1.6u10 which
should handle
an out of memory problem more graceful. Also, how is the
resource
consumption on the server when this error occurs?
> What I don't understand is why my JRun memory usage
continually creeps up and
> up over the course of the process.
Do you have debugging enabled? If so, switch it off.
Jochem
Jochem van Dieten
Adobe Community Expert for ColdFusion