PD4ML getting out of memory errors

#26237

We have a few slices over at EngineYard and it seems like PD4ML is failing quite miserably, 95% of the time. We have about a gig and a half worth of RAM, but from our local tests it looks like PD4ML uses ~70mb when we run it. So we don’t understand how we’re getting out of memory errors.

Here’s the exact error:

Error occurred during initialization of VM
Could not reserve enough space for object heap
Could not create the Java virtual machine.

We also get another one pertaining to “out of memory”.

The java version we have on our local machine (which PD4ML runs flawlessly) is “1.5.0_16” (32bit) and the one installed on our slices is “1.6.0_11” (64bit).

Any idea’s why this would be happening?

#27426

As far as I remember the default max heap size of JVM is limited by 64mb (which matches “~70mb”, you got).

In order to increase it you need to run your java runtime or appserver with a JVM command-line parameter like the following:

-Xmx512m
(defines 512 mb of heap size, sufficient for most of PD4ML applications)

#27427

Hi PD4ML Support team,
I have a question regarding my “out of memory” problem.
My environment:
– OS: AIX
– Apps Server: Weblogic 6 (JDK 1.3)
– PD4ML 350 Pro

I tried to use PD4ML 350 Pro to print chinese character in PDF by accessing TTF file.
The TTF file size is 22 MB.

The parameter of java options in weblogic start script is the following
-Xms=128m -Xmx=256m -> always throwing error “out of memory” when converting to PDF
-Xms=512m -Xmx=512m -> always throwing error “out of memory” when converting to PDF
-Xms=128m -Xmx=512m -> working properly

My question is why when I set the heap size -Xms=512m -Xmx=512m, always throwing out of memory
and -Xms=128m -Xmx=512m it can work properly.

Is my setting above correct (-Xms=128m -Xmx=512m) ?

Thank you

[Author]

Posts