[ 
https://issues.apache.org/jira/browse/PDFBOX-4041?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16309930#comment-16309930
 ] 

Tilman Hausherr commented on PDFBOX-4041:
-----------------------------------------

I ran your program and had it do a few loops without trouble in release mode 
with the trunk. I set {{-Xmx250m 
-Dsun.java2d.cmm=sun.java2d.cmm.kcms.KcmsServiceProvider 
-Dorg.apache.pdfbox.rendering.UsePureJavaCMYKConversion=true}}. {{-Xmx250m}} is 
a pretty low value IMHO, I use {{-Xmx2g}} for my tests.

Maybe you had some troubles in your production environment, and then thought 
that what you saw in profiling was the cause?

{quote}
in my production environment the memory is not released for a very long time 
(some time hours)
{quote}
Are you using some individual settings for the gc in production? What jdk are 
you using? I am using 1.8.0_151 on windows 7.


> Memory Leak while converting pdf to images
> ------------------------------------------
>
>                 Key: PDFBOX-4041
>                 URL: https://issues.apache.org/jira/browse/PDFBOX-4041
>             Project: PDFBox
>          Issue Type: Bug
>    Affects Versions: 2.0.7
>            Reporter: Ajit
>         Attachments: File2_5Mb.pdf, PDFBOX-4041_MemProfile.png, 
> PDFConverter.java, heap.jpg, output.txt
>
>
> I am observing memory leak in PDF to images conversion simple program. I 
> realise the PDF that I am converting is bit bigger in size.
> Check the heap image and you will see that leak is happening even though I 
> have flushed the buffered image and set it to null explicitly. I have also 
> closed the document too.
> !heap.jpg|thumbnail! If you notice the image, the process was running at 2 
> GBs in only two runs. I am not able to attach the heap, because of its bigger 
> size. But note the image at the bottom and you will see that BufferedImage is 
> taking about 500 MBs.
> I have also provided the source code (few lines) and hoping I haven't done 
> any stupid mistake. 
> * I have gone through getting started 
> [https://pdfbox.apache.org/2.0/getting-started.html]  and set 
> -Dsun.java2d.cmm=sun.java2d.cmm.kcms.KcmsServiceProvider 
> -Dorg.apache.pdfbox.rendering.UsePureJavaCMYKConversion=true the options. It 
> does not help much
> * Maven Dependencies - pdfbox-tools, jai-imageio-core, jai-imageio-jpeg2000, 
> levigo-jbig2-imageio
> There are some debug messages related to ScratchBuffer, I searched about it 
> and it seems it is harmless. I came across the bug PDFBOX-3388. From the 
> comments, It seems it is relevant but harmless.



--
This message was sent by Atlassian JIRA
(v6.4.14#64029)

---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]

Reply via email to