You are viewing a plain text version of this content. The canonical link for it is here.
Posted to dev@pdfbox.apache.org by "Thomas Chojecki (JIRA)" <ji...@apache.org> on 2014/11/20 16:04:33 UTC

[jira] [Created] (PDFBOX-2512) OutOfMemory while signing large documents

Thomas Chojecki created PDFBOX-2512:
---------------------------------------

             Summary: OutOfMemory while signing large documents
                 Key: PDFBOX-2512
                 URL: https://issues.apache.org/jira/browse/PDFBOX-2512
             Project: PDFBox
          Issue Type: Bug
          Components: Parsing, Signing
    Affects Versions: 1.8.7
            Reporter: Thomas Chojecki
            Assignee: Thomas Chojecki


While working with large documents, we found some memory issues.

1. The method close() in the COSDocument, clones the objectpool and does not clean it properly. The cloning in getObjects() cause a OutOfMemory exception.

2.The COSWriter copy the whole pdf into the memory for signing and does not use BufferedInputStream for the FileInputStream which also has a big performance impact. (PDFBOX-1798)

3. The cloning of COSStreams cause a OutOfMemory exception

I used the CreateSignature example with a about 150 MB big document from here:
https://cdn-reichelt.de/bilder/downloads/reichelt_01-2015_DE_B_HQ.pdf

Additionaly I add a RandomAccessFile to the PDDocument.load in the CreateSignature class.

PDDocument doc = PDDocument.load(document,new RandomAccessFile(new File("d:\\temp.bin"), "rw")); (this prevent the OOM for the third case)

The use of a BuffedInputStream in case two, will increase the signing speed from more than 5 minutes to less than 1 minute. 



--
This message was sent by Atlassian JIRA
(v6.3.4#6332)