Background

Symphony OCR crashes unexpectedly with the user interface showing java errors.  If you open the "C:\Program Files (x86)\Trumpet\SymphonyOCR\logs\maestro.log" log file, you'll notice an entry that looks like this:

2020-06-01 06:43:15,637 [Analyzer-2] ERROR com.trumpetinc.maestro.processormanager.ProcessorThreadPoolExecutor - Processing task failed with an error: Java heap space
java.lang.OutOfMemoryError: Java heap space

Cause

Symphony OCR requires a substantial amount of memory to be available for analyzing and processing documents.  Some documents require more memory than what Symphony OCR is able to allocate to analyzing and processing the document.

Resolution

• In version 7.2.49 and higher, Symphony OCR increased the default memory allocation to 1024M.  If you haven't already done so, go ahead and update to this version or later
• Open the "C:\Program Files (x86)\Trumpet\SymphonyOCR\launch.ini" file in Notepad.  Ensure the 'MaxHeap' entry is set to '1024m'.  If it is not, change the value and save the file.  Restart the Symphony OCR program/service after the change is made to verify the Analysis queue is processed.
  
• Lastly, if the above don't resolve your issue, there still may not be enough memory allocated to process the very small subset of documents.  We've seen this with large topographical maps for example. The following steps can be used to manually ignore offending documents. This process may need to be repeated many times until the analysis backlog has been processed:
  
  • Stop the Symphony OCR program/service
  • Open the config file "C:\Program Files (x86)\Trumpet\SymphonyOCR\config\settings.xml" in Notepad (this is the default location, your system could be different)
  • Find the following lines and change the 'autostart' value from "true" to "false"
    
    <ocrManager autostart="true" maxThreads="0"/>
    <analyzerManager autostart="true" maxThreads="0"/>
    
    
  • Start the Symphony OCR program/service
  • Open the log file "C:\Program Files (x86)\Trumpet\SymphonyOCR\logs\maestro.log" in Notepad
  • Go to the bottom of the file and locate the last document with the status INPROCESS — Copy the path or Document ID number to the clipboard
    
  • Select 'Lookup Document' on the main Symphony OCR interface, and paste the path or document id number from the clipboard and click the "Query" button
  • In the details of the document, choose "Ignore"
    
  • After the file is ignored, restart the Analyzer — If Symphony OCR crashes again, repeat the above process until the Analyzer backlog is empty.
  • After the Analyzer queue has been fully processed, reset the default values in the "C:\Program Files (x86)\Trumpet\SymphonyOCR\config\settings.xml" config file — Find the following lines and change the 'autostart' value from "false" to "true"
    
    <ocrManager autostart="true" maxThreads="0"/>
    <analyzerManager autostart="true" maxThreads="0"/>
  
  

If you are unable to find the offending file(s), feel free to reach out to us at support@trumpetinc.com and we'll be happy to help.

If you do find an offending file is causing the problem, we'd love to get a representative sample to investigate in detail.  If possible, you can upload the file here:  https://extranet.trumpetinc.com/upload/clientdata