Platform: java, Version: 17.4
Occasionally, all of my server threads lock up indefinitely while converting PDF documents with an HOCR callback. All of the RUNNABLE threads report the following stacktrace.
java.lang.Thread.State: RUNNABLE
at com.aspose.pdf.internal.ms.System.Collections.Generic.z2.m1(Unknown Source)
at com.aspose.pdf.internal.ms.System.Collections.Generic.z2.addItem(Unknown Source)
at com.aspose.pdf.internal.p62.z31.m1(Unknown Source)
at com.aspose.pdf.internal.p65.z1.m1(Unknown Source)
at com.aspose.pdf.TextSegment.setText(Unknown Source)
at com.aspose.pdf.TextSegment.m1(Unknown Source)
at com.aspose.pdf.TextBuilder.m1(Unknown Source)
at com.aspose.pdf.TextBuilder.appendParagraph(Unknown Source)
at com.aspose.pdf.TextBuilder.appendParagraph(Unknown Source)
at com.aspose.pdf.internal.p107.z10.m1(Unknown Source)
at com.aspose.pdf.internal.p107.z10.m1(Unknown Source)
at com.aspose.pdf.ADocument.convert(Unknown Source)
at com.aspose.pdf.Document.convert(Unknown Source)
at com.aspose.pdf.IDocument$convert$3.call(Unknown Source)
at com.platinumids.storage.ocr.pdf.PdfHocrProcessor$_process_closure1.doCall(PdfHocrProcessor.groovy:35)
Killing the threads and reprocessing the same documents fixes the issue. It doesn’t appear to be content related. The server has plenty of java heap available according to our instrumentation. How can I troubleshoot the cause further? There doesn’t appear to be any logging.