java - 文件系统交互异常: Could not access target file while using documents4j

Question

我正在使用documents4j将word文档转换为pdf，有时我遇到了异常

2016-03-28 09:29:16.982  INFO 3660 --- [pool-1-thread-2] c.d.c.msoffice.MicrosoftWordBridge       : Requested conversion from C:\conversion-temp\2b33637b-b74a-4aaa-ac65-a5ebc1eb3efc\temp3 (application/msword) to C:\conversion-temp\2b33637b-b74a-4aaa-ac65-a5ebc1eb3efc\temp4 (application/pdf)
2016-03-28 09:29:17.372 ERROR 3660 --- [http-nio-8080-exec-9] c.s.c.e.mappers.ExceptionMapper          : Exception while handling request

com.documents4j.throwables.FileSystemInteractionException: Could not access target file
  at com.documents4j.util.Reaction$FileSystemInteractionExceptionBuilder.make(Reaction.java:180) ~[documents4j-util-all-1.0.2.jar:na]
  at com.documents4j.util.Reaction$ExceptionalReaction.apply(Reaction.java:75) ~[documents4j-util-all-1.0.2.jar:na]
  at com.documents4j.conversion.ExternalConverterScriptResult.resolve(ExternalConverterScriptResult.java:70) ~[documents4j-transformer-api-1.0.2.jar:na]
  at com.documents4j.conversion.ProcessFutureWrapper.evaluateExitValue(ProcessFutureWrapper.java:48) ~[documents4j-util-transformer-process-1.0.2.jar:na]
  at com.documents4j.conversion.ProcessFutureWrapper.get(ProcessFutureWrapper.java:36) ~[documents4j-util-transformer-process-1.0.2.jar:na]
  at com.documents4j.conversion.ProcessFutureWrapper.get(ProcessFutureWrapper.java:11) ~[documents4j-util-transformer-process-1.0.2.jar:na]
  at com.documents4j.job.AbstractFutureWrappingPriorityFuture.run(AbstractFutureWrappingPriorityFuture.java:78) ~[documents4j-util-conversion-1.0.2.jar:na]
  at java.util.concurrent.ThreadPoolExecutor.runWorker(ThreadPoolExecutor.java:1142) ~[na:1.8.0_74]
  at java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor.java:617) ~[na:1.8.0_74]
  at java.lang.Thread.run(Thread.java:745) ~[na:1.8.0_74]

在此异常之后，documents4j 库将拒绝任何进一步的请求，并出现以下异常

com.documents4j.throwables.ConverterAccessException: The converter seems to be shut down
  at com.documents4j.util.Reaction$ConverterAccessExceptionBuilder.make(Reaction.java:117) ~[documents4j-util-all-1.0.2.jar:na]
  at com.documents4j.util.Reaction$ExceptionalReaction.apply(Reaction.java:75) ~[documents4j-util-all-1.0.2.jar:na]
  at com.documents4j.conversion.ExternalConverterScriptResult.resolve(ExternalConverterScriptResult.java:70) ~[documents4j-transformer-api-1.0.2.jar:na]
  at com.documents4j.conversion.ProcessFutureWrapper.evaluateExitValue(ProcessFutureWrapper.java:48) ~[documents4j-util-transformer-process-1.0.2.jar:na]
  at com.documents4j.conversion.ProcessFutureWrapper.get(ProcessFutureWrapper.java:36) ~[documents4j-util-transformer-process-1.0.2.jar:na]
  at com.documents4j.conversion.ProcessFutureWrapper.get(ProcessFutureWrapper.java:11) ~[documents4j-util-transformer-process-1.0.2.jar:na]
  at com.documents4j.job.AbstractFutureWrappingPriorityFuture.run(AbstractFutureWrappingPriorityFuture.java:78) ~[documents4j-util-conversion-1.0.2.jar:na]
  at java.util.concurrent.ThreadPoolExecutor.runWorker(ThreadPoolExecutor.java:1142) ~[na:1.8.0_74]
  at java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor.java:617) ~[na:1.8.0_74]
  at java.lang.Thread.run(Thread.java:745) ~[na:1.8.0_74]

这就是我进行文档转换的方式。

我正在实例化 LocalConverter 的实例

LocalConverter.builder().workerPool(corePoolSize, maximumPoolSize, keepAliveTime, TimeUnit.MINUTES).baseFolder(baseFolder).processTimeout(processTimeout, TimeUnit.SECONDS).build();

corePoolSize is 5
maximumPoolSize is 10
keepAliveTime is 3 minutes
processTimeout is 20 minutes

我正在使用这个实例

public File convertFile(MultipartFile file) throws ConversionException {
        try(InputStream docStream = file.getInputStream(); ByteArrayOutputStream pdfStream = new ByteArrayOutputStream()) {
            boolean status = iConverter.convert(docStream, false).as(DocumentType.DOC).to(pdfStream, false).as(DocumentType.PDF).execute();
            if(status) {
                // conversion is success, send the response
                File response = new File();
                //InputStream responseStream = new ByteArrayInputStream(pdfStream.toByteArray());
                response.setContentLength(pdfStream.size());
                //response.setInputStream(responseStream);
                response.setOutputStream(pdfStream);
                return response;
            } else {
                LOGGER.error("Failed to convert word to pdf, conversion status is {}", status);
                throw new ConversionException("failed to convert word to pdf");
            }
        } catch (FileSystemInteractionException fsie) {
            LOGGER.error("documents4j file system interaction exception", fsie);
            throw new ConversionException("File system exception", fsie);
        } catch(IOException ioe) {
            throw new ConversionException("Cannot read the input stream of file", ioe);
        }
    }

此多部分文件是 spring 多部分文件。我检查了documents4j用于转换的vb脚本，发现当wordDocument未正确关闭时会出现此错误。以下是 vb 脚本的片段，它是此错误的来源

' Close the source document.
wordDocument.Close WdDoNotSaveChanges
If Err <> 0 Then
    WScript.Quit -3
End If
On Error GoTo 0

我不确定为什么会收到 FileSystemInteractionException。

我能想到两个假设

1. 我同时发送多个请求，文件被其他线程删除
2. 我从 MultipartFile 对象获取输入流，并且多部分文件是临时文件，根据文档，用户负责将内容复制到持久存储。

spring official docs

如何解决此错误以及此错误的根本原因是什么。

Answer 1

导致此错误的原因可能有多种:

com.documents4j.throwables.FileSystemInteractionException: Could not access target file

Exception documentation here

您是否尝试将上传的多部分文件保存到临时文件，然后将此临时文件传递给转换器？我知道这是不必要的开销。但是，如果这有效，那么我们可以放心地假设当 IConverter 实例尝试访问输入“docstream”时，该输入并未完全填充，因此会出现错误。在这种情况下，您应该确保在尝试转换之前填充输入流，这应该可以解决您的问题。

如果即使在“基于文件”的转换场景中您也遇到此错误，请尝试以下步骤:

1. 确保 MS Office 应用程序未运行(因为您在外部打开了 Word 文档)
2. 确保物理机(而不仅仅是 JVM)上运行唯一一个 IConverter 实例
3. 如果您将 Tomcat 作为服务运行(我假设您将其部署在 Tomcat 上)，则您不是将 Tomcat 作为系统帐户服务运行，而是作为本地用户帐户运行。
4. 在 Web 应用程序中，您应该创建一次 IConverter 实例(就像在单例类中一样)，并且只要您的业务方法之一请求，它就应该返回相同的实例。另外，如果您预计会同时出现文档转换请求，请不要关闭转换器。

理想情况下，这些步骤之一应该可以解决您手头的问题，如果您仍然遇到此问题，请在评论中告诉我。