I have the following line of code in one of my models:
我的一个模型中有以下代码行：
open("|cd lib && /home/rails/jruby/bin/jruby -J-Djava.awt.headless=true -J-Xmx500m -J-Xms500m -J-Djruby.jit.logging=true jasper_order_invoice.rb #{order_id} #{db} #{user} #{pwd} > ../log/jasper.log")
jasper_order_invoice.rb is a script which calls the Java JasperReports library and creates an invoice PDF in tmp/. This used to work very well until I introduced Capistrano as deployment method. The app is now in /home/rails/rails-production/current instead of /home/rails/app. However, the PDF creation doesn't work anymore. No PDF is being created. The strange thing is: It works on my local dev machine and also on the server when I call it directly from the shell, paste the line into irb, or copy the line into an otherwise empty .rb script.
jasper_order_invoice.rb是一个调用Java JasperReports库并在tmp /中创建发票PDF的脚本。在我将Capistrano作为部署方法介绍之前，这曾经非常有效。该应用程序现在位于/ home / rails / rails-production / current而不是/ home / rails / app。但是，PDF创建不再起作用。没有创建PDF。奇怪的是：当我直接从shell调用它时，它可以在我的本地开发机器上运行，也可以在服务器上运行，将行粘贴到irb中，或者将行复制到一个空的.rb脚本中。
I can see that a java thread starts for about 2 seconds when Rails calls Jruby. It gives me enough time to check the parameters and env variables with ps -exa. Everything seems to be correct. The order_id, db, user and pwd are being passed and the envs are also the same as if I would call the code directly from irb. Paths are also correct.
我可以看到当Rails调用Jruby时，java线程启动大约2秒。它给了我足够的时间来检查ps -exa的参数和env变量。一切似乎都是正确的。正在传递order_id，db，user和pwd，并且envs也与我直接从irb调用代码相同。路径也是正确的。
The log/jasper.log file is unfortunately empty when the Code is being called by Rails. But strange to say it does contain output when I call it from irb.
遗憾的是，当Rails调用Code时，log / jasper.log文件是空的。但奇怪的是，当我从irb调用它时，它确实包含输出。
I have no idea how I could debug this any further. I would especially be very interested in seeing the stdout and stderr, but I don't know why the log file is empty.
我不知道如何进一步调试这个。我特别感兴趣的是看到stdout和stderr，但我不知道为什么日志文件是空的。
1 个解决方案
#1
0
The problem was a new environment variable RUBYOPT introduced by the Capistrano setup. This somehow confused jRuby. Adding export RUBYOPT= to the open call before executing jRuby helped.
问题是Capistrano设置引入的新环境变量RUBYOPT。这在某种程度上混淆了jRuby。在执行jRuby之前将导出RUBYOPT =添加到open调用帮助。
#1
0
The problem was a new environment variable RUBYOPT introduced by the Capistrano setup. This somehow confused jRuby. Adding export RUBYOPT= to the open call before executing jRuby helped.
问题是Capistrano设置引入的新环境变量RUBYOPT。这在某种程度上混淆了jRuby。在执行jRuby之前将导出RUBYOPT =添加到open调用帮助。