Details
-
Improvement
-
Status: Reopened
-
Minor
-
Resolution: Unresolved
-
Nightly Builds
-
None
-
Operating System: All
Platform: All
-
38686
Description
It's really a bug but a explanation.
The default VFS cache is SoftRefFilesCache.
In my process, I have a FileManipulator which is simply a singleton. This
singleton create one (and only one max) instance of FileSystemManager using
VFS.getManager().
This process is called every 10mn to perform a copy from a FileObject to
another. In earch loop, I perform something which looks like :
FIleManipulator fileManipulator = FileManipulator.getInstance();
fileManipulator.copy("tgz:http://remote/archive.tar.gz!/dir/file", "/tmp/file");
Either the tgz:http://remote/archive.tar.gz!/dir/file has not change, in the
vfs_cache directory, I have :
tmp_19071_file
tmp_19076_file
So a copy if performed every loop and neither release which take a lot of space
on the filesystem.
How can I avoid this behavior ? Should I create a FileManipulator (and so a
FileSystemManager) and use the same in the loop ? Do I need to define my own
filesystemmanager with another FileCache system (such as NullFileCache) ? Is it
possible to define another VFS cache system for the default filesystem manager
got with VFS.getManager() ?
Many thanks for your help.