More efficient file:// path handling?

Handle file:// path resources directly when reading, without using the temp path.

I might have missed something, because I don’t know if this option exists. I did check the parameters once again thoroughly and didn’t find what I were looking for. If this already exists great.

When running restore / test runs, which are incredibly slow, I noticed that disk I/O is being wasted in the most devastating way. All files are read from A, written to B and then read from B again. Fixing this would save at least 2/3 of unnecessary disk I/O. - Is there an option to avoid this? - Also when saving data, similar performance improvement would be nice. Yet the option use-move-for-put seems to mostly do the trick. When this option is enabled the data written to temp, isn’t copied to the destination directory, but is moved as file system reference to it. One trick to slightly work-a-round this problem is to use the temp with high performance SSD, but it’s utter resource waste. Yet the temp path need to be configured per operation, because if alternate disk is used for temp, then the move-for-put option won’t work.

Any thoughts?

Yet I do acknowledge the common attitude with many software developers, where all inefficient I/O code can be fixed by installing high performance SSDs. - Don’t take it too seriously, this statement isn’t limited to this project alone.

  • Thank you