On Mon, 2010-03-08 at 15:53 -0300, Luigi Castro Cardeles wrote:
2010/3/8 Patrick O'Callaghan <pocallaghan(a)gmail.com>
> [Please don't top-post on this list]
>
> First of all, is this device a real hard disk, i.e. a mechanically
> rotating magnetic surface, or is it some kind of Flash drive?
Hi,
sorry for top post and sorry about the poor english.
It is not a flash drive, is an real HD.
I search a lot on www about this issue and the only clue i find is
about sync/async option when mount the device and verify if it is
connected to usb 2.0 hub.
The strange thing is if i use windows to copy this diretory (drag on
drop), the full operation last more or less two six (copy directory
and umount - "eject" - the device),
The same operation under fedora last a lot longer. (last day, i let
the copy last for a full day and it is not finished - i pause in the
middle).
It would be interesting to see what happens with a different filesystem,
e.g. ext3 or ext4, that doesn't use fuse. If that's not practical
because the disk contains real data, perhaps you could create a small
partition with one of the kernel-supported filesystems and measure that.
If that shows a big difference, check if fuse has enough working memory,
or if there's any way to instrument it (I know very little about fuse so
I can't help there). Or even try it with a vm-based Windows installation
(I know it sounds wierd, but it could be interesting to try).
poc