Strange import behaviour, 70%, 80%, 90%, back to 70%, and "oscillating" used space

[Cartesio]

I tried for the fist time today to import an external disk (NTFS formatted), in FreeNAS 11.3U1. The disk (a TOSHIBA MK5075GSX rebranded Verbatim in a USB 3.0 enclosure), was correctly identified and mounted, connected to a USB 3.0 port.

When I started the import task, at first everything looked OK, but then the task completion percentage started to move back and forth, reaching over 90% to get back to 70% and so on. The used space on the target dataset. as seen using df -h was "oscillating" too, sometime increasing sometime decreasing - and it looked larger than the occupied space as I can see from Windows (around 108GB, df shown about 170GB used in a empty dataset). Eventually, I stopped the task sending it a SIGTERM from a console, because I wasn't sure it would ever finished, and I can see no way to stop it from the GUI. The disk looks fine under Windows 10, and a chkdsk found nothing.

I see it uses rsync to import the disk, is there any rsync log anywhere? When I stopped the task it just reported the exitcode and little else. Is there any other log I should check for more info about the operation?

Anyone ever saw such a behaviour? Before posting a bug I'm trying to eliminate as much variables as possible (disks, cables, USB ports) but I don't have many disks to test with.

 

[Cartesio]

Don't mind... I found the rsync log, actually from the task pane pressing the "Download log" downloads a log different from what is shown in the dialog box, and it's the complete one. It looks to be some kind of rsync behaviour, and anyway rsync + FUSE is so slow to be almost unusable unless you have a lot of time to wait for the transfer to complete. That disk can easily reach 90MB/s under Windows, but it barely reaches 30 in FreeNAS. I see this is an effect of how rsync works, even for "local" copies. Maybe import should let users choose a "copy" method depending on their needs.