[Mldonkey-users] Problems with slow disks

[Robert Reither]

I'm using mldonkey 4 a long time and i'm quite satisfied with it, many thanks 
to the coder(s).
Currently using the self-compiled cvs version from October 02.

I've just one big problem :

I'm downloading on a disk connected via usb1.1 and even worse it's mounted with 
the
'sync' option (linux 2.4 system). i canot change this, i have no root rights on 
the
machine. So i only have about 200 kB/s rate over the USB line (but 100MBit 
network rate ;)).

The problem is i guess the verify_chunk stuff at startup. This is killing my 
download
rate ( only just  1kb/s) for about 1 hour, (a few GB to rehash) because it 
consumes the
full traffic rate over the usb.

I changed the compute_md4_delay to higher values (loading one 9 MB chunk from 
the disk
takes about 40 secs), so i used a delay rate of 100, which leaves me about 1 
minute for
download and the rest for loading the chunk.

But i encountered problems :

It looks like mldonkey cannot load fresh chunks for a file while the file_md4 
isn't fully
recalced (the status changes from downloading to paused), so all my tricks with
the md4_delay doesn't really work.

So my questions : Is it possible to bring mldonkey to download fresh chunks 
while still
                  rehashing ?

                  Or maybe an option to split the rehash-chunk-reading into 
smaller parts,
                  e.g. don't read the full chunk in a piece, use smaller 
sub-chunk-readings
                  maybe this can help to get a balance between downloading and 
getting
                  chunk data for rehash accessing the disk.


Btw i noticed that the rehash status display of a file in the telnet interface 
disappered
in the 1.99 versions (was in in 1.15 as far as i can remember (the !!__? 
stuff)).
The source-client info too.
Why is this gone ? It was always a big help 4 me, to see how long it will still 
take to
complete the rehash.

Robert

############################################################################
Robert Reither                8925573 E754 (ja wirklich schon !!)
TU-Vienna
AUSTRIA
############################################################################