| 1 |
Grant Edwards <grant.b.edwards@×××××.com> wrote: |
| 2 |
|
| 3 |
> On 2012-02-13, Michael Orlitzky <michael@××××××××.com> wrote: |
| 4 |
> > On 02/13/12 05:49, Helmut Jarausch wrote: |
| 5 |
> >> |
| 6 |
> >> I've written a small Python program which outputs the file names in |
| 7 |
> >> i-node order. If this is fed into tar or cpio nearly no seeks are |
| 8 |
> >> required during copying. |
| 9 |
> > |
| 10 |
> > What makes you think the inodes are sequential on-disk? |
| 11 |
> |
| 12 |
> Even if the i-nodes are sequential on-disk, there's no reason to think |
| 13 |
> that the data blocks associated with the inodes are in any particular |
| 14 |
> order with respect to the i-nodes themselves. |
| 15 |
|
| 16 |
Correct, there is however a really fast method using "star -copy". |
| 17 |
|
| 18 |
This works because there are two decoupled processes, shared memory between |
| 19 |
them and the fact that star reads names from directories in one big chunk. |
| 20 |
|
| 21 |
Jörg |
| 22 |
|
| 23 |
-- |
| 24 |
EMail:joerg@××××××××××××××××××××××××.de (home) Jörg Schilling D-13353 Berlin |
| 25 |
js@××××××××××××.de (uni) |
| 26 |
joerg.schilling@××××××××××××××××.de (work) Blog: http://schily.blogspot.com/ |
| 27 |
URL: http://cdrecord.berlios.de/private/ ftp://ftp.berlios.de/pub/schily |
Archives