1 |
Tom schrieb: |
2 |
> Thanks for your answer! |
3 |
> |
4 |
>> Last time I checked, ext2 didn't work with Truecrypt on Windows due to |
5 |
>> a bug. If you use another solution (or the problem is fixed), I'd |
6 |
>> recommend ext3 or ext4 without extents (so it can still be mounted as |
7 |
>> ext2 by the Windows driver). |
8 |
>> I would use NTFS. I dislike using non-journalling filesystems like FAT |
9 |
>> or ext2 on such big disks. However, using the fuse implementation |
10 |
>> under Linux causes a rather high CPU utilization. Together with the |
11 |
>> encryption it could slow down less beefy systems. |
12 |
> |
13 |
> Fat is really out of the question, I just listed it for completeness. |
14 |
> Regarding NTFS, the performance overhead was exactly the reason I was |
15 |
> thinking of rather sticking to a ext based filesystem. |
16 |
|
17 |
I would simply test it. Most likely, USB will be the bottle neck, not |
18 |
CPU time. |
19 |
|
20 |
> I've considered only writing to the disk from within a |
21 |
> linux-environment and only mounting it readonly (as ext2) from within |
22 |
> windows, but as you mentioned, using a non-journalling fileystem is an |
23 |
> obvious risk. However, does this still apply when its in readonly-mode? |
24 |
|
25 |
No write-action, no need for a journal. |
26 |
|
27 |
By the way: On my external hard disk I have made two partitions: 20GB |
28 |
NTFS for exchanging data with Windows hosts and the rest (230GB) |
29 |
encrypted (LUKS) ext3. It works great, but only because I seldom need |
30 |
the NTFS-partition. |