1 |
On Wednesday 24 April 2002 03:46 pm, Miguel S. Filipe wrote: |
2 |
|
3 |
> I never understood why gentoo says this about reiserfs. |
4 |
> reiserfs is actively mantained in the kernel tree, it's allways getting |
5 |
> updated and bugs are fixed every release, this for me is a signal that |
6 |
> it's being very tested, and the fact that it's in the main kernel tree |
7 |
> for some time, and I never saw in kerneltrap.com, or in any other web |
8 |
> page related to the kernel, or the kernel mailing list. |
9 |
|
10 |
I can't speak for gentoo, but I have personally witnessed data loss with |
11 |
reiserfs on a number of occasions ... a friend of mine had the dubious |
12 |
privelege of having that experience just a couple of weeks ago. |
13 |
|
14 |
So while I cannot offer any insight into why the gentoo developers may have |
15 |
warned people away from reiser, I can tell you from personal experience that |
16 |
I'm glad they did, and only wish other distros had done the same. |
17 |
|
18 |
As a user of XFS I too am frustrated at the occasional hiccup in support |
19 |
(which will undoubtably change once it is in the 2.5.x kernel), although even |
20 |
so I have never experienced a single bit of data loss ... just the annoying |
21 |
himem bug during install which I work around by installing onto an ext2 |
22 |
partition and then copying over to a new xfs partition after I've booted a |
23 |
kernel with himem not compiled in. |
24 |
|
25 |
One can add performance patches by hand to an XFS capable kernel, or make a |
26 |
custom ebuild to do so. I've had good luck with the preemtable patch as well |
27 |
as the low latency patch (though I haven't gotten around to trying either |
28 |
against the current 2.4.19-r1 (xfs-sources) kernel yet). |
29 |
|
30 |
Jean. |