1 |
Nikos Chantziaras wrote: |
2 |
> Dale wrote: |
3 |
>> Nikos Chantziaras wrote: |
4 |
>>> OK, I once again verified that fragmentation seems to be a big issue |
5 |
>>> even on Linux. I just migrated to ext4, and in order to do that I had |
6 |
>>> to rsync, format and rsync back. The result is similar to the last |
7 |
>>> time I did this (over 8 months ago): |
8 |
>>> |
9 |
>>> emerge --sync takes 15 seconds (at least 3 minutes yesterday) |
10 |
>>> update-eix takes 2 seconds (20 seconds yesterday) |
11 |
>>> |
12 |
>>> And I don't believe it's due to ext4. It's a nice speed-up from ext3, |
13 |
>>> but not THAT nice. |
14 |
>> |
15 |
>> Well, try as I may, I could not get mine past 10% on resiserfs. |
16 |
>> Fragmentation happens on any file system but I think the point is that |
17 |
>> Linux doesn't get as bad as the windoze file system. 10% or so is not |
18 |
>> to bad depending on the size of the files. Files that are large will |
19 |
>> have to be fragmented no matter what file system you use. |
20 |
>> I posted in another the reply right after a copy to another drive. I |
21 |
>> think that was before I even booted into the OS and was still on the |
22 |
>> CD. It is around 2% or so. I doubt given that condition that it could |
23 |
>> get any better. |
24 |
> |
25 |
> I think the main problem may not be so much fragmentation of files, |
26 |
> but rather their position on disk. Even if files are not fragmented, |
27 |
> if they are located too far from each other even though they're |
28 |
> related (same directory for example) or there's simply too much empty |
29 |
> space between files (I think this is intentional in order to reduce |
30 |
> fragmentation) then seek times get really bad. After I rsync the data |
31 |
> back, it's nicely and sequentially laid out on disk. I guess over |
32 |
> time it starts to get further apart again (to combat fragmentation) |
33 |
> and emerge --sync goes up from 15 seconds to 2 minutes again. Even |
34 |
> though the files aren't fragmented at all. |
35 |
> |
36 |
> Some defrag apps for Windoze actually offer to put the files back |
37 |
> closer together without trying to defragment at all. I guess this is |
38 |
> why :P |
39 |
> |
40 |
> |
41 |
> |
42 |
|
43 |
|
44 |
Well, this is what I got on my rig. Sort of interesting in a way. |
45 |
|
46 |
root@smoker / # mount |
47 |
/dev/hda6 on / type reiserfs (rw) |
48 |
/proc on /proc type proc (rw) |
49 |
sysfs on /sys type sysfs (rw,nosuid,nodev,noexec) |
50 |
udev on /dev type tmpfs (rw,nosuid) |
51 |
devpts on /dev/pts type devpts (rw,nosuid,noexec) |
52 |
/dev/hda1 on /boot type ext2 (rw) |
53 |
/dev/hda7 on /home type reiserfs (rw) |
54 |
/dev/hda8 on /usr/portage type ext2 (rw) |
55 |
/dev/hda9 on /data type reiserfs (rw) |
56 |
none on /dev/shm type tmpfs (rw) |
57 |
usbfs on /proc/bus/usb type usbfs (rw,noexec,nosuid,devmode=0664,devgid=85) |
58 |
binfmt_misc on /proc/sys/fs/binfmt_misc type binfmt_misc |
59 |
(rw,noexec,nosuid,nodev) |
60 |
/dev/hdd on /media/hdd type udf |
61 |
(rw,noexec,nosuid,nodev,uid=0,gid=0,umask=007) |
62 |
root@smoker / # |
63 |
|
64 |
root@smoker / # /root/fragck.pl / |
65 |
3.10978840211776% non contiguous files, 1.08156705459019 average fragments. |
66 |
root@smoker / # /root/fragck.pl /usr/portage/ |
67 |
0.0276657266269232% non contiguous files, 1.00029450612216 average |
68 |
fragments. |
69 |
root@smoker / # /root/fragck.pl /boot/ |
70 |
6.25% non contiguous files, 1.0625 average fragments. |
71 |
root@smoker / # /root/fragck.pl /home/ |
72 |
3.2440588457186% non contiguous files, 1.16408902301018 average fragments. |
73 |
root@smoker / # /root/fragck.pl /data/ |
74 |
5.56267766568196% non contiguous files, 1.06797837355777 average fragments. |
75 |
root@smoker / # |
76 |
|
77 |
Now keep in mind that the first one includes all the others. I'm logged |
78 |
into a GUI so I can't umount /home at least. May do that in single mode |
79 |
someday. |
80 |
|
81 |
I think sometimes the files are just to big to fit on one section. I |
82 |
know I have some files that are pretty big. I got a couple videos that |
83 |
are big that came off my camera and one video that is a hour or so |
84 |
long. I think there are a lot of variables that without a microscope we |
85 |
can never see and know for sure. |
86 |
|
87 |
Dale |
88 |
|
89 |
:-) :-) |