| 1 |
Hello, |
| 2 |
|
| 3 |
Le Lundi 06 Février 2006 19:50, Duncan a écrit : |
| 4 |
> Fragmentation doesn't tend to be as much of an issue on Linux, with "real" |
| 5 |
> filesystems, as on MSWormOS, particularly FAT/FAT32. I'm running all |
| 6 |
> reiserfs here, FWIW. It doesn't have a compaction tool (defrag, on |
| 7 |
> MSWormOS), but I've not noticed any issues as a result. |
| 8 |
|
| 9 |
Fragmentation seems to be a myth for anyone on Linux, and I was enclined to |
| 10 |
believe that myth until I started to use Gentoo. |
| 11 |
|
| 12 |
At first, a brand new gentoo system is fast, but after a few months and a |
| 13 |
dozen emerge -uDN world, things tend to slow down to a point that is barely |
| 14 |
acceptable. In fact, the first time I tought that maybe I installed too many |
| 15 |
things, and that my system was crippled with cruft. |
| 16 |
But then I had to repartition my hard drive, so I made a backup (tar zcvpf) of |
| 17 |
my different partitions, fdisk, mkfs, and tar zxvpf. |
| 18 |
The system was exactly the same as before, just the partition size had |
| 19 |
changed. |
| 20 |
But then emerge -S was much faster than before the operation, as well as |
| 21 |
common portage operations. |
| 22 |
|
| 23 |
Since then, I've tried to do the same on several servers, without the fdisk |
| 24 |
operation, just tar cp, mkfs, tar xp, and I've always noticed an appreciable |
| 25 |
speedup. |
| 26 |
|
| 27 |
The only explanation that comes from this experiment is fragmentation. |
| 28 |
And I think Gentoo is more sensible to fragmentation than binary distributions |
| 29 |
because it has to deal with many small files, often changing, during |
| 30 |
compilation and rsynchronisation. |
| 31 |
|
| 32 |
So the directories sensible to fragmentation are IMHO, /var/tmp |
| 33 |
and /usr/portage, and they are the ones to put on different partitions. |
| 34 |
|
| 35 |
Now, I don't have exact numbers to prove my sayings, but anyone can make the |
| 36 |
test themself, if they already have /var/tmp and/or /usr/portage on separate |
| 37 |
partitions. |
| 38 |
|
| 39 |
I didn't have time yet to sort out what kind of filesystem is more or less |
| 40 |
sensible to fragmentation, but from my experience, ext[23] is not a good |
| 41 |
candidate for /var/tmp or /usr/postage. Reiser3 has proven to fragment too, |
| 42 |
and one of the last system I installed was formated with XFS, which I will |
| 43 |
"defragment" in a few weeks. Hopefully I could then come with numbers. |
| 44 |
|
| 45 |
BTW, does someone know of a tools to show the fragmentation level of a *nix |
| 46 |
filesystem ? |
| 47 |
|
| 48 |
David |
| 49 |
|
| 50 |
|
| 51 |
|
| 52 |
|
| 53 |
-- |
| 54 |
gentoo-amd64@g.o mailing list |
Archives