| 1 |
On Tue, May 19, 2009 at 15:52, Branko Badrljica <brankob@××××××××××.com> wrote: |
| 2 |
> Alex Alexander wrote: |
| 3 |
> |
| 4 |
>> most packages build fine with a 768M tmpfs :) |
| 5 |
>> if you plan on compiling big stuff like gcc you'll need to make it |
| 6 |
>> larger or unmount it though. |
| 7 |
>> |
| 8 |
>> |
| 9 |
> With ext4 useable, is it still practicall to fiddle with thmfs for |
| 10 |
> building ? |
| 11 |
> Ext4 can be configured so that it defers writing for significant amounts |
| 12 |
> of time, which means that it should perform closer to tmpfs performance, |
| 13 |
> without headaches, associated with it. |
| 14 |
> |
| 15 |
> Using aggressive settings might increase risk of data loss, but notebook |
| 16 |
> is by its nature battery backed and so prone to most of the issues... |
| 17 |
|
| 18 |
you would need a separate partition for /var/tmp/portage with special |
| 19 |
settings to be safe... even then, if you can afford the ram, tmpfs is |
| 20 |
probably a better solution. |
| 21 |
|
| 22 |
disk io is one of the worst bottlenecks in today's systems... |
| 23 |
no disk io > * =] |
| 24 |
|
| 25 |
-- |
| 26 |
Alex Alexander || wired |
| 27 |
Gentoo QT && KDE Herd Tester |
| 28 |
http://www.linuxized.com |
Archives