1 |
On Sat, Sep 26, 2015 at 10:08 AM, James <wireless@×××××××××××.com> wrote: |
2 |
> Florian Gamböck <ml <at> floga.de> writes: |
3 |
>> Now, before I try some crazy stunts like bind-mounting $D and $ED on "preinst" |
4 |
>> and cleaning up in "postinst", I wanted to know if some of you guys did |
5 |
>> similar experiments and/or have some advice that you could share with me. |
6 |
> |
7 |
> I have been following 'bcache' as an interesting addition to complex |
8 |
> compiling scenarios. I'm not certain how it will help your 'wild ideas', |
9 |
> but it is worth a look, imho [1] |
10 |
|
11 |
Honestly, I'm hard-pressed to think of situations where you're better |
12 |
off with anything other than either tmpfs or whatever filesystem your |
13 |
/usr happens to be on. |
14 |
|
15 |
However, if you do want to build on something like bcachefs and you're |
16 |
using an ssd you might just want to use a pure log-based filesystem |
17 |
like f2fs. You can afford to go far more experimental for temp space. |
18 |
I haven't studied the design of bcachefs closely, so I'm not sure how |
19 |
it would compare for this use. I know it has made a bunch of noise |
20 |
lately, though honestly I'm not impressed with some of the hype (not |
21 |
something I blame the authors for - more the groupies). |
22 |
|
23 |
One thing that both zfs and bcachefs seem to be doing (not that I'm in |
24 |
an expert in either) is taking a multi-tiered approach to storage. |
25 |
That is, writes can go into a durable log on a separate device and |
26 |
then be consolidated onto more long-term storage. I suspect that this |
27 |
might help with some of the fragmentation issues that you run into |
28 |
with COW. Then again, I don't know how its performance would compare |
29 |
to just dumping everything right to a single tier like btrfs and then |
30 |
just periodically having some kind of smart defragmenter that just |
31 |
cleans up the stuff that really needs it. |
32 |
|
33 |
-- |
34 |
Rich |