1 |
Alan McKinnon wrote: |
2 |
> On Tuesday 31 October 2006 11:04, Uwe Thiem wrote: |
3 |
> |
4 |
>> On 31 October 2006 09:17, Alan McKinnon wrote: |
5 |
>> |
6 |
>>> I find it useful to keep in mind that XFS is a file-system (i.e. a |
7 |
>>> system for files), and not necessarily a severly disk-bound |
8 |
>>> filesystem |
9 |
>>> |
10 |
>> Would you mind to elaborate on this? I simply do not get your point. |
11 |
>> |
12 |
> |
13 |
> Historically SGI was very strong in graphics, and those applicatiosn |
14 |
> tended to generate massive amounts of temporary files that had a short |
15 |
> life and only the final version needs to be written to persistent |
16 |
> storage, very well suited to aggressive caching and other similar |
17 |
> speedups. |
18 |
> |
19 |
> SGI's engineers could get away with this because they could guarantee |
20 |
> that power loss to the machine wouldn't happen, so the potential data |
21 |
> loss on a power outage didn't happen either. This sounds a bit weird to |
22 |
> those of us raised on Intel where we pay close attention to getting |
23 |
> everything on disk ASAP with as little performance loss as possible, |
24 |
> but it's a perfectly reasonable system for an engineer to implement on |
25 |
> the kind of hardware SGI were building. |
26 |
> |
27 |
> That's why I say XFS is designed to not be tightly bound to the physical |
28 |
> disk if the admin chooses to set it up that way, and the file system |
29 |
> becomes more of a collection of directories and files that might never |
30 |
> even be stored on a disk at all |
31 |
> |
32 |
> alan |
33 |
> |
34 |
|
35 |
"..we pay close attention to getting everything on disk ASAP with as little performance loss as possible.." |
36 |
|
37 |
Then I would propose you to use "hdparm -W0 /dev/(what-ever)" to disable the write caching (no matter which FS you use). Nothing can give 100% guarantee against power failure. |
38 |
|
39 |
|
40 |
-- |
41 |
Best regards, |
42 |
Daniel |
43 |
|
44 |
|
45 |
-- |
46 |
gentoo-user@g.o mailing list |