1 |
If you are so concerned with the awesomeness of XFS's caching... why not |
2 |
turn on data-journaling? Then data (not just meta-data) is committed to |
3 |
the journal. |
4 |
|
5 |
You can also tune XFS to not wait so long to hold cached data. |
6 |
|
7 |
Alan McKinnon wrote: |
8 |
> On Saturday 28 October 2006 16:41, b.n. wrote: |
9 |
> |
10 |
>> Dale ha scritto: |
11 |
>> |
12 |
>>> If you use XFS, make sure you have good power. XFS does not like |
13 |
>>> power failures at all. I have had to reinstall on a second rig |
14 |
>>> because of this very problem. If you have a UPS, that may be OK. |
15 |
>>> |
16 |
>> Thanks a lot for the advice. Power outages do happen and I don't have |
17 |
>> an UPS. Why does it happen? Isn't XFS journaled? |
18 |
>> |
19 |
> |
20 |
> Yes it is journaled but it also allows data to be very aggressively |
21 |
> cached. Make that VERY aggressively cached. With the result that data |
22 |
> can be held in a huge cache somewhere and the kernel can be convinced |
23 |
> it has been written to disk. |
24 |
> |
25 |
> Consider XFS's pedigree - SGI wrote it for their graphics machines. |
26 |
> These were big monsters backed up with high grade UPSs and such - the |
27 |
> logic was that if you spend a brazillion bucks on hardware, a mega UPS |
28 |
> is part of the deal, along with the wages to pay the army of admins you |
29 |
> also need. |
30 |
> |
31 |
> And, when doing video rendering, it turns out that it's easier to simply |
32 |
> re-render a frame when the filesystems does something odd with the data |
33 |
> rather than go to the effort of writing an FS that is 100% reliable. So |
34 |
> SGI sacrificed something that doesn't actually matter for their use |
35 |
> case to gain a significant performace increase (which does matter a |
36 |
> great deal) |
37 |
> |
38 |
> alan |
39 |
> |
40 |
|
41 |
-- |
42 |
gentoo-user@g.o mailing list |