1 |
Am 17.09.2013 20:11, schrieb covici@××××××××××.com: |
2 |
> Volker Armin Hemmann <volkerarmin@××××××××××.com> wrote: |
3 |
> |
4 |
>> Am 17.09.2013 09:20, schrieb Grant: |
5 |
>>> I'm convinced I need 3-disk RAID1 so I can lose 2 drives and keep |
6 |
>>> running. I'd also like to stripe for performance, resulting in |
7 |
>>> RAID10. It sounds like most hardware controllers do not support |
8 |
>>> 6-disk RAID10 so ZFS looks very interesting. |
9 |
>>> |
10 |
>>> Can I operate ZFS RAID without a hardware RAID controller? |
11 |
>>> |
12 |
>>> >From a RAID perspective only, is ZFS a better choice than conventional |
13 |
>>> software RAID? |
14 |
>>> |
15 |
>>> ZFS seems to have many excellent features and I'd like to ease into |
16 |
>>> them slowly (like an old man into a nice warm bath). Does ZFS allow |
17 |
>>> you to set up additional features later (e.g. snapshots, encryption, |
18 |
>>> deduplication, compression) or is some forethought required when first |
19 |
>>> making the filesystem? |
20 |
>>> |
21 |
>>> It looks like there are comprehensive ZFS Gentoo docs |
22 |
>>> (http://wiki.gentoo.org/wiki/ZFS) but can anyone tell me from the real |
23 |
>>> world about how much extra difficulty/complexity is added to |
24 |
>>> installation and ongoing administration when choosing ZFS over ext4? |
25 |
>>> |
26 |
>>> Performance doesn't seem to be one of ZFS's strong points. Is it |
27 |
>>> considered suitable for a high-performance server? |
28 |
>>> |
29 |
>>> http://www.phoronix.com/scan.php?page=news_item&px=MTM1NTA |
30 |
>>> |
31 |
>>> Besides performance, are there any drawbacks to ZFS compared to ext4? |
32 |
>>> |
33 |
>> do yourself three favours: |
34 |
>> |
35 |
>> use ECC ram. Lots of it. 16GB DDR3 1600 ECC ram cost you less than 170€. |
36 |
>> And it is worth it. ZFS showed me just how many silent corruptions can |
37 |
>> happen on a 'stable' system. Errors never seen neither detected thanks |
38 |
>> to using 'standard' ram. |
39 |
>> |
40 |
>> turn off readahead. ZFS' own readahead and the kernel's clash - badly. |
41 |
>> Turn off kernel's readahead for a visible performance boon. |
42 |
>> |
43 |
>> use noop as io-scheduler. |
44 |
> How do you turnoff read ahead? |
45 |
> |
46 |
set it with blockdev to 8 (for example). Doesn't turn it off. Just makes |
47 |
it none-obstrusive. |