1 |
Am 2010-12-10 16:41, schrieb Volker Armin Hemmann: |
2 |
> On Friday 10 December 2010, Stefan G. Weichinger wrote: |
3 |
>> Am 09.12.2010 18:21, schrieb Volker Armin Hemmann: |
4 |
>>> Put /boot on raid1, / on raid6. Don't bother with lvm - it is |
5 |
>>> just another layer that can go wrong. |
6 |
>> |
7 |
>> You mean "don't use lvm for /" ? ... for other stuff it's very |
8 |
>> useful, isn't it? |
9 |
> |
10 |
> no, I mean don't use lvm. It is a very compley, easily broken layer |
11 |
> reducing data safety. |
12 |
|
13 |
ok .... hmmm |
14 |
|
15 |
> There is no need for lvm with bind mounting and ln -s. |
16 |
|
17 |
I don't fully get that yet ... |
18 |
|
19 |
>>> If Raid6 is like raid5 you should be able to have the kernel auto |
20 |
>>> assemble everything, so no initrd is needed. Pay attention to the |
21 |
>>> metadata format when creating the raid. Also have a look at |
22 |
>>> stripe sizes. And stripe_cache_size. |
23 |
>> |
24 |
>> Thanks for your reminders. What exactly do you think of with |
25 |
>> stripes and sizes? You point at the performance-impact? large files |
26 |
>> vs. small files etc? |
27 |
> |
28 |
> stripe size has nothing to do with big and small files. But choosing |
29 |
> the wrong stripe size can impact your performance very, very badly. |
30 |
> We are talking about abysmal performance, Challenger depht abysmal. |
31 |
> XFS and ext4 - for both is a lot of documentation available about |
32 |
> choosing the right stripe size. |
33 |
> |
34 |
> stripe_cache_size can be set in /sys and has a 4-5x performance |
35 |
> impact on raid5 (where I tried it). A good size for me is 8192. For |
36 |
> example. |
37 |
|
38 |
Ah, yes, already found that and I will check that out soon. Thanks! |