1 |
On 29/06/2021 14:56, Frank Steinmetzger wrote: |
2 |
> Hello fellows |
3 |
> |
4 |
> This is not really a Gentoo question, but at least my NAS (which this mail |
5 |
> is about) is running Gentoo. :) |
6 |
> |
7 |
> There are some people amongst this esteemed group that know their stuff |
8 |
> about storage and servers and things, so I thought I might try my luck here. |
9 |
> I’ve already looked on the Webs, but my question is a wee bit specific and I |
10 |
> wasn’t able to find the exact answer (yet). And I’m a bit hesitant to ask |
11 |
> this newbie-ish question in a ZFS expert forum. ;-) |
12 |
> |
13 |
> Prologue: |
14 |
> Due to how records are distributed across blocks in a parity-based ZFS vdev, |
15 |
> it is recommended to use 2^n data disks. Technically, it is perfectly fine |
16 |
> to deviate from it, but for performance reasons (mostly space efficiency) it |
17 |
> is not the recommended way. That’s because the (default) maximum record size |
18 |
> of 128 k itself is a power of 2 and thus can be distributed evenly on all |
19 |
> drives. At least that’s my understanding. Is that correct? |
20 |
> |
21 |
> So here’s the question: |
22 |
> If I had three data drives, (c|w)ould I get around that problem by setting a |
23 |
> record size that is divisible by 3, like 96 k, or even 3 M? |
24 |
> |
25 |
> |
26 |
> |
27 |
> Here’s the background of my question: |
28 |
> Said NAS is based on a Mini-ITX case which has only four drive slots (which |
29 |
> is the most common configuration for a case of this formfactor). I started |
30 |
> with two 6 TB drives, running in a mirror configuration. One year later |
31 |
> space was running out and I filled the remaining slots. To maximise |
32 |
> reliability, I went with RaidZ2. |
33 |
> |
34 |
> I reached 80 % usage (which is the recommended maximum for ZFS) and am |
35 |
> now evaluating my options for the coming years. |
36 |
> 1) Reduce use of space by re-encoding. My payload is mainly movies, among |
37 |
> which are 3 TB of DVDs which can be shrunk by at least ⅔ by re-encoding. |
38 |
> → this takes time and computing effort, but is a long-term goal anyway. |
39 |
> 2) Replace all drives with bigger ones. There are three counter arguments: |
40 |
> • 1000 € for four 10 TB drives (the biggest size available w/o helium) |
41 |
> • they are only available with 7200 rpm (more power, noise and heat) |
42 |
> • I am left with four perfectly fine 6 TB drives |
43 |
> 3) Go for 4+2 RaidZ2. This requires a bigger case (with new PSU due to |
44 |
> different form factor) and a SATA expansion card b/c the Mobo only has |
45 |
> six connectors (I need at least one more for the system drive), costing |
46 |
> 250 € plus drives. |
47 |
> 4) Convert to RaidZ1. Gain space of one drive at the cost of resilience. I |
48 |
> can live with the latter; the server only runs occasionally and not for |
49 |
> very long at a time. *** This option brings me to my question above, |
50 |
> because it is easy to achieve and costs no €€€. |
51 |
> |
52 |
5) Dunno if this is possible but ... replace one 6TB by a 12TB (any |
53 |
reason you don't like Helium?) and raid-0 two of the remaining 6's |
54 |
together. Dunno anything about what the raidZ's are but I presume this |
55 |
would give you 12TB of mirrored storage. It would also only use 3 slots, |
56 |
so you could use the 4th for eg your videos, and back them up on |
57 |
external storage ie the drive you've just removed :-) |
58 |
|
59 |
(The raid-0, I'd probably stripe rather than linear for performance.) |
60 |
|
61 |
Cheers, |
62 |
Wol |