1 |
On Thu, Dec 23, 2021 at 11:56 AM Mark Knecht <markknecht@×××××.com> wrote: |
2 |
> |
3 |
> Thanks for the responses. I'll post a single response here. I had |
4 |
> thought of the need to mirror the ZIL but didn't have enough physical |
5 |
> disk slots in the backup machine for the 2nd SSD. I do think this is a |
6 |
> critical point if I was to use the ZIL at all. |
7 |
|
8 |
Yeah, I wouldn't run ZIL non-mirrored, especially if your underlying |
9 |
storage is mirrored. The whole point of sync is to sacrifice |
10 |
performance for reliability, and if all it does is force the write to |
11 |
the one device in the array that isn't mirrored that isn't helping. |
12 |
Plus if you're doing a lot of syncs then that ZIL could have a lot of |
13 |
data on it. Plus it is an SSD that you're forcing a lot of writes |
14 |
through, so that is going to increase your risk of failure at some |
15 |
point. |
16 |
|
17 |
Nobody advocates for non-mirrored ZIL, at least if your array itself |
18 |
is mirrored. |
19 |
|
20 |
> Instead |
21 |
> of a ZIL in machine 1 the SSD becomes a ZLOG cache most likely holding |
22 |
> a cached copy of the currently active astrophotography projects. |
23 |
|
24 |
I think you're talking about L2ARC. I don't think "ZLOG" is a thing, |
25 |
and a log device in ZFS is just another name for ZIL (since that's |
26 |
what it is - a high performance data journal). |
27 |
|
28 |
L2ARC drives don't need to be mirrored and their failure is harmless. |
29 |
They generally only improve things, but of course they do nothing to |
30 |
improve write performance - just read performance. |
31 |
|
32 |
> As always I'm interested in your comments about what works or |
33 |
> doesn't work about this sort of setup. |
34 |
|
35 |
Ultimately it all comes down to your requirements and how you use |
36 |
stuff. What is the impact to you if you lose this real-time audio |
37 |
recording? If you will just have to record something over again but |
38 |
that isn't a big deal, then what you're doing sounds fine to me. If |
39 |
you are recording stuff that is mission-critical and can't be repeated |
40 |
and you're going to lose a lot of money or reputation if you lose a |
41 |
recording, then I'd have that recording machine be pretty reliable |
42 |
which means redundant everything (server grade hardware with fault |
43 |
tolerance and RAID/etc, or split the recording onto two redundant sets |
44 |
of cheap consumer hardware). |
45 |
|
46 |
I do something similar - all the storage I care about is on |
47 |
Linux/ZFS/lizardfs with redundancy and backup. I do process |
48 |
photos/video on a windows box on an NVMe, but that is almost never the |
49 |
only copy of my data. I might offload media to the windows box from |
50 |
my camera, but if I lose that then I still have the camera. I might |
51 |
do some processing on windows like generating thumbnails/etc on NVMe |
52 |
before I move it to network storage. In the end though it goes to zfs |
53 |
on linux and gets backed up and so on. If I need to process some |
54 |
videos I might copy data back to a windows NVMe for more performance |
55 |
if I don't want to directly spool stuff off the network, but my risks |
56 |
are pretty minimal if that goes down at any point. And this is just |
57 |
personal stuff - I care about it and don't want to lose it, but it |
58 |
isn't going to damage my career if I lose it. If I were dealing with |
59 |
data professionally it still wouldn't be a bad arrangement but I might |
60 |
invest in a few things differently. |
61 |
|
62 |
Just ask yourself what hardware needs to fail for you to lose |
63 |
something you care about at any moment of time. If you can tolerate |
64 |
the loss of just about any individual piece of hardware that's a |
65 |
pretty good first step for just about anything, and is really all you |
66 |
need for most consumer stuff. Backups are fine as long as they're |
67 |
recent enough and you don't mind redoing work. |
68 |
|
69 |
-- |
70 |
Rich |