1 |
On Fri, Jul 30, 2021 at 11:50 PM Wols Lists <antlists@××××××××××××.uk> wrote: |
2 |
> |
3 |
> btw, you're scrubbing over USB? Are you running a raid over USB? Bad |
4 |
> things are likely to happen ... |
5 |
|
6 |
So, USB hosts vary in quality I'm sure, but I've been running USB3 |
7 |
drives on lizardfs for a while now with zero issues. |
8 |
|
9 |
At first I was shucking them and using LSI HBAs. That was a pain for |
10 |
a bunch of reasons, and I would have issues probably due to the HBAs |
11 |
being old or maybe cheap cable issues (and new SAS hardware carries a |
12 |
hefty price tag). |
13 |
|
14 |
Then I decided to just try running a drive on USB3 and it worked fine. |
15 |
This isn't for heavy use, but it basically performs identically to |
16 |
SATA. I did the math and for spinning disks you can get 2 drives per |
17 |
host before the data rate starts to become a concern. This is for a |
18 |
distributed filesystem and I'm just using gigabit ethernet, and the |
19 |
cluster is needed more for capacity than IOPS, so USB3 isn't the |
20 |
bottleneck anyway. |
21 |
|
22 |
I have yet to have a USB drive have any sort of issue, or drop a |
23 |
connection. And they're running on cheap Pi4s for the most part |
24 |
(which have two USB3 hosts). If for some reason a drive or host |
25 |
dropped the filesystem is redundant at the host level, and it also |
26 |
gracefully recovers data if a host shows back up, but I have yet to |
27 |
see that even happen due to a USB issue. I've had far more issues |
28 |
when I was trying to use LSI HBAs on RockPro64 SBCs (which have a PCIe |
29 |
slot - I had to also use a powered riser). |
30 |
|
31 |
Now, if you want to do something where you're going to be pulling |
32 |
closer to max bandwidth out of all your disks at once and you have |
33 |
more than a few disks and you have it on 10GbE or faster, then USB3 |
34 |
could be a bottleneck unless you have a lot of hosts (though even then |
35 |
adding USB3 hosts to the motherboard might not be any harder than |
36 |
adding SATA hosts). |
37 |
|
38 |
-- |
39 |
Rich |