| 1 |
On 03/05/2020 18:55, Caveman Al Toraboran wrote: |
| 2 |
> On Sunday, May 3, 2020 1:23 PM, Wols Lists <antlists@××××××××××××.uk> wrote: |
| 3 |
> |
| 4 |
>> For anything above raid 1, MAKE SURE your drives support SCT/ERC. For |
| 5 |
>> example, Seagate Barracudas are very popular desktop drives, but I guess |
| 6 |
>> maybe HALF of the emails asking for help recovering an array on the raid |
| 7 |
>> list involve them dying ... |
| 8 |
>> |
| 9 |
>> (I've got two :-( but my new system - when I get it running - has |
| 10 |
>> ironwolves instead.) |
| 11 |
> |
| 12 |
> that's very scary. |
| 13 |
> |
| 14 |
> just to double check: are those help emails about |
| 15 |
> linux's software RAID? or is it about hardware |
| 16 |
> RAIDs? |
| 17 |
|
| 18 |
They are about linux software raid. Hardware raid won't be any better. |
| 19 |
> |
| 20 |
> the reason i ask about software vs. hardware, is |
| 21 |
> because of this wiki article [1] which seems to |
| 22 |
> suggest that mdadm handles error recovery by |
| 23 |
> waiting for up to 30 seconds (set in |
| 24 |
> /sys/block/sd*/device/timeout) after which the |
| 25 |
> device is reset. |
| 26 |
|
| 27 |
Which if your drive does not support SCT/ERC then goes *badly* wrong. |
| 28 |
> |
| 29 |
> am i missing something? |
| 30 |
|
| 31 |
Yes ... |
| 32 |
|
| 33 |
> to me it seems that [1] |
| 34 |
> seems to suggest that linux software raid has a |
| 35 |
> reliable way to handle the issue? |
| 36 |
|
| 37 |
Well, if the paragraph below were true, it would. |
| 38 |
|
| 39 |
> since i guess all disks support resetting well? |
| 40 |
|
| 41 |
That's the point. THEY DON'T! That's why you need SCT/ERC ... |
| 42 |
> |
| 43 |
> [1] https://en.wikipedia.org/wiki/Error_recovery_control#Software_RAID |
| 44 |
> |
| 45 |
https://raid.wiki.kernel.org/index.php/Choosing_your_hardware,_and_what_is_a_device%3F#Desktop_and_Enterprise_drives |
| 46 |
|
| 47 |
https://raid.wiki.kernel.org/index.php/Timeout_Mismatch |
| 48 |
|
| 49 |
Cheers, |
| 50 |
Wol |
Archives