I suspect that will be somewhat dependent on how the drive manages wear 
leveling. Letting the firmware know certain areas of flash are no longer in use 
via TRIM doesn’t necessarily mean that the drive will immediately re-use (and 
thus wear out) that area.

IIUC, pretty much all writes to flash have their block addresses rewritten by 
the firmware as part of the wear leveling, so the particular LBAs being written 
to (as seen by the host) doesn’t really matter as much in terms of wear as much 
as the total number of writes. TRIM should still be beneficial since it should 
minimize the amount of read-erase-write cycles the flash has to do.


From: Bob Friesenhahn <[email protected]>
Date: Tuesday, April 27, 2021 at 1:13 PM
To: illumos-discuss <[email protected]>
Subject: Re: [discuss] zpool add -o autotrim=on doesn't work?
On Tue, 27 Apr 2021, Kent Watsen wrote:
>
>  - “tank" is a raidz2 made up of spinning rust.
>  - I want to add a ZIL/Slog comprised of SSDs (INTEL 3700s)
>  - Being SSDs, "autotrim=on” should be set (right?)

It seems to me that the ZIL is not like a filesystem (where data is
continually allocated and freed) and it should be expected to appear
to write to the same regions over and over.  The FLASH should do
wear-leveling as erase areas start to wear out due to the number of
writes.

Doing a trim of the ZIL might be counter-productive.  Doing a trim of
normal pool disks makes a lot more sense.

Bob
--
Bob Friesenhahn
[email protected], http://www.simplesystems.org/users/bfriesen/
GraphicsMagick Maintainer,    http://www.GraphicsMagick.org/
Public Key,     http://www.simplesystems.org/users/bfriesen/public-key.txt
------------------------------------------
illumos: illumos-discuss
Permalink: 
https://illumos.topicbox.com/groups/discuss/T6ef0a71646a80e63-Mcbfcf954daddbf865d28809e
Delivery options: https://illumos.topicbox.com/groups/discuss/subscription

Reply via email to