-
Notifications
You must be signed in to change notification settings - Fork 1.8k
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Missing good documentation for TRIM on L2ARC cache devices #16644
Comments
Persistent L2ARC is enabled by default, unless you disable it via module parameter. But AFAIK it should have nothing to do with TRIM. Generally TRIM for L2ARC is a questionable concept, since L2ARC is written sequentially, and so in some opinions should not cause excessive flash fragmentation to solve with TRIM. But if you wish, according to zpoolprops(7): "TRIM on L2ARC devices is enabled by setting l2arc_trim_ahead > 0" and "autotrim=on". |
Thank you, but where can I read about this? Especially about autotrim=on in conjunction with cache devices. |
Is section for autotrim in |
To me it is not clear that the autotrim property includes log and cache devices. If that is the case a "zpool trim" command would also do the trick. But I have 4 HDD in the pool plus the SSD for cache. And when I initiate a zpool trim for that pool, I get the error that trim is not supported.
Why would autotrim for that pool than work? PS |
Both |
Thank you for your explanations. Very helpful for me. I will close this issue now. |
I just added an SSD as cache device to my main pool (4 HDD RAID10). But I did not find any good documentation about how the trim operation is performaned on SSD cache devices.
I read about persistent and non-persistent L2ARC. I read about parameters like l2arc_trim_ahead and l2arc_write_max and I also read PR #9789
But I still have no clue if and how trim is performed on my new cache SSD. I still do not understand what I need to configure. etc. etc. Should I go for a persistent l2arc? Why? How?
From my point view the easiest setup would be if a trim is performant on every import of a pool.
The text was updated successfully, but these errors were encountered: