You are not logged in.
So, I've set up a new system on my laptop after purchasing a new SSD. Everything works wonderfully, but I'm wondering if the continuous trim is actually working...
The reason for me to be inclined to thinking that my filesystem is not actually continuously trimming is because when I do a fstrim / after a fresh boot I always get something like:
/: 429,4 GiB (461036777472 bytes) trimmed
My current setup is just 2 partitions a non-encrypted /boot and an encrypted /. This is my /etc/fstab:
# /dev/sda1
UUID=6BE1-2DC5 /boot vfat rw,noatime,utf8 0 2
# /dev/mapper/cryptroot
UUID=0d73fd89-bc3b-4759-80eb-135ddbb33022 / ext4 rw,noatime,discard 0 1
I'm also passing the correct arguments via the kernel parameters, as indicated by Flags in cryptsetup luksDump /dev/sda2:
LUKS header information
Version: 2
Epoch: 4
Metadata area: 12288 bytes
UUID: 81f38daa-9632-4817-b237-0de941cafff8
Label: (no label)
Subsystem: (no subsystem)
Flags: allow-discards
AFAIK there's no easy way for me to test if the continuous trim is working, because I cannot do the common procedure of writing to a file, running hdparm --fibmap [file], deleting the file and checking with hdparm --read-sector, as hdparm doesn't work with device-mappers.
Does anyone knows of an alternative way to check if my ext4 filesystem is actually doing the continuous trim? If it is, why is that I always get a positive result when running fstrim then?
Last edited by esauvisky (2018-03-27 20:24:04)
Offline
So, I've set up a new system on my laptop after purchasing a new SSD. Everything works wonderfully, but I'm wondering if the continuous trim is actually working...
The reason for me to be inclined to thinking that my filesystem is not actually continuously trimming is because when I do a fstrim / after a fresh boot I always get something like:
/: 429,4 GiB (461036777472 bytes) trimmed
That's supposed to do that AFAIK, ext4 does not save the information about what has been trimmed between reboots so it will always trim the free space.
I guess that if luks was not passing the trim request down to the disk you might get an error instead of the space that has been trimmed, I haven't checked though so it might just fail silently.
R00KIE
Tm90aGluZyB0byBzZWUgaGVyZSwgbW92ZSBhbG9uZy4K
Offline
You can test it: https://unix.stackexchange.com/a/85880/30851
There is no problem with fstrim, either: https://unix.stackexchange.com/a/371665/30851
On discard vs. fstrim vs. no trim at all: https://unix.stackexchange.com/a/218083/30851
Online
Wow, thanks a lot! I could confirm my continuous trim is working!
As as a bonus I'm also now aware of fstrim's output meaning (or lack thereof) and the differences between each method.
As always, Arch Linux and its awesome community making me learn more and more.
Offline