To optimize SSD performance on Linux, first enable TRIM by checking support with sudo hdparm -I /dev/sdX | grep TRIM, then ensure the weekly fstrim timer is active with systemctl status fstrim.timer, or enable it manually using sudo systemctl enable fstrim.timer, and optionally use manual trimming via sudo fstrim -av; second, optimize mount options in /etc/fstab by adding noatime and avoiding relatime to reduce writes, using defaults,noatime for periodic TRIM or adding discard only if real-time TRIM is desired; third, select the appropriate I/O scheduler by checking current settings via cat /sys/block/sdX/queue/scheduler and setting none for NVMe or kyber/mq-deadline for SATA SSDs, making it permanent by adding elevator=none to GRUB boot parameters and running sudo update-grub; fourth, reduce unnecessary writes by mounting /tmp and /var/log as tmpfs with entries in /etc/fstab, limiting journal size in /etc/systemd/journald.conf to SystemMaxUse=100M, and minimizing swap usage by either disabling it or enabling zram with sudo systemctl enable --now systemd-zram-setup@zram0; fifth, choose SSD-friendly filesystems like ext4 with mkfs.ext4 -E discard, XFS for high throughput, or Btrfs with compression while avoiding defragmentation; finally, monitor SSD health regularly using sudo smartctl -a /dev/sdX after installing smartmontools, checking attributes like Percentage Used for NVMe or Media_Wearout_Indicator for SATA, ensuring the drive remains healthy and performs optimally over time.

Solid State Drives (SSDs) offer faster read/write speeds and lower latency than traditional hard drives, but to get the most out of them on Linux, a few optimizations are essential. Without proper tuning, you may not fully leverage the performance benefits—or worse, risk reduced lifespan due to unnecessary wear. Here’s how to optimize SSD performance on a Linux system.

1. Enable TRIM for Garbage Collection

SSDs perform best when they know which blocks are no longer in use. The TRIM command allows the OS to inform the SSD which data blocks can be wiped internally, improving write performance and longevity.

Check if your SSD supports TRIM:

sudo hdparm -I /dev/sdX | grep TRIM

Enable periodic TRIM (weekly on most distros):
Most modern Linux distributions enable fstrim weekly via a systemd timer.

Check if it's active:

systemctl status fstrim.timer

Manually run TRIM:

sudo fstrim -av

Enable automatic TRIM (if not already):

sudo systemctl enable fstrim.timer

? Note: For continuous TRIM, you can mount with the discard option in /etc/fstab, but this can cause performance overhead. Periodic TRIM is generally preferred.

2. Optimize Mount Options in /etc/fstab

Tweaking filesystem mount options can reduce write amplification and improve performance.

Edit /etc/fstab and add these options for your SSD partitions:

defaults,noatime,discard

Or, if you're using periodic TRIM instead of real-time:

defaults,noatime

• noatime: Prevents the system from updating file access times on every read, reducing unnecessary writes.
• discard: Enables online TRIM (use with caution—better to rely on fstrim timer).

Example line:

UUID=abcd-1234 / ext4 defaults,noatime 0 1

?? Avoid relatime if you want maximum write reduction—noatime is better for SSDs unless you specifically need access time tracking.

3. Choose the Right I/O Scheduler

The I/O scheduler manages how read/write requests are sent to the storage device. SSDs don’t benefit from complex reordering like HDDs.

Common schedulers: none, mq-deadline, kyber, bfq.

For SSDs, especially NVMe, use:

• none (aka none for blk-mq devices) – Best for NVMe and high-performance SSDs.
• kyber or mq-deadline – Good for SATA SSDs with multi-queue support.

Check current scheduler:

cat /sys/block/sdX/queue/scheduler

Temporarily set scheduler (e.g., for nvme0n1):

echo none | sudo tee /sys/block/nvme0n1/queue/scheduler

Make it permanent:
Use kernel boot parameters via GRUB:

elevator=none

Add this to GRUB_CMDLINE_LINUX in /etc/default/grub, then run:

sudo update-grub

4. Reduce Unnecessary Writes

Minimizing writes extends SSD life and maintains performance.

Move temporary directories to RAM:

Use tmpfs for /tmp, /var/log, or browser caches.

Add to /etc/fstab:

tmpfs /tmp tmpfs defaults,noatime,nosuid,size=2G 0 0
tmpfs /var/log tmpfs defaults,noatime,nosuid,size=512M 0 0

?? Ensure adequate RAM and consider log persistence if needed.

Limit journal logging:

Reduce systemd journal size in /etc/systemd/journald.conf:

SystemMaxUse=100M

Avoid swap (or use zram):

If you have enough RAM, disable swap or use compressed RAM instead:

sudo systemctl disable --now dev-sda2.swap  # if using swap partition

Or enable zram:

sudo systemctl enable --now systemd-zram-setup@zram0

5. Filesystem Choice and Formatting

Use modern filesystems optimized for SSDs:

• ext4: Reliable, widely supported. Enable options:

  mkfs.ext4 -E discard /dev/sdX

• XFS: Great for large files and high throughput.
• Btrfs: Offers built-in compression (compress=zstd) and copy-on-write, but use with caution on critical systems.

Avoid excessive defragmentation—SSDs don’t need it. Filesystems like ext4 and XFS handle fragmentation well.

Bonus: Monitor SSD Health

Use smartctl to check health and wear:

sudo smartctl -a /dev/sdX

Look for:

• Percentage Used (NVMe)
• Wear_Leveling_Count or Media_Wearout_Indicator (SATA)

Install smartmontools if not present.

Optimizing SSD performance on Linux isn’t complicated, but it does require attention to mount options, TRIM, I/O scheduling, and minimizing writes. With these tweaks, your SSD will stay fast and last longer. Basically, just trim regularly, use noatime, pick the right scheduler, and avoid writing unnecessary data.

The above is the detailed content of Optimizing SSD Performance on a Linux System. For more information, please follow other related articles on the PHP Chinese website!