This is a tutorial on how to make fstrim run more often or on days that you wish. This could be handy if you have stuff continuously writing to the drives which then later gets deleted, which would cause it to create a large amount of unused data. For example, I have a Minecraft server running in Docker in an LXC container. It creates backups every 15 minutes and generates a lot of unused data at the end. Fstrim didn't run and left me with fully used space which interrupted my Minecraft server by locking it at read only. This was a pain and by doing this I was finally able to get back up without this issue!
Let's get started.
Here, first check if it doesn't contain
After that, you can modify the
You're probably editing using nano, so use Ctrl+O and then press Enter to save the edited file. Press Ctrl+X to exit. If you are somehow using another text editor then you'd have to make sure to save it the proper way and exit there.
Then, we restart the fstrim timer using:
As final step, we'd check if the schedule was properly set by checking the timer's status:
Check here if it runs successfully at all and importantly also check the
If this was done for a host only then you're done, congratulations!
We'd then as before find something like
Then we do the reloading of the systemd daemon once again:
We restart the fstrim service:
And finally, we check the status of the fstrim service to see if it runs properly:
If it now runs without any errors, then you're good to go and you're done!
I hope y'all enjoyed this little guide and it helps you further to customize how often fstrim runs and allowing it to run in containers and VMs automatically too.
Let's get started.
Step 1: Editing the Systemd Service Files of fstrim.timer (and fstrim.service if Container)
The first step would be to enter the shell on your guest or host and log into the machine using root (or add a sudo prefix before each command here if you'd like to). Run the following command to edit the fstrim.timer file:
Bash:
systemctl edit --full fstrim.timerHere, first check if it doesn't contain
ConditionVirtualization=!<THISCOULDBEANYTHINGLIKECONTAINERKVMORNO> when doing this for a container or host. If it does, you should comment it out, e.g. like this for a container:
Code:
#ConditionVirtualization=!containerAfter that, you can modify the
OnCalendar option to set the schedule as you wish. For example, we change it from weekly to run on Tuesday, Friday and Sunday (so thrice a week, you may change the description accordingly as well or generalize it):
Code:
OnCalendar=Tue,Fri,SunYou're probably editing using nano, so use Ctrl+O and then press Enter to save the edited file. Press Ctrl+X to exit. If you are somehow using another text editor then you'd have to make sure to save it the proper way and exit there.
Step 2: Reloading and Properly Loading the New Changes
The second step would be to do the final steps to reload things and check whether it runs the proper way according to the schedule. Still from the shell, we run the following command to reload systemd's daemon to make sure the changes are properly detected by systemd:
Bash:
systemctl daemon-reloadThen, we restart the fstrim timer using:
Bash:
systemctl restart fstrim.timerAs final step, we'd check if the schedule was properly set by checking the timer's status:
Bash:
systemctl status fstrim.timerCheck here if it runs successfully at all and importantly also check the
Trigger part. This will tell you when it will trigger the fstrim service for the next time. This should be according to your set up schedule. If it's not set properly or incorrectly put, then you should check for more information on it or ask me here!If this was done for a host only then you're done, congratulations!
Step 3 (Only for Containers and VMs): Making the fstrim Service Run
The final step for containers and VMs would be to allow the fstrim service to run. By default or in some cases, due to a condition it seems to be disabled (same one as above). So just like above, you should edit the fstrim file but this time not the timer but the actual fstrim service:
Bash:
systemctl edit --full fstrim.serviceWe'd then as before find something like
ConditionVirtualization=!<THISCOULDBEANYTHINGLIKECONTAINERKVMORNO> and then comment it out, e.g. for Proxmox VE VMs which use the KVM hypervisor:
Code:
#ConditionVirtualization=!kvmThen we do the reloading of the systemd daemon once again:
Bash:
systemctl daemon-reloadWe restart the fstrim service:
Bash:
systemctl restart fstrim.serviceAnd finally, we check the status of the fstrim service to see if it runs properly:
Bash:
systemctl status fstrim.serviceIf it now runs without any errors, then you're good to go and you're done!
I hope y'all enjoyed this little guide and it helps you further to customize how often fstrim runs and allowing it to run in containers and VMs automatically too.
