Well the little system I use has both NVME (sata type) and a sata port.
I have a larg accumulation of hardware sitting around which I’d rather use than collect dust.
So I would have a 128Gb NVME for / then use a much larger SSD/HDD for use with proxy/update accelerator.
It’s mount point would probably be deeper than just /var but I’d work out the default storage location.
This is a home setup and I don’t really need a RAID mirror.
AFAIK, there is M.2/B.2 connector (that supports NVMe and SATA respectively)…
So should not be NVMe and SATA at the same time, while some hardware have both slots.
There is an exception though. Using ext2 over ext4 can be beneficial in cases where the storage medium lacks effective wear leveling (like in SD cards). A non-journaled file system like ext2 can extend the lifespan of the medium by having less writing events per cycle, but sacrifices data reliability due to the absence of error correction.
it can be done after the install, by changing the fstab.
Procedure:
Find the UUID of the new disk by running as a root: blkid
Note down the UUID for the disk you want to mount as /var.
Backup the existing /etc/fstab file: cp /etc/fstab /etc/fstab.backup
Open /etc/fstab for editing: nano /etc/fstab
Add a new line for the new disk: UUID=<Your-UUID> /var ext4 defaults 0 0
Save and exit the editor.
Copy existing /var data to the new disk: rsync -a /var/ /path/to/new/disk/
Unmount the new disk, if it’s already mounted: umount /path/to/new/disk
Test the fstab entry: mount -a
If no errors, proceed to reboot the system: reboot
After reboot, /var should be on the new disk. Confirm with df -h.
Note: The old /var directory becomes hidden by the new mount point. Data in the old /var remains on the original disk but is not accessible unless the new disk is unmounted. You should leave the old files there.
If you are like me obsessive/compulsive and cannot sleep thinking about that unnecessary duplication, here is the procedure to delete the old /var (BE VERY CAREFUL, have at least a backup ready in case you screw this out!!!).
Ensure the new disk is mounted at /var, after reboot.
Open a terminal as root user.
Create a temporary directory: mkdir /mnt/temp_var
Mount the original root filesystem to the temporary directory: mount / /mnt/temp_var
Navigate to the old /var: cd /mnt/temp_var/var
Delete the old files: rm -rf * <— if you destroy your system, don’t blame me.
Change directory cd /
Unmount the original root filesystem: umount /mnt/temp_var
Remove the temporary directory: rmdir /mnt/temp_var
Thanks for that I’ll take a look when I get a chance.
Still I think it would be grate to have an expert option during install.
Thanks for all the other tips from people too.
Booting from SD cards has never ended well for me even after taking into consideration the quality of the card, the file system used and other tricks such as only make your partition the size you need. So e.g 32Gb SD card you might only have an 8Gb partition in use, the idea being that due to ware leveling data will still be written and read from the whole of the card. The theory being it will last longer.
Personally I just stick with HDD’s\SSD’s for systems other than install media of course. There’s also a healthy market for 2nd had enterprise grate storage, well mainly sata/sas’s.
I recently got a bunch of intel ssd’s for my TrueNAS to act as SLOG and Metadata caches. Had to buy an enterprise NVMe from new for ARC cache mind but the price was not too bad for a PCIe 4 2Tb.
Again thanks to all.
So far this community has been very helpful and friendly
for what it is worth - I think the general direction is to simplify things. There are few development volunteers and there is LOTS to do. So simplifying things helps keep the to-do list shorter.
Your instruction worked all most flawlessly.
I started the procedure on IPFire but became a little confused at step 7. Probably more my bad than the instructions but the disk was not mounted so i first tried to mount it in the web ui and failed. I know i could have used the mount command but I’m no terminal guru. More a copy n past man.
Instead i just booted the system with ubuntu, mounted the drives then ran the rsync command.
Re-booted and all has worked grate. So thanks again.
As for the the KIS approach i get that and it shows as IPFire does just work.
Respect to the developers for making it that way.
I think my limited knowledge probably means its not as simple a thing to do as I might have thought. Plus I suppose it creates another variable for things to go wrong when you have limited resources.
my bad, I should have listed step 9 before step 7. Congratulations on working around the problem. Did you delete the old files or did you go for the sane route of not risking the obliteration your IPFire?