Is it possible to install Linux on different drives?

Question

linux

Look, I have in my computer:
2 ssds for 120gb
1 hard drive for 500gb

Now I installed linux on a separate ssd partitioned like this:
/ - 75gb
swap - 8gb
/ home 35gb

+ - something like this.

I have composer, node_modules and docker just gobbled up 50GB of space on the root when I deployed projects + more programs are put there too.
Why the hell then this home is needed.
And I thought, maybe there is an opportunity to install the entire system on a separate ssd, except for the home folder.

That is, the breakdown will be something like this:
sda1 - 120gb /
sdb1 - 50gb / home

On the ssd all Linux, on the hard partition / home

Everywhere in the internet they write that allocate 15-30 GB to the system partition.
But I don’t have enough, I had it for an hour while I was unfolding projects. What I installed: chrome, tg, phpstorm, nginx, docker, npm, nodejs, composer and deployed 1 project which weighs in total +- 400mb

Answer 1

It would be nice to understand where, what and in what volumes will be stored before partitioning the disk.
If you don't understand, make 2 partitions - root / and swap.

Answer 2

I have composer, node_modules and docker just gobbled up 50gb root space when I deployed the projects

Apparently, the projects are for some reason deployed to /var/www - which is not the slightest need. I generally have them on a terabyte HDD mounted in / media / section.
Programs don't eat that much, and it's completely optional to put them in /usr or /opt, especially JAVA programs like PhpStorm, which generally don't give a damn where to be located (for example, I have them in ~/JAVA).
The root partition, if you don’t throw everything into it, but at least sort it out a little, 75 gigabytes - for the eyes and ears.
If it becomes crowded, check if /var/log is not polluted with logs and if the bases are swollen to the point of indecentness (their files can also be moved from the default folder to an arbitrary location, elementary).

Answer 3

there is no need to scatter the system over different disks (the only thing for which it would make sense is the convenience of reinstalling and transferring user files between installations, i.e. the windows way), even a swap partition is not necessary, since it can be placed in file. Even /boot is not required if a standard file system is used without non-standard llvm/raid/bcache/... configs, this is only relevant for mbr, gpt will still need a 1mb biosboot partition if the boot type is legacy mbr, and regardless of the partition type , if efi is used, then a fat efiboot partition is needed)
You place the root on ssd, and mount hdd to some technical directory (I usually put something like /o, /g or /i in the root, that is, the shortest possible name) and then, if necessary, transfer large directories using symlinks to a subdirectory inside hdd, i.e. manual file placement management
For example, npm or gradle package file caches are at least tens of gigabytes and do not require quick access
ps install btrfs, it turns on zstd compression, it works so fast (experiment with the compression level) and it's good that it's shocking, plus docker has support for btrfs snapshots, which speeds up the work even more

Answer 4

You can mount a drive on /home in /etc/fstab. mount will be done at system startup. Something like

/dev/sdb1    /home/yourname    ext4    defaults    0    1

Answer 5

Abstract from the concept of disks and use modern methods of working with the data constructor, lvm, zfs/btrfs pools.