Is it possible to mount LUNs as NFS?

Question

linux

Hello!
Question out of idle curiosity: there are 3 servers (prod, stby and test)
There is a SAN on which 3 LUNs are cut and mapped to each of the servers.
The contents of the /dev/mapper directory on each of the servers:

# ls /dev/mapper/
control  mpathb  mpathc  mpathd
[[email protected]test disk]#

On one of the servers, we made file systems from these multipass disks and set up labels:

[[email protected] by-label]# ll
total 0
lrwxrwxrwx 1 root root 10 Mar  4 14:37 ARCH -> ../../dm-4
lrwxrwxrwx 1 root root 10 Mar  4 14:14 BOOT -> ../../sda1
lrwxrwxrwx 1 root root 10 Mar  4 14:35 PROD -> ../../dm-2
lrwxrwxrwx 1 root root 10 Mar  4 14:36 ROOT -> ../../sdl1
lrwxrwxrwx 1 root root 10 Mar  4 14:37 SWAP -> ../../sdj1
lrwxrwxrwx 1 root root 10 Mar  4 14:36 TEST -> ../../dm-5
[[email protected] by-label]#

As you can see from the block, these settings were made on a production machine. Question: Will the other 2 servers "see" these changes? Should they see them?
The idea to mount SAN LUNs as NFS

Answer 1

No, they won't. More precisely, they will see, but after a while there will be a mess of data. The thing is that the file system lives not only on the disk, but also in the memory of the computer that made the changes. 1) Deploy GPFS (IBM, paid) or OCFS (Oracle) - these are shared FS types (one storage is used, changes are reported to everyone)
You need a distributed file system to do this, where the computer that made data changes communicates them to its partners, and the partners are already updating the changed data.
Let's imagine a situation, A created a file in a directory, and B read this directory before that. B has cached the contents of the directory and will never know that a new file has appeared in it. In the case of distributed FS, A after creating the file tells B to reread the directory.
In your case, there are several options.
2) Deploy Gluster, luster, CephFS or something similar - distributed FS (separate storages are used, changes are communicated to everyone, data can be replicated).
3) If the data is not tied to files, use distributed storage / databases such as Ceph, S3, Mongo, etc.

Answer 2

No