7

Back before Solaris SunOS had a cool feature called transparent file system. The basic concept allowed disk files systems to be stacked like transparencies on an overhead projector. And allowed the presenter to draw on the top transparency.

What is a modern equivalent of the transparent filesystem? I want/need to stack multiple drive filesystem.

poige
  • 9,171
  • 2
  • 24
  • 50
peter cooke
  • 213
  • 1
  • 8
  • 2
    I have experimented with TFS once and thought cool but would never use it – peter cooke Sep 27 '19 at 15:51
  • 3
    It looks like you have already answered your own question! But it was called "translucent file system". – Michael Hampton Sep 27 '19 at 20:02
  • 3
    @MichaelHampton from your description it sound like it predates (as there is a paper called ["The Translucent File Service"](http://mcvoy.com/lm/papers/SunOS.tfs.pdf) from 1988) the concepts that ["union mount"](https://en.wikipedia.org/wiki/Union_mount) filesystem on other OSes mimic today... – Anon Sep 29 '19 at 09:31
  • @Anon Yes, other OSes have similar technologies, but Sun dropped this and as others have already answered, never replaced it with anything that serves the same purpose, not even after being sold to Oracle. – Michael Hampton Sep 29 '19 at 21:07
  • thanks for the union mount – peter cooke Sep 30 '19 at 04:58

4 Answers4

7

This is deprecated feature. Unfortunately.

RiGiD5
  • 837
  • 1
  • 6
  • 10
2

On Solaris I think you're out of luck, but Linux has overlayFS, AuFS and some others.

wazoox
  • 6,782
  • 4
  • 30
  • 62
1

It's none on Solaris or its derivatives: "… The translucent file system (TFS) type has been withdrawn from the SunOS release 5.7 software …"

In despite in this quote it's referred as "file system" actually it seems used to be a service running in user-space.

poige
  • 9,171
  • 2
  • 24
  • 50
0

You should be able to replicate most of the behavior of such "transparent" filesystems by creating ZFS clones and making appropriate changes to the clone(s).

See Overview of ZFS Clones:

A clone is a writable volume or file system whose initial contents are the same as the dataset from which it was created. As with snapshots, creating a clone is nearly instantaneous and initially consumes no additional disk space. In addition, you can snapshot a clone.

Clones can only be created from a snapshot. When a snapshot is cloned, an implicit dependency is created between the clone and snapshot. Even though the clone is created somewhere else in the dataset hierarchy, the original snapshot cannot be destroyed as long as the clone exists.

Depending on the scope of the changes, there might be a substantial time needed to generate the new image.

Note also that changes to the original file system will not be reflected in the cloned filesystem(s), so that might be a problem for your needs. Or it might be a useful feature.

One use for such a system is using a base file system as, for example, the root file system image of a farm of diskless servers. Each server would use its clone of the base file system, and the base file system image can be updated without impacting running servers, and new clones created from the updated image that the servers can be rebooted to.

Andrew Henle
  • 1,232
  • 9
  • 11