How to convert Multifilesystem to Singleton Mount and Vice Versa in DBFS
(Doc ID 1361627.1)
Last updated on JANUARY 24, 2020
Applies to:
Oracle Database - Enterprise Edition - Version 11.2.0.2 to 11.2.0.2 [Release 11.2]Information in this document applies to any platform.
Goal
DBFS has been designed to support :
(a) Per-user filesystems (i.e. each RDBMS user's DBFS namespace and storage is isolated from that of other RDBMS users)
(b) Each user can simultaneously create/use/drop multiple filesystems (where one filesystem's namespace and storage is
isolated from that of other filesystems---for any given RDBMS user).
Point (b) is critical in this context. This means that if RDBMS userA creates filesystems A, B, and C, and wants to
simultaneously access all of them, DBFS _must_ create a distinction between these filesystems using the only means at
its disposal, namely the _filesystem namespace_.
So the RDBMS-side (i.e. ContentAPI) pathnames for the root directory of the various filesystems are identified as
"/A", "/B", "/C", etc. (or something similar where the prefix is derived in an obvious and straight-forward manner from
the name of the underlying filesystem), and the FUSE/dbfs_client mountpoints are similarly identified as "/mnt/dbfs/A",
"/mnt/dbfs/B", "/mnt/dbfs/C" (assuming that the Linux mountpoint prefix is "/mnt/dbfs").
The mountpoint root itself is a "virtual" directory that cannot be used to store anything and attempts to write into this
directory will fail with errors such as "Permission denied" as seen below example.
Solution
To view full details, sign in with your My Oracle Support account. |
|
Don't have a My Oracle Support account? Click to get started! |
In this Document
Goal |
Solution |