File System Usage and Metering
This topic describes how usage and metering are calculated for your file systems, to help you understand and manage your service costs. This topic also describes different ways to see your file systemand snapshot utilization and the differences in reporting that can occur depending on which method you use.
File Storage service provisioning is fully managed and automatic as your utilization scales. For more information, see Space Allocation.
Here are the methods you can use to view your file system and snapshot usage:
- The File Storage service reports metered file system utilization and updates hourly. The metered file system utilization comes from the
meteredBytesvalue in the API, and represents the authoritative utilization value that is used to count your service cost. You can access the reported utilization for each of your file systems using the Console, the Command Line Interface (CLI), or the API. For more information, see the following section File System Metered Utilization.
- The File Storage service supports Network File System (NFS) protocol, so you can use the
ducommand from your instance command line tool to see usage for mounted file systems. However, the usage reported by
ducan differ from both the
meteredBytesvalue and the
dfvalue. For more information, see Using DF and DU Commands.
The File Storage service allocates space in blocks of variable size in a way that minimizes total customer cost and optimizes performance. Other storage systems might allocate blocks differently than Oracle Cloud Infrastructure File Storage. If you copy files from another storage device to your Oracle Cloud Infrastructure file system, you might see minor differences when you compare the physical file size before and after copying.
Metering and Service Cost
This section describes aspects of file system usage and how they affect your overall service costs.
File System Metered Utilization
The File Storage service reports metered utilization size for each file system. The metered utilization size is updated on an hourly cycle. You can see the metered Utilization size in the Console on the Details page of the file system. This value comes from the File Storage service API
meteredBytes property which is the total number of bytes consumed by the file system. The
meteredBytes value is updated asynchronously with respect to updates to the file system. Your usage charges are calculated based on the
You can also use the CLI or API to obtain this information. See Managing File Systems for instructions about how to view your file system utilization.
When you add or remove files from your file system, it can take the File Storage service up to one hour to report the change in metered size.
Snapshot Metered Utilization
A snapshot is a point-in-time view of your file system. Snapshots initially consume no additional usage in the file system, because they reference the original data instead of duplicating it, limiting usage cost. A snapshot doesn't change which blocks it references after it's taken.
Snapshot data usage is metered against differentiated data only. If nothing has changed within the file system since the last snapshot was taken, a new snapshot does not consume more storage. The metered size of snapshots is included in the reported
meteredBytes value of the file system it belongs to.
- Let's say you create a file system called "MyFileSystem" and add
"File1". The new file system now contains 1 GB including metadata. After
the hourly update cycle is complete, the total
meteredBytesshown by the File Storage service is 1 GB.
Next, you create a snapshot of "MyFileSystem" named "Snapshot1". After the hourly update cycle is complete, the total
meteredBytesshown by the File Storage service remains at 1 GB, because there's no differentiated data yet.
You then overwrite the first 0.5 GB of "File1". Now, "MyFileSystem" has a file that is different than the version previously captured in "Snapshot1". The
meteredBytesvalue is 1.5 GB, because the differentiated data between the live file system and the snapshot is 0.5 GB.
1 GB (snapshot) + 0.5 GB (differentiated data) = 1.5 GB
- If you then delete "File1". "MyFileSystem" now has a
meteredBytesvalue of 1 GB, which represents just the usage for "Snapshot1".
- Finally, delete "Snapshot1". Deleting the snapshot removes its references to the file data. Provided no other snapshots reference the file data, the space is relinquished and utilization returns to zero.
Metadata Metered Utilization
Files in the file system require space to be allocated for metadata.
512 bytes are required for each directory entry, and
8192 bytes are required for each symlink. Multiple hardlinks to a file create multiple directory entries for the file, and increases the metadata utilization. This utilization is included in the
meteredBytes value of the file system it belongs to.
Using DF and DU Commands
You can use
du commands from your instance command-line application to view usage information about your file system. To use these commands to view file system usage, the file system must first be mounted to the instance. See Mounting File Systems for instructions on mounting your file system.
How the Commands Work
dfprovides the amount of storage metered for your file system. Results are returned quickly, but can be up to 1 hour out of date.
duprovides the storage used by a directory hierarchy. The
ducommand walks the directory tree, and if your hierarchy is large, it can take a long time to run and return results.
How Results can Differ
DF and DU report snapshot utilization differently
A snapshot is a point-in-time view of a file system. Snapshots reference unchanged file system data instead of duplicating it. The file system blocks that the snapshot references don't count towards the snapshot utilization. Only differentiated data increases the snapshot utilization.
dfcommand retrieves information provided by the File Storage service using the NFS FSSTAT call. The NFS FSSTAT call accounts correctly for the way that snapshots reference file system data. Only utilization caused by differentiated data is reported.
ducommand descends the file system tree and uses each file's size attribute to total up the space used. When you create a snapshot, it copies the original size attribute for each file. So, if you run the
ducommand, the snapshot reports the file system's size at the time the snapshot was taken not necessarily the snapshot's actual current utilization.
Let's say you create file system called "MyFileSystem". You then add a 1 GB file called "FileA" to the file system. Here's how each command would report size.
For.. du reports...
FileA 1 GB 1 GB MyFileSystem 1 GB 1 GB
You then create "Snapshot1". The snapshot is placed in the
/.snapshotfolder of MyFileSystem. Here's how each command would report size:
For... du reports...
FileA 1 GB 1 GB snapshot1 1 GB 0 GB
2 GB 1 GB
dfreports 0 GB for Snapshot1 because the data hasn't changed yet, so there is no space allocated for differentiated data.
dureports 1 GB for Snapshot1 because it reports the copied file size attribute of FileA, which is 1 GB.
Charges are calculated using the
meteredBytesvalue. The utilization size reported by
ducan be much larger than
dfreports the same value as
meteredBytes, so you can use it to accurately view the file system size.
DF and DU count hard links differently
dfcounts each file only once.
dumay count files with hard links more than once.
DF and DU count symlinks and metadata differently
dfreports the utilization of bytes required by File Storage for metadata and symlinks, even on empty files.
dureports empty files as using zero bytes. It doesn't accurately report the bytes being used by File Storage for metadata and symlinks.