Folders And Virtual Disks

Folders

A Syneto Storage Folder is a ZFS filesystem. They have a fixed mount point and they can be regarded as a traditional ‘directory’ from any other operating system.

1. Creating a folder

The image below illustrates how to add a folder:

On the left side you can explore a tree like view of the folders. After selecting a desired folder (the place where you will create your sub-folder) click the “New folder” button. The image below contains the options available at creation:

The previously selected folder appears as the parent folder.

Now choose a name for your folder and press Add.

2. Case sensitivity

Case sensitivity refers to the file names allowed to be written inside the folder.

The following types of case sensitivity may be selected:

Sensitive - Files that differ only in upper case/lower case are permitted (e.g. Readme, README and readme)
Insensitive - Files that have differences in letter case are considered identical (e.g. Readme is the same file as README)
Mixed - Folder behaves like a case insensitive folder, with the difference that it is possible to write multiple files with the same name from the case insensitive perspective. For example a directory may contain the files Readme and README, but when reading the file called READme the storage will always return one of Readme or README. While there is no guarantee on which of the two will be read, it will always be the same file on subsequent reads/writes.

3. Setting a maximum size (or quota)

When a folder or virtual disk is created its available size is equal to the size of its parent folder (the folder on top of which is created). The maximum allocation size for a folder can be modified to a user defined value. You can do this by inserting a number in the Quota field.

4. Reserving a specific size

If you want to secure a minimum space available for the folder you will insert a value in the Reservation field. What it does is pre-allocate a specific number of bytes for that folder.

5. Compression

ZFS compression is a feature that once enabled not only allows you to save space but also improves overall system performance. The reason this happens is that the time it takes to compress and decompress data is quicker than the time it takes to read and write the uncompressed data to disk. Your CPU may suffer a bit and we would recommend watching it when enabling compression on large folders or virtual disks.

Also, is always a good idea to enable compression on backup folders.

6. Deduplication

Enabling deduplication will configure the storage to save only one copy of data on the same disk pool, rather than storing redundant copies of identical data.

Deduplication can be enabled or disabled per folder or virtual disk, but once enabled it is a property of the entire disk pool. This means that when new data is written, the storage will look for duplicated data on the whole disk pool, not only on the folder or virtual disk that has deduplication enabled.

Deduplication is done real-time and for all written data to the folders or virtual disks on which it is enabled. This means that a ‘Data Deduplication Table’ (DDT) must be kept in the machine’s RAM. The size of this table is directly proportional with the amount of used data and block sizes of the pool. All the used data on the pool will be indexed in DDT. One block uses about 250 bytes of data in DDT. The block size of a pool is always automatic and variable. The system will use larger block sizes for big files and smaller blocks for tiny files. The size of the blocks can wary between 512 bytes and 128 KBytes. On a usual system the block size is between 45-65 KBytes.

As a rule of thumb, on average, the system will require 5GB of memory for every 1TB of data allocated for a good performance. Also, it will automatically know how to allocate RAM and SSD Cache for ARC and L2ARC, the place where DDT resides. So, with a little performance penalty, you can add SSD cache to your system and it will also be used for the DDT.

When deduplication is enabled, the maximum size of DDT is automatically guessed by the system. If your data requires more than this value, the system will check the blocks directly, by reading their checksums from the disk. This will lead to many reads on the system and will severely impact performance.

When you have a SSD cache, this is mainly used for read-cache. This offers faster sequential reads and improves performance. When you will use the same SSD for the DDT also, the available space for read-cache will be automatically adjusted. A very big DDT will reduce read-cache size to an almost zero, thus impacting general read performance on the system.

You can check the value of deduplication ratio on the Disk pools page.

For example in the screenshot above, if the deduplication ratio for the tank pool would be 2.25x, it would mean that the actual space usage is 2.25 times less that what would normally be required to store the same data without deduplication.

WARNING: on a big folder with deduplication turned on, when you delete a folder all the input/output is blocked until the folder gets deleted. For this to happen the system will search all the filesystem blocks, process which cannot be stopped even with a reboot. E.g. if you delete a 10k folder from 80 Terabytes your system may be blocked on deletion for one week.

7. Removing a folder

Go to the folder you wish to delete and click on it. The following display will appear:

Now push the X on the right of the folder name and confirm the directory you intend to delete. After confirming a new message will notify you of the deletion: “Dataset deleted.”

Setting folder permissions

Syneto Storage allows fine grained access to folder permissions.

When you click the “Access Control” tab you’ll be able to modify these permissions. Here you can also change the owner or the group of the folder and set its sticky bit.

The ACL inherit option is set on folder creation.

You can allow special privileges by adding ACL rules, they take precedence over the permissions. You can do this by pressing the “New ACL rule” button.

ACL rules can be deleted by hovering over the rule and pressing the X button.

Folder permissions and CIFS sharing

Please refer to the Active Directory document on how to set folder permissions to set CIFS sharing scenarios.

Virtual disks

A Syneto Storage virtual disk is created on top of a folder. Their functionality can be regarded as the functionality of a traditional ‘disk’. Besides that, virtual disks include also all the benefits of a folder (scrubbing, reduplication, etc …).

1. Creating a virtual disk

Virtual disks can be created on top of a parent folder. You cannot create a virtual disk on top of another virtual disk. To create a virtual disk select the parent folder and press the “New virtual disk” button.

Add a name for the virtual disk and choose the desired size. The size is mandatory and will be its visible size. The size occupied on disk depends on whether you choose the Thin or Thick provisioning format:

There are two possible ways to create a virtual disk:

Thin = No (aka Thick provisioning): where the entire size specified for the virtual disk is allocated on creation.

Thin = Yes (aka Thin provisioning): the virtual disk size is not allocated on creation, instead an ‘on-demand’ mechanism is used to allocated space only when it is needed. This method supports ‘over-allocation’ which means that a thin-provisioned virtual disk can be created with a size greater than the size of its parent filesystem. This allows future growth of a virtual disk.

For Compression and Deduplication please refer to their descriptions in Folders.

2. Removing a virtual disk

Go to the virtual disk you wish to delete and click on it. The following display will appear:

Now push the X on the right and confirm that you wish to delete the vdisk. After confirming a new message will notify you of the deletion.