Managing Nfs; Creating A File System On A Storage Array - Dell EMC VxBlock 240 Administration Manual

Managing NFS

Creating a file system on a storage array

Use this procedure to create a file system on a storage array.
Before you begin
Verify that there is sufficient space in the NAS pool to create the file system.
About this task
Use the following guidelines for file system slicing:
•
When using VP pool volumes (TDEVS/Thin LUNs), slicing is permitted, except with CLARiiON
(VNX, CX) VP pools, which are FAST VP enabled. Slicing is permitted on thick VP pool LUNs or
RAID group LUNs on CLARiiON (VNX, CX).
•
Be familiar with replication requirements that may restrict slicing.
TF/SRDF requires full volumes; slicing is not permitted.
—
Celerra Replicator replicates the file system, regardless of the underlying storage volume
—
configuration. Slicing is permitted.
When using traditional volumes (STDs/thick LUNs on RAID groups), slicing is permitted.
—
•
Do not thin a file system with a thin (TDEV or thin VP Pool LUN) backend volume. Only thin the
file system with a thick volume on the backend. A backend volume is a traditional thick LUN
(STD, fully allocated or thick VP Pool LUN, or RG volume).
•
Do not use auto-extend unless specifically requested by the customer.
The auto-extended file system displays the maximum size of the file system on the NFS
—
client.
Use of auto-extend on many file systems can decrease the performance of the file systems
—
due to the control station managing the file system extension rather than the storage array.
Use of auto-extend reduces the possibility of a file system full event, but requires more
—
monitoring and space management.
•
The recommended file system size is no larger than what can be backed up and restored while
maintaining SLAs. The maximum size is 16 TB.
•
Enable uncached write on the NFS file systems that are to be used as VMware datastores.
You can increase performance by caching writes in the X-Blade and not requiring
—
downstream acknowledgements from the array before acknowledging the write to the host.
These techniques create data integrity issues. The X-Blade is not designed to preserve
cached writes if an extended power outage occurs.
If the customer desires increased performance by not bypassing the X-Blade cache, make
—
sure that they understand the risks involved.
Managing NFS | 70