nav[aria-label="Primary Navigation"] { padding: 0; & ul { list-style: none; width: 100%; display: flex; flex-direction: row; justify-content: start; align-items: start; gap: 30px; padding: 0; & li { margin: 0; } & ul li { list-style: none; } } }

Understanding NAS with Flexpod

Sir-Loin-of-Beef

As a new user of Flexpod I have discovered that the FAS can run as both a SAN & NAS. How is this possible? NAS is nothing more than a bunch of disks in a RAID, shared out on a network (VLAN'ed, of course). NAS has none of the SANs redundacny or resiliancy, or does it in Flexpod?

Can someone please help explain this? I have a SAP implimentation that recommends NAS and that makes zero sense to me.

Find more posts tagged with

Accepted answers

Sir-Loin-of-Beef

Thanks. That helps a great deal. I've been working on SAN/NAS for over ten years but I have very little experience with NetApp. I still don't understand why anyone would want to use NAS for a DB application.

All comments

SeanHatfield

ONTAP provides storage services over a variety of protocols, both SAN and NAS. SAN can be FC, FCoE, or iSCSI. It can also provide file services over NFS and CIFS. The same FAS that provides FCoE boot luns to UCS, can provide NFS datastores to VMware or SMBv3 storage to Hyper-V, and iSCSI to a guest directly, along with general file services to clients on the network wether they favor CIFS with NTFS permissions, NFS with unix permisions, or even both, just to throw out some examples. The underlying storage system is resilient regardless of the protocol.

Sir-Loin-of-Beef

So NAS gains the same benefits as a SAN?

SeanHatfield

Each protocol has pros and cons depending on the use case, but with ONTAP you can leverage whichever one is most advantageous for your scenario.

AlexDawson

ONTAP aggregates are collections of disks in RAID groups.

From that point, you can either create volumes in your aggregates for NAS use (CIFS-SMB/NFS) or create LUNs in them for SAN use (FC/FCoE/iSCSI), or both.

At a simplified level, the LUNs live inside volumes too - but LUN access over SAN protocols bypasses the usual latency/access control of a filesystem, and writes direct to the RAID.

For SAN, redundancy at a client level is through FC/iSCSI multipathing through all controllers in a cluster, and for NAS, the IPs on the NetApp move between equivalent ports on the same node where possible, or onto another node in the cluster.

Hope this helps - please feel free to ask any followup questions.

Sir-Loin-of-Beef

AlexDawson

I know this is a little late of a follow up, but if you're interested in the NAS vs SAN use-case for databases - on a NetApp NAS system, files inside snapshots can be accessed through the directory tree, eliminating the need to mount/unmount LUN snapshots, etc. This can simplify backup and recovery. ONTAP 8.2 in cluster mode support SMB3, Continuous Availability and ODX, all useful technologies for MSSQL.

OZWALKERZ

I'm probably 2 years too late to be a useful answer, but NetApp NAS is far more than a bunch of disks!

Think of all the redundancy features you expect from a top notch SAN, and then appliy that to NAS file services.

All the disks are in double-parity raid groups, almost every component is duplicated, and the software offers rich features to assist with cross site duplication and management just to name two....

The new flash (SSD) arrays come with deduplication and compression enabled as standard, and are blindingly fast!

We serve Oracle databases and VMWare datastores over NFS, and although you do need to make sure your network has enough bandwidth, it's just as fast as FC. (We've got at 4 x 10GBps on some nodes and 2 x 40GBps connections on others!)

Add in the AutoSupport function which allows the NAS to effectively order it's own replacement disks when one fails, and it makes life very easy! Very nice function when you're managing 700 disks...

Quick Links

All Categories
Recent Posts
Activity
Unanswered
Groups
Help
Best Of