Tip:
Highlight text to annotate it
X
Hello. My Name is Chris Klosterman and I’m a Senior Solution Architect with EMC's Isilon
Storage Division. Today, I’m going to talk about EMC Isilon's
SnapshotIQ licensed feature, and how it can be part of a comprehensive data protection
strategy. Customers find that today unstructured, or
file data, is growing significantly faster than structured or block data.
This presents two challenges. First, file data is becoming more business
critical. And second, because file data is becoming
more business critical, users are demanding more frequent backups and rapid restores.
In this video, we will discuss how SnapshotIQ can help address these challenges.
A snapshot is a point-in-time picture of a directory and all the contents therein.
At the point-in-time when a snapshot is taken, it consumes no space, but as the live data
diverges from that point-in-time, the snapshot grows based upon the blocks of data that have
changed. Why is a snapshot needed?
The easiest way to explain why a snap is needed is to tell a story.
Let’s take for example a given directory of data.
At 1 a.m., the directory is backed up to tape. At 8 a.m., users come into the office and
start working. And then at noon with SnapshotIQ on its scheduled
basis, we take a snap. When users return from lunch at about 12:45,
one of them accidentally deletes or corrupts some portion of their data.
At this point we have two options for recovery. The first is we can recover from a backup
from tape from 1 a.m. The second is we can allow the user to perform
a self-service restore from the snapshot at noon.
This results in much faster recovery and much less data loss.
Let’s use a simple example to explain how SnapshotIQ works.
Imagine if you will that each character in the word, “R-A-N-D-U-M” is represented
by one block. At 9 a.m. we take a snapshot of the word.
At the point-in-time when that snapshot is taken, it is effectively a series of pointers
looking at the live data. Obviously, the user has made a typo.
At 9:30, the user realizing their typo, replaces the “U” with an “O”.
And using “copy on first write”, we will move that “U” down to the snapshot, so
if we ever needed to recover that data, we still have that consistent point-in-time from
9 a.m. The snapshot also now consumes one block of
space. On Isilon’s OneFS file system, snapshots
are stored in the same namespace as any other file.
There’s no need to reserve space specifically for snapshots, and once they are deleted,
the space is available to be overwritten by any other file or folder.
Because snapshots do not demand performance, we’re able to store them on a separate near-line
tier of storage from the production data that may demand higher performance.
Protecting data with SnapshotIQ is done like most things on Isilon’s OneFS filesystem:
at a directory level. And because a directory has a common dataset,
such as home directories, group shares, or applications, each of these workflows usually
has a common RPO, or recovery point objective. Therefore one example schedule might be hourly:
take one snapshot and keep a rolling 24 for one day.
Daily: take one snapshot and keep a rolling seven for one week.
And weekly: take one snapshot and keep a rolling four for one month.
SnapshotIQ offers options to directly recover both from Microsoft Windows as well as Unix
or Linux clients. For Microsoft Windows clients over SMB, we
offer VSS integration, which allows users to right-click on a file or folder, choose
the previous versions tab, and then choose the files that they wish to restore.
From Linux or Unix clients over NFS, you can simply “cd” to “.snapshot”, and choose
the files that you wish to restore. These options reduce the time to recover as
well as eliminate the need to frequently recall tapes.
To maintain security, these options can be enabled or disabled at the discretion of the
NAS administrator. Snapshots when created either manually or
via schedule should have an expiration date and time set.
When that expiration date and time is met, the snapshot is marked for deletion.
Automatically OneFS will run a job in the background called SnapshotDelete, which frees
up the space used by the snap and allows it to be used for other files or folders.
Snapshots are by their very nature read-only. Think about a photograph: once it is taken
it cannot be changed. This ensures data integrity, which is the
point of the snapshot to begin with. OneFS offers the capability, however, of what
we refer to as a file clone, which allows an administrator to create a read/writeable
copy based upon a read-only snapshot. The purpose of this is to allow application
testing without impacting live production data.
For more information about implementing SnapshotIQ in your environment, please contact your EMC
or partner account team. Further details are available in the OneFS
Administration Guide on EMC Support website. Thank you for watching.