|
| | |
|
Snapshot Technology for GuardianOS™
Snapshots and How to Use Them With The GuardianOS Platform
Click here for detailed
pdf of this article
Introduction
In the digital age, an ever-increasing percentage
of a company’s critical business data is stored electronically. With
prohibitively high costs of network and data downtime, high availability of this
critical asset is a necessity for any network infrastructure. This data must be
well protected to enable easy and rapid recovery in the event of a file loss,
data corruption, or user error. Snap Server™ by Adaptec, built on GuardianOS by
Adaptec provides storage solutions designed to protect a company’s data assets
with its integrated proprietary snapshot technology.
Overview
GuardianOS Provides Powerful, Flexible Snapshot
Technology
The snapshot
implementation of GuardianOS enables users to take instant, disk-based
point-in-time images of any volume on a GuardianOS-powered Snap Server. Key
features of a snapshot include:
Key Features
•Instant virtual volume— The snapshot
instantly captures a copy of the live volume.
•Point-in-time image— The snapshot is a
virtual image of the live volume as it appeared at the point-in-time when the
snapshot was captured. It creates a nearly instant map of pointers to the actual
blocks of data found in the live volume.
•Disk-based storage— The snapshot images
are stored entirely on disk. Blocks of data that have not changed since the
completion of the snapshot remain in the live volume, while the original
contents of data blocks that have changed are stored in designated “snapshot”
spaces.
•Sustained client access to the live volume— With the exception of a momentary pause as the live volume is frozen and
captured in the snapshot image, network clients maintain read/write access.
•Read-only access to the snapshot volume— Users have read-only access to the individual snapshot volumes. Since these
snapshot images are true point-in-time copies of live volumes, users cannot
manually modify them, protecting the snapshot from inadvertent user error.
•Performance maintained during concurrent
snapshots— The GuardianOS server can accommodate multiple concurrent
snapshots per unit while continuing to maintain optimal performance.
How It Works
Creation of Point-in-Time Images
The snapshot technology utilizes an optimized
copy-on-write implementation to create pointin-time images of the live volume.
When a snapshot is initiated:
1. The snapshot technology momentarily “quiesces”
or freezes the live volume.
2. The virtual copy of the live volume is
near-instantly created, as a database of pointers to the actual blocks of data
(which presently exist in the live volume) is created.
3. When client-initiated write requests are made,
the snapshot technology intercepts these requests, reads the data in the soon
to-be-overwritten blocks, and saves the “original data”in the snapshot volume.
4. After a copy of the original data is made, the
client-initiated write requests are completed.
5. Further write requests to the live volume
trigger steps 3 and 4, as a copy of the original data is made when new write
requests are initiated.
Please note that if the volume receives further
requests to overwrite blocks of data which have already been overwritten and the
original contents have already been captured in a snapshot volume, those
requests will be ignored by the snapshot subsystem. Instead, only the very first
overwrite of a set of data blocks will trigger the copyon- write operation
described above.
6. Upon receiving client requests to read the
contents of individual snapshots, the snapshot technology simply substitutes the
original data blocks for the changed data blocks
Applications
High-Performance/High-Availability Storage
Solutions for Demanding Environments
When deployed in any given application or
environment, the GuardianOS storage solution’s optimized snapshot implementation
provides:
Online backup.
Each
snapshot is a consistent image or virtual copy of the live data volume at a
given point in time. The GuardianOS snapshot technology enables users to
maintain multiple concurrent snapshot images as online disk-based archives.
Furthermore, each of these snapshot images may be shared for user access, and
subsequently backed up to tape or alternate media for archival purposes.
An elegant means of side-stepping the backup
window.
In traditional network environments, the
network administrator performs regular backups of critical server data to tape
for both backup and archival purposes. Typically, this includes any combination
of daily incremental backups and weekly/monthly full backups. To minimize
business impact, the daily backup operations are usually performed, with the
servers taken offline, during non-business hours, or the so-called “backup
window.”
In today’s global marketplace,
information systems infrastructure and digital business assets must be made
available 24x7. The cost of downtime can be prohibitive, particularly in the
ecommerce, financial, and banking industries. When network storage is deployed
in such transaction- and/or I/O-intensive applications, the protection of
digital business assets and valuable data is more crucial than ever before.
The GuardianOS snapshot
implementation provides the means to elegantly sidestep the backup window by
allowing network administrators to more quickly and easily perform daily
incremental backups and weekly/monthly full backups to tape. The snapshot images
of the volume can be backed up throughout the course of the day as needed.
Because snapshots are chained, system performance is maintained whether a volume
contains one snapshot or many. In addition, all of the snapshot backup and
archival operations are transparent to users.
Instant recovery of user data.
The GuardianOS snapshot technology enables the storage and
maintenance of multiple snapshot images. Each of these snapshot images remains
available online as a disk-based archive. Furthermore, the network administrator
can configure each snapshot image to be shared over the network. This enables
individual users read-only access to the snapshot images from which they can
copy any inadvertently modified, deleted, and/or corrupted user data and files
back into the live volume.
Offline volume for reporting and revision
testing.
The snapshot volume contains the collection
of all completed and active snapshots. Third-party or SRM-provided reporting of
storage trends by volumes, file types, and applications may be run on individual
snapshot images. This enables the administrator to proactively monitor storage
utilization and plan for data growth, while removing the access contention and
performance overhead associated with running reports on a publicly accessed live
volume. The ability to maintain a collection of snapshot images, corresponding
to multiple points-in-time, makes the snapshot technology ideal for revision
management and testing. Multiple copies of software code or documentation can be
easily tracked, maintained, accessed, and audited for testing purposes.
Benefits
Reducing Costs While Enhancing Data Availability
Regardless of the application and the environment
in which the GuardianOS-based solution is deployed, it is easy to see that
snapshot technology provides many real benefits to administrators and end users.
These benefits include:
•
Increased availability of both storage resources and
user/system data.
The use
of multiple snapshots as online backups provides maintenance of multiple,
consistent copies of data at various points-in-time.
•Enhanced protection of critical data.The use of both snapshots as disk-based “virtual volumes” and regular
incremental backups of daily snapshots to tape provides multiple levels of
protection for critical business data.
•Instant recoverability of user and system
data from disk.This capability provides time-to-data advantages over
traditional data retrievals from tape. Users and network administrators can
recover specific data from disk-based snapshots almost instantly.
•
Lower total cost of ownership (TCO); higher return on
investment (ROI).
The
increased uptime, availability of critical business assets and user access to
data translates into higher user productivity. The easy administration and use
of snapshots for a myriad of applications further contribute to lower TCO and
increased ROI.
How To Create Snapshots
Simple, Flexible Implementation
Set up snapshots for on-demand or scheduled execution The
administrator has control over the following options:
•
Name
of the snapshot.
•Source Volume to Snapshot.The
administrator can specify any available volume.
•Start Date and Timeto the nearest half
hour.
•Repeat Intervalsto schedule recurrence
and frequency. Multiple concurrent snapshots can be maintained, enabling an
administrator to keep up to approximately two weeks’ worth of daily and weekly
snapshots.
•Durationof the individual snapshots.
The administrator sets a snapshot’s lifetime or duration, at the end of which
the individual snapshot is automatically deleted to conserve storage capacity.
This is typically used to dispose of older snapshots that have been properly
backed up to tape.
•Create Recovery Fileto capture volume
attributes and settings. This enables the administrator to capture volume
settings and extended attributes such as ACLs and quotas. Note that optimal
performance and the number of stored snapshots depends on size of the snapshot
storage pool the administrator has created and the amount of data being backed
up. That said, many administrators have found the optimal balance between data
protection and storage space to be roughly 15-20 concurrent snapshots.
Restoring Data From Snapshots
Accessing Snapshots/Instant Data Recovery
Once completed, snapshots are listed in order of creation, with
a display of pertinent information (Figure 4). Completed snapshots can be
accessed by the network administrator and any clients with appropriate security.
The network administrator first needs to create a network share, named SS_Share1
by default, of the snapshot volume, corresponding to the network share, Share1,
of the live volume. When viewed from Windows Explorer, the contents of the share
SS_Share1 appear as shown in Figure 5. Should administrators or end users
experience data loss from data corruption or user error, they can simply copy
the needed data from the last available snapshot back into the live volume.
Snapshot Backup Strategies
Easy Integration With Your Existing Data
Protection Plan
Backing Up a Snapshot.
An integral part of a company-wide data
protection plan, the snapshot technology lends itself readily to providing both
online backups and subsequent backups to tape. Using the network backup software
(BakBone NetVault [included with OS], VERITAS NetBackup/Backup Exec, Legato
NetWorker, or CA ARCserve) and a tape product of choice, the administrator may
continue to do weekly and/or monthly full backups to tape on the weekends. On
all other days of the week, they can perform daily incremental backups of the
snapshot volume to tape.
In the example on the previous page, the network
administrator can do weekly full backups to tape every Saturday at 12:00 a.m.
Assuming that the weekly full backup requires up to 48 hours to complete, they
can create daily snapshots of the live volume: snapshots snap0, snap1…through
snap5, which are captured at precisely 12:00 a.m. on Monday, Tuesday…through
Friday.
Since the snapshot process can take as little as
a fraction of a second to complete, the administrator can safely initiate daily
incremental backups of the snapshot volume any time after 12:00 a.m. on each of
those days. Once configured for daily recurrence, the incremental backup job
will automatically backup individual snapshots: snap0 on Monday, snap1 on
Tuesday, etc. Consequently, the daily backups can take up to 24 hours to
complete, enabling the administrator to sidestep the “backup window.”
Restoring Data from Tape.
To restore data from tape, administrators can simply use the
backup software to select individual or groups of folders/files for restoration
to the live volume. They can do so by using the latest backup job to find the
required data for restoration. It may be necessary for the administrator to set
the target for the redirection of the data restoration; by default, most backup
software restores data to the original location. Since administrators (and end
users) have read-only access to the snapshot volume, they must restore the data
to the live volume instead.
Using snapshots with NDMP.
An NDMP backup operation automatically initiates a snapshot. If
the snapshot pool does not have sufficient space to file the new snapshot, it
will create space for it by removing the oldest snasphots.
Using snapshots with iSCSI.
It is not recommended that snapshots be used to backup iSCSI
volumes.
Conclusion
Speeding Data Recovery While Minimizing Costs and
Complexity
Storage solutions powered by GuardianOS provide snapshot
technology. This easy-touse yet powerful feature enables the capture of
point-in-time images of the live volume to provide the network administrator
with a means for effectively eliminating or sidestepping the backup window and
give all users instant recoverability of lost or corrupted data. This feature,
along with hardware component redundancy, enhanced system security, high
performance, fault tolerance, and ease-of-deployment and administration
contributes to the overall system and data reliability, availability, and
serviceability. As a result, companies can achieve low TCO and high ROI.
| |
| | |
| 
|
|