This is the most fundamental configuration for data replication. A single
computer on the LAN or WAN has its data replicated to another computer,
either local or remote. This provides protection of the source
computer against catastrophic failure of the computer itself.
In a Fan-In configuration, multiple computers on the LAN or WAN have their data
replicated to a single computer, either local or remote. This provides
protection of all of the source computers against catastrophic failure, while
maximizing the use of resources by directing all the data to a single
destination computer.
On Windows, most of the configuration of replication and Recovery Point
options can be accomplished from the
Fan-In tab of the Agent Properties on the destination
computer, and these settings are automatically applied to all the source
computer. On UNIX, replication and Recovery Point options must be
configured on each source computer.
Although the scalability of a Fan-In setup can vary based on network and system
resources, it is recommended that each Fan-In setup contains no more than 100
source clients.
For maximum performance and robustness, the total number of Replication Pairs
configured for the same source volume should be kept to a minimum. If multiple
Replication Pairs must be configured for the same source volume, the recommended
upper limit is five.
Recovery Points created for a Fan-In configuration use VSS or ONTAP as the
snap engine for creating snapshots. The use of snap engine is based on the
destination being used. When the destination is a fixed volume then VSS is used
and when the destination is a filer or iSCSI device then ONTAP is used for the creating
snapshots. When destination volume is ONTAP LUN, Use ONTAP snapshot for ONTAP
LUN destination option should be selected for creating snapshots with ONTAP
snapshot engine. In case if option is not selected, VSS snapshot engine will be
used.
Consider the following for ONTAP snapshots:
Specify the user authentication details to be used for creating ONTAP
snapshots. The user information must be specified both in the
General tab of
the replication set properties on the source and in the General tab of the
agent properties on the destination.
Manual mounting of ONTAP snapshots is not supported in a
Fan-In configuration. If Auto-mount option is selected, ONTAP snapshots
will be automatically shared.
The following section provides the steps required to use CDR for data replication
and recovery in a Fan-In configuration based on multiple source computers and a
single destination.
Install CDR software on all the designated source and destination
computer.
Verify that all selected computers meet the System Requirements.
For more information see:
On UNIX, replication and Recovery Point options must be configured on
each source Replication Set.
On Windows, the options specified in the previous step - configure
replication and recovery point options - should not
be specified when configuring the Replication Sets, e.g., Recovery Point
options should only be set on the Fan-In tab of the destination
Agent Properties dialog box,
not on the Replication Set Properties dialog box on the source computer.
All source computers configured to Fan-In to the same destination
computer.
Each Replication Pair must use the same destination volume but a
different destination path.
Optionally, configure Throttling.
Note the following:
Throttling can be used to allow the destination computer sufficient
time to replay the logs it receives from all sources.
On Windows, this can
also help prevent the destination from running out of log space.
For Fan-In
configurations in particular, it is recommended that you configure the destination
computer to start throttling the source computers when its log space runs
low. For more information, see
Replication Activity
Throttling.
Configure the Throttling Amount on a source computer
to limit the maximum amount of data (in MB/second) it can transfer to the
destination, thus preventing a particular source from overwhelming the destination
during periods of high I/O.
However, if the source computer is throttled
too much, it may not be able to transfer all its Replication Logs quickly
enough, and it could run out of space. For more information, see
Network Bandwidth
Throttling.
Recovery Points are configured on the destination. These settings will be
applied to all Replication Sets for all source computers that use this computer
as a destination. See
Create a Recovery Point in a Fan-In Configuration for step-by-step instructions.
Recovery Points created for a Fan-In configuration use VSS or ONTAP as the
snap engine for creating snapshots. The use of snap engine is based on the
destination being used. When the destination is a fixed volume then VSS is used
and when the destination is a filer or iSCSI device then ONTAP is used for the creating
snapshots. When destination volume is ONTAP LUN, Use ONTAP snapshot for ONTAP
LUN destination option should be selected for creating snapshots with ONTAP
snapshot engine. In case if option is not selected, VSS snapshot engine will be
used.
Consider the following for ONTAP snapshots:
Specify the user authentication details to be used for creating ONTAP
snapshots. The user information must be specified both in the
General tab of
the replication set properties on the source and in the General tab of the
agent properties on the destination.
Manual mounting of ONTAP snapshots is not supported in a
Fan-In configuration. If Auto-mount option is selected, ONTAP snapshots
will be automatically shared.
Recovery Points are configured on each Replication Set on each source computer.
Thus, any Replication Set sharing the same destination can create Recovery Points.
For instructions, see
Configure Recovery Points and
Create
a Recovery Point.
To perform backups of Recovery Points, you must also install the appropriate
file system iDataAgent on the destination
computer. Note that you cannot replicate Windows data to a UNIX computer, nor
the converse. For the file system iDataAgent
installation instructions, see:
This configuration for data replication, referred to as "Fan-Out", adds
significantly to the protection afforded to the source computer, because
of the redundancy. A single computer on the LAN or WAN has its data replicated
to multiple computers, any of which can be either local or remote. This
provides protection against catastrophic failure of an entire
site, as well as the source computer itself.
A snapshot of the source volume for each Replication Pair is created
during the SmartSync Scan phase, which lead to significant space
requirements in a Fan-Out configuration, since a separate Replication Set is
required for each different destination. As a simple illustration, if you
have 5 destinations for the same source path, thus 5 Replication Sets each
having one Replication Pair, 5 snapshots will be created of the source
volume. Further, if you have 10 volumes which are each being replicated to
those 5 destinations, and you start replication for all of them
simultaneously, 50 snapshots will be created during the SmartSync Scan
phase. It is recommended in such circumstances to avoid starting all
Replication Pairs simultaneously.
The VSS cache can be configured using the vssadmin add shadowstorage
command from a command line prompt. Refer to Microsoft documentation for
details.
QSnap COW Cache space considerations:
As the SmartSync phases complete for each Replication Pair, these snapshots
are deleted by the system, but only in the order they were created for each
volume. If the first snapshot of a given volume was created for a Replication
Set using a slower destination computer, and it is the last one to complete
the SmartSync phases, no other snapshots of that volume can be deleted until
this first one is deleted.
By default, the location for the QSnap COW Cache is on the same volume
of which the snapshot is being created, but you can also specify a separate
volume to be used for all snapshots, in the
Client Properties (Advanced)
tab. See
Change
the COW Cache Location.
For Windows, note that since each destination computers communicates with the source
computer to indicate when it is finished with a log, a given log will not be
automatically deleted on a source computer until all destination computers
are finished with it. If one or more destinations are unavailable for any reason,
(or planning to be) in Fan-Out scenarios for prolonged period of time, the Replication
Pairs for that destination should be aborted, or the source will eventually
run out of space as all the replication logs for the offline destination(s)
accumulate.