Written By: Murray White – Solution Architect with Vision Solutions

INTRODUCTION

This is a demonstration of how Double-Take Availability complements Equallogic array-based replication.

First, let’s review what Double-Take Availability provides. It provides two key functions.
1) Data synchronization and replication
2) Rapid failover of services or the entire server, in a physical or virtual environment.


Data Replication
With respect to Double-Take replication, Double-Take can replicate individual files, folders, volumes, or entire servers. Where array-based replication is not available, for example with a source server using MD3200 storage and a target standby server using EqualLogic storage (so no SAN replication possible), Double-Take can replicate everything that is required for the server to be protected to a standby server… both System-State and data.
When SAN/array-based replication is available, for example where a source server is utilizing EqualLogic storage for its data and its target standby server is also utilizing EqualLogic storage, Double-Take can be configured to exclude the data that is being replicated SAN-to-SAN (by the EqualLogic devices) and simply replicate the data that the array based replication does not.


Failover of services or the entire server
With SAN/array-based replication, the data on the SAN is replicated and is available on the target SAN but requires certain steps to make that data available following a failover condition. In the case of losing the server(s) that were connected to the failed storage, it is often a time consuming process re-provisioning servers to access the data that has been replicated SAN-to-SAN.
Double-Take Availability bridges the gap with this, and provides a means to failover the server (or servers) that have failed and need to make use of the data on the failed over SAN. Failing over servers with Double-Take Availability provides a rapid recovery time and greatly reduces the administrative effort required to make the failed over EqualLogic data available for use. In a matter of a few button clicks, you are able to function on a failed over server, accessing the SAN replicated data.


OVERVIEW
Let’s take a look at the environment we will be using for this demonstration.

[Untitled]

Source Server / Source SAN

We have a Dell R710 PowerEdge server running as a vSphere Virtual Host.

We also have an EqualLogic PS6000XV SAN storage device to host data for the Virtual Machine(s) running on the R710 vSphere Host.

The guest Virtual Machine named “VISION1” is running on the vSphere Host. The VISION1 guest Virtual Machine has a C:\ and F:\ drive.

C:\ contains the Operating System and other System-State (Program Files, User Profiles, etc), including Double-Take Availability. This drive is held on a Datastore on the vSphere Host.

F:\ drive is a data volume and its storage is a volume held on the PS6000XV EqualLogic SAN.
We have the F:\ drive accessing the EqualLogic SAN via an iSCSI Initiator connection that is configured from within the guest Operating System. In vSphere terms, this is considered a “Raw Device Mapping” (RDM). This is an alternative method of attaching to storage, versus having the storage maintained in a vSphere Datastore. There are arguments to using one over the other but when it comes to Double-Take complementing EqualLogic storage on a server, our testing resulted in the Raw Device Mapping being the preferred method to connect to the EqualLogic storage. RDM is the preferred method for three reasons…


1) The re-connection to the EqualLogic device post server failover is an easy process.

2) Consider that generally speaking Raw Device Mapping can offer a performance improvement versus offering the SAN storage via a Datastore.

3) Using the Raw Device Mapping approach, the connection to the EqualLogic storage is made from within the guest Operating System via the Operating System’s iSCSI Initiator feature. So, the same process can be used when using physical servers, as physical servers would have no choice but to use an iSCSI Initiator connection to the SAN storage.



Target Server / Target SAN (“DR”)

There is also a second PowerEdge R710 PowerEdge Server, running as a vSphere Host. This host contains a guest Virtual Machine (VISION2) that will act as a target/standby server for VISION1.

A second EqualLogic PS6000XV SAN storage device will maintain a replica of the Source-side PS6000XV volumes.

The VISION2 guest Virtual Machine is configured with just a C:\ drive.

C:\ contains the Operating System and other System-State (Program Files, User Profiles, etc), including Double-Take Availability. This drive is held on a Datastore on the vSphere Host.

There is no F:\ drive at this stage, because the VISION1 F:\ drive (an EQL volume) is being directly replicated from PS6000VX-to-PS6000XV, where the target-side PS6000XV is maintaining a “Replica” of the source-side volume. The Replica is an un-promoted volume, and therefore not accessible to the server until such time it is promoted (ie: following a failover of the storage).

Following a failover of VISION1 to this server (VISION2), you will need to manually reconnect the failed over server to the SAN storage (the promoted volume on the failed over EqualLogic device) with a quick and simple iSCSI Initiator connection from within the guest. This is shown in the live demonstration.




LIVE DEMONSTRATION

Please take a moment to view the following video. It is a live demonstration of the environment I have detailed here.

The video walks through the steps to configure Double-Take Availability for full server protection, with EqualLogic SAN replication handling protection of the SAN data. The live demonstration will conclude with a server failover in conjunction with the source-side EqualLogic PS6000XV failing over to its replica on the target-side EqualLogic PS6000XV, with the failed over server accessing that SAN data.
So, it is possible to failover one without the other, but this demonstration is failing over both server and storage and shows how easy it is to make the server and storage rapidly available in the event of a disaster.