Arcserve Replication & High Availability introduction

So far in my blog I have been discussing Arcserve UDP. But did you know that Arcserve also provides a solution that has the ability to provide you with zero data loss and instant fail-over capabilities! And did you know that this solution is already in the Arcserve family for 11 years!  Well, I think it is time to introduce you Arcserve Replication & High Availability aka RHA.

Back in the 2006 CA acquired XOsoft which had two products WANsync and WANsync HA, these products where introduced into the Arcserve family as RHA. When Arcserve became independent from CA in 2014, RHA stayed in the Arcserve portfolio. Today many companies in the world rely on this solution to provide high availability for their physical and virtual servers, and RHA can do a lot more than just HA..

What is RHA?

Arcserve RHA is an easy to use application that provides high availability and continues data protection for Windows, Linux and Unix environments on both physical and virtual servers. It provides RPO and RTO in mere seconds. That is pretty cool.

RHA first synchronizes the data from your Windows, Linux, and UNIX production servers to a second replica server (physical or virtual) that you provision locally or at any remote location, or even in the cloud.

Once synchronized, RHA continuously replicates all ongoing byte-level changes from your production server to the replica server. It provides application-aware replication, which means automated configuration and transactional integrity for applications such as Exchange, SQL Server, SharePoint, IIS, Microsoft Dynamics CRM and Oracle. You can protect other Windows applications by using the Custom Application Protection wizard or with a simple script (you can protect Linux and UNIX applications with a simple script as well).

RHA provides real-time server and application monitoring, automatic and push-button failover, automated end-user network redirection, push-button failback functionality and automated integrity testing, which will help you reduce any system downtime.  It will provide you granular, application-aware protection, recovery and availability.

RHA works with scenarios that you can setup using a step by step wizard, I will highlight the most common used scenarios in this blog, and will go deeper into each scenario in the next few posts.

  • Full system HA – the mostly used scenario, to protect any physical and virtual machine which will replicate to a virtual target. Full System HA supports cross hypervisor migration between VMware, Hyper-V and Xen.
  • Application HA – Failover applications between a master and a standby server, mostly used for Oracle, SQL and customer made applications, supports p2p, p2v, v2p
  • Content distribution – is aimed at addressing the need of reliably spreading and managing information across a highly distributed environment. The Content Distribution solution is also a powerful content delivery and web publishing solution that can serve your external customers. Through portals and web sites, you can deliver to your customers any information that is stored in files, from music to movies to documents to news.

In my blog, I will focus on Full system HA and Application HA.

Components

RHA is a very easy to deploy solution that does not require a lot of your infrastructure, the following components are used in a typical deployment:

  • Control service – This is the single point of control in replication and HA operations. It contains the dataset of the scenarios and it communicates with the engines and managers
  • Engine – This is the agent installed into the master and replica server, an engine is deployed via the scenario wizard. And it captures the data and replicates to the replica server
  • Management Center – This component provides the overview page, the manager to setup and monitor the scenarios and has a report center
  • FSHA Appliance – for Full system HA which use VMware or Citrix as a target, a virtual windows server has to be created with the engine installed. This functions as the replica target. A FSHA appliance can be used for multiple master replica’s. If a Hyper-V target is used, we don’t need to install a FSHA appliance as we can directly write to disk with Hyper-V
  • Powershell (optional) – If you (like me) love powershell, you can manage the whole solution using CLI

 Scenarios

Full System HA

So let me explain two scenarios that we often use with RHA, first of full system HA

In this scenario, we will protect a physical server, however it can also be a virtual server.

1 – first we setup the scenario using the management center and the control service, the scenario installs the engine on the physical server, and as we are using VMware as a target, we installed a FSHA appliance it gets its engine also deployed via the control service. This engine only needs to be installed once.

2 – the initial synchronization starts and it replicates all data to the target FSHA appliance

3 – during the initial sync the data will be written to a VMDK disk(s) that will is attached to the FSHA appliance. This disk is created during the initial setup of the wizard and for each disk on the master it will create a disk on the target site.

4 – once the initial sync is done, a heartbeat will monitor the master service, from now on every change made on the master is asynchronous replicated to the replica. Heartbeat settings can also be configured in the scenario setup.

Now this is done, you are protected, if anything happens to the master you can fail-over to the replica. This process can be manual or automatic.

So lets see what happens in case of a fail-over

1 – the replica engine monitor detects that the master is no longer available and depending on the settings made in the scenario it will either notify you or start a fail-over after an amount time configured before failing over

2 – when the fail-over is initiated, the vmdk disk(s) attached the FSHA appliance are detached,

3 – a VM will be started using the vmdk files. Once started, network will be redirected for the end users (in case of different subnets) and users can work again on the virtual server, the RTO is the boot time of the server.

As you see this is a pretty straight forward operation and very easy to install.

Application HA

The second scenario I want to cover is application HA

This scenario works different than full system HA as we protect the application and its data.

In this scenario, I am using SQL server as an example.

1 – Again first we setup the scenario using the management center and the control service, the scenario installs the engine on the master SQL server and installs the engine on the standby SQL server. The standby server must be the same operating system and must have the same application (in this case SQL server) installed. With application HA, network redirection will have DNS redirection on by default, which means that standby server will get the master IP address in case of fail-over

2 – as with the full system HA, the initial sync starts

3 – when the initial sync is finished, the standby server will monitor the master user

The setup is very straight forward and you can protect your application server in minutes. So what happens if the master servers fails?

1 – the replica engine monitor detects that the master is no longer available and depending on the settings made in the scenario it will either notify you or start a fail-over after an amount time configured before failing over

2 – the services on the standby server will be started and the network will be redirected to the standby server for the end users, this process will take a few seconds in total. If the master server comes back online the users still will continue to use the standby server, until you do a reverse replication and switchback.

Every Tuesday we do a webinar around RHA, below a recording of this webinar that I have done a while back demoing Application HA using a SQL server:

So there you have it, RHA is gem of a solution which provides a lot functionality but more important helps you to sleep as you can be assured that you can recover from any outage easy and quickly.

I will create some posts soon where I will go deeper into scenario creation and doing fail-overs.

Download UDP: http://okt.to/eXmslE
Arcserve High Availability Live Webcast: Every Tuesday, 10:00 BST: Register: http://okt.to/mR4fEN
Arcserve UDP Live Demo: Every Friday, 10:00 BST: Register: http://okt.to/fALlGx
Please follow and like vDutchy:
LinkedIn
FACEBOOK
Google+
Google+
http://vdutchy.com/arcserve-replication-high-availability-introduction/
RSS
Follow by Email

Leave a Reply