Arcserve UDP Part 3 – Off-site replication

Many organisations these days have an offsite requirement for their data backups. In the “old” days we used to backup to tape and on Monday someone will pick up the tapes and store them at an offsite location. Times has changed a lot since those days, as many organisations are using their another office or cloud to store their backup data.

Most of us in IT would use the “Backup Rule of Three” also known as the 3-2-1 Rule. Which stands for:

  • 3 copies of data that is important to the business (Original data + 2 backup copies)
  • 2 copies on different media
  • 1 copy stored offsite

Besides the main office to be protected, organisations also want their remote offices to be protected, ideally all with same solution and if possible in a unified console. As this is part of my UDP series I will explain what options Arcserve UDP provides to store data off site using replication. I will explain the possibilities using scenarios.

Scenario 1 – Main office and remote office

In this scenario, company X has two offices Main and Remote, both offices have a VMware cluster running with production data. Company X uses UDP to protect both locations and replicate data between the two sites.

In this scenario, a gateway is deployed at site two. Arcserve UDP uses the gateway to establish a secure connection between the Nodes and Console. The Gateway and RPS can be installed on the same server.

To install a remote office, go to the resources tab on the console and select sites (under infrastructure). Here you can add a new site;

When adding a new site, a wizard start, name the site and enter the console URL

Next step is to install the gateway software on the remote server, the wizard provides yo with a link to download the software or alternatively sends the administrator a link with the URL. Also in this screen (and email) is the authorisation code which binds the RPS server to the console and sets a secure SSL connection between the nodes and the console.

Last step of the wizard is the conformation

The remote site is now available in the console, and from here you can now install / configure the RPS server for the remote site. In my lab the RPS server is installed on the same server as the gateway.

Company X can use the replicate task in his plans for the both the main site and the remote site. Adding the replication task is very simple create a new plan or modify an existing plan, add a new task and this new task will use the source from Task 1.  Then select the destination for replication. In the other tabs you can set different retention for the replica data

In our scenario they have two plans:

Plan A – Protect Main Office

  • Task 1 – Agentless backup of virtual machines
  • Task 2 – Replicate backup to secondary site with different retention

Plan B – Protect remote Office

  • Task 1 – Agentless backup of virtual machines
  • Task 2 – Replicate backup to Main office with different retention

The beauty of UDP replication is that it uses deduplication from source. Which means that only blocks not known are replicated which can result in massive bandwidth reduction.

Below a screenshot I did earlier, I took a full backup of a SQL server in my lab and replicated it to my main location and because most blocks where known only a small portion was send across (655MB from 16GB full backup)

Scenario 2 – Main Office to MSP DR Site

In this scenario Company X would like to protect their main office and replicate the data to their MSP DR site, also they would like to have virtual standby (VSB). VSB creates a clone of the production machine and injects it with the last recovery points. If anything happens with the source machine you can start the VSB up with the same performance as the source. You can always run an instant VM (IVM) too, the difference is that an IVM runs from the backup image and a VSB is pre-provisioned on the hypervisor of choice.

Please note: I am using a MSP as an example in this post, however it can also be a different site from the same company…

In our scenario, the DR location runs a different hypervisor as the main office as it is on a MSP location.

As the MSP has already a console the steps are a bit different than when using a gateway.

  • Step 1 – MSP – Create a user account at the MSP site
  • Step 2 – MSP – Create plan to define destination
  • Step 3 – MSP – Map the user to the plan
  • Step 4 – MSP – Send the plan and the user account to the source admin via email or other means
  • Step 5 – Local Admin – Create Replication plan

I will go through these steps one by one. The first 4 steps are for the MSP and the last step for the local administrator.

At the MSP a user account need to be created to receive the source data, the source admin uses this account to connect to the destination console.

Once done, the MSP sets up a plan to receive the data from Customer X where the MSP defines the data store and merge schedule.

The next step for the MSP is map the plan just created to the user created in step 1.

Now this is done, the MSP sends the data to Customer X so that they can amend or create a new plan with the details from the MSP. The information that needs to be send is:

  • URL the console
  • Network details (i.e. ports to opened, VPN details if this is used etc)
  • Account details from Step 1

Customer X can now add the task replicate to managed console, in our scenario the plan looks like this:

  • Task 1 – Agentless backup of virtual machines
  • Task 2 – Replicate to remotely managed RPS

In this example I will amend an existing plan and add task 2 replicate to a remotely managed RPS server with the details received from the MSP.

Now this is done, Customer X can run his plan and first task 1 will be performed (backup) and then the replication starts.

Now the data is on the MSP site, the MSP can amend the plan to add virtual standby of the nodes protected. When a MSP goes into the plan he can simply add a new task with VSB.

Lastly, the Assured Recovery task is also available, which means that the nodes from Customer X can be tested as an instant VM or mounted as virtual disk to make sure that nodes are actually recoverable as a virtual machine or that you can recover files.

Arcserve UDP has also cloud integration with Amazon, Azure and of course Arcserve Cloud. This functionality will be covered in later posts.

Please follow and like vDutchy:
Visit Us
Follow by Email


  1. Arcserve UDP has also cloud integration with Amazon, Azure and of course Arcserve Cloud.
    looking for post cover this functionality.

  2. why replication taks is slower than backup tasks?

    I use udp 6.5, my backups are fast. I am trying to replicate the data to another server, but it is very slow. The servers are on the same network

    1. Hi Alessandro, replication to another RPS server should have good performance. However performance could depend on a lot of things such as network speed, the replication target disk performance etc. How are you replicating? How is your data store configured at the replicate site, how is dedupe configured etc. I would recommend to contact Arcserve Support or your local sales team for a quick health check of your environment. Cheers Harold

Leave a Reply