Hi guys -

I have a production Lync voice deployment in a multi-pool arrangement. I'm writing up the DR plan at the moment, and have worked through the official MS documentation on restoring service in the event of a complete site failure, where both the Lync Front Ends and the backend SQL database have gone offline.

As I have multiple pools with failover configured, limited voice services will be retained in a DR scenario. I'm just trying to understand the best way to restore full services at the DR site. As I see it, there are 3 options:

1) Do a forced move of all user accounts to Pool 1 -> Pool 2. Import the user contacts from backup. Re-create the response groups from backup. Assumes Pool 2 has capacity to host all users from Pool 1 as well (it does)

2) Restore the Lync Front End Servers from backup/snashot. Re-create the empty database for Pool #1 on the remaining SQL cluster at DR site (same hostname\path). Import the user contacts from backup. Re-create the response groups from backup

3) Restore the Lync Front End Servers from backup/snashot. Re-create the empty database for Pool #1 on the remaining SQL cluster at DR site (same hostname\path). Restore the RTC database from backup over the top of empty database. Re-create the response groups from backup

Have I understood the options correctly? Does anyone have experience with this sort of DR configuration, and can they provide me with their preferences/feedback/real world experience on the pros and cons of restoring the Pool 1 infrastructure versus just moving all users to a second pool and recovering their contacts?

Note: Let's assume the CMS is out of scope for this discussion.

Regards,

Here is a blog you can refer to.

http://blogs.technet.com/b/toninof/archive/2011/06/04/lync-server-2010-disaster-recovery-scenarios-restoring-enterprise-edition-back-end-server.aspx

Yes, I've already read that blog several times, but really after some thoughts from people with real world experience with DR planning for Lync 2010. Regards, James

James, we are running Lync 2010 in a similar environment tan you (three main Enterprise Edition pools in HR mode, plus multiple SBA installed in Branch Offices), we experienced a couple of huge outages in the last three years, but nothing that forced us to develop a plan as you are doing.

 

First of all, you must to reconsider the definition of Disaster Recovery, based on the functionalities that Lync provide: for sure, you can afford a few hours in Reduced Functionality Mode, which is provided by Lync just having more than on Registrar Pool deployed: users will be registered to other available pool, and they will be able to IM and call any other user, and if you planned your PSTN Outgoing Routing carefully, they also will be able to make and receive external calls. Obviously, they will lose some feaures like Conferences, Response Groups and Autoattendant, but Im sure that you can manage to live without those features during a catastrophic outage, or if you not, well, there is a couple of VoIP tricks that you could implement if you hire the right VoIP Engineer J

 

Instead to spend time and many in a DR plan that will be very (or impossible) to implement in the middle of the battle, better you should start thinking to have your Front End in HR mode, and your BE based in a reliable NAS with a quality proven backup system. If you are in the cheaper sidewalk and you implemented you FEs on Standard Edition, they you should think to move then to a Virtual environment, and have a reasonable backup schema on the farm hosts.

 

In our experience, Lync offer enough failover mechanisms to allow business continuity in almost any scenario. Trying to implement complex DR procedures is just a waste of time, time which, I repeat, you could invest in a very robust platform.

________

Sebastian