Hi experts, * Single Domain - Management server placed in 3 data centers * SQL DB and DW are clustered/ Reporting services installed on a standalone server * Total we have 4000 agents (Including Unix, Linux & Network Devices etc..) * we have planned to setup DR site Scenario 1: 1. If i am setting up a SQL log shipping for SCOM 2012 database and i got two management server setup in DR. Q: Incase if there is a outage, how the management server in DR site will take over the DR log shipped SQL standalone server. Is it automatic or we need to any manual change. Scenario 2: 1. In case if there is a WAN bandwidth issue, how effectively can we change the setup. Regards, Pratap

HI, Regarding the SCOM HA or DR settings, you could refer to the following links: SCOM 2012 & Multi site SQL HA http://social.technet.microsoft.com/Forums/en-US/operationsmanagerdeployment/thread/c69ca77b-fa03-4cb1-8373-201b0d51956c/ Configure SQL Server Log Shipping Guide for the System Center Operations Manager 2007 R2/2012 Operational Database http://social.technet.microsoft.com/wiki/contents/articles/11372.configure-sql-server-log-shipping-guide-for-the-system-center-operations-manager-2007-r22012-operational-database.aspxAlex Zhao TechNet Community Support

Hi Pratap, You may refer to below thread as well apart from links posted by Alex http://social.technet.microsoft.com/Forums/en-US/operationsmanagergeneral/thread/b449e1f5-84d2-472c-8434-d65797109bb1 And http://blogs.technet.com/b/mgoedtel/archive/tags/ha+guidance/ Thanks, Varun

Hi, when using log shipping to a warm standby server somewhere else. that other server will have a different name. if primary site fails (with sql and some managements ervers) you need to tell the management servers to start using that other sql server. (manual change, stop services, change registry key, start services). SSRS and such are the same story basically. Its quite do-able. At one customer I am working for they are looking at a scenario where they want to use log shipping to another server and use SQL aliasses (is a per server setting). Means that on any applicable server we dont need to make 1 to many changes here and there on the box, but just adjust the SQL alias in one spot and point it to the other ip address so to say. Single point of change on the machines (like management servers). We are still in testing to see if this will work for SCOM as well. Testing will take some time here, so cant give you an answer if this works as they thought it would. In any case I hope you dont have a bandwidth issue mate. At least between ANY management server and the main database server for SCOM you should have low latency. Otherwise you cant place management servers there. You could think that if those dont have any agents and such to monitor that it is a little bit less important. But also that means you should use resource pool changes to force the behavior for loading of workload for certain components. And doing that means you need to change those in the case the primary site goes down. Manual changes again, not preferred as far as im concerned. Bandwidth and latency are importan factors when you want to be able to use the other datacenters for placing SQL or management servers for your management group and ensure monitoring will continue.Bob Cornelissen - BICTT (My Blog about SCOM) - MVP 2012 and Microsoft Community Contributor 2011 Recipient

To add to this story: a lot of things you can do for DR means you are creating complexity in your environment and stuff can go wrong. for instance you might find that you are (or if lucky... are not...)... troubleshooting log shipping a lot. But creating more and more security in your DR and possibilities to run the whole infra from somewhere else creates complexity in your current network and more error possibilities. SOme things are worth it and are easy and robust. Other things are not. As John and others point out in some of the referenced links above you could go for a second management group located in the other datacenter and multihome your agents. I assume the machines you monitor do have the resources to double-monitor. Both methods would give you some form of DR with some amount of work before and during the "site crash". First go for the primary processes and components you really need. Some others go for using virtual machines for all scom components and next check out the DR/replication/HA possibilities of the hypervisor. Its also a possibility.Bob Cornelissen - BICTT (My Blog about SCOM) - MVP 2012 and Microsoft Community Contributor 2011 Recipient