none
Scom 2019 deployment across two datacenters RRS feed

  • Question

  • Hi. I am designing a Scom 2019 solution using sql always on with a 2 node active cluster in 1 datacenter and a passive / failover node in the second datacenter. This witll monitor about 500 servers across both DCs. Pls i need a best practice advice on how to monitor both DCs even if the 1st dc goes down the 2nd dc will still be active. Can i achieve this using one database or will need two?

    Need advice on how many management groups will be required for high availability. Also can VMware SRM be utilised to failover servers from 1st DC to 2nd DC. Not sure if scom supports this

    • Edited by J8ggy Tuesday, June 25, 2019 4:50 AM Typo
    Tuesday, June 25, 2019 4:46 AM

All replies

  • Hello,

    It’s enough to have a single management group, and have all SCOM agents reporting to it, your SCOM environment will still work whether a Domain Controller (DC) goes down or not.

    For high availability I’d recommend to have a SCOM management server at each site, and of course your SCOM databases either clustered/AlwaysOn.

    Another approach is to use the multiple management groups, in your case one (1) management group at each site. Then configure your SCOM agents for multi-homing, meaning they report to both management groups.

    However I see no need in having multiple management groups here, so I would recommend using only one (1) management group for less management overhead.

    SCOM provides automatic failover for the SCOM management servers, if for example: if one management server goes down or becomes unreachable, or when let’s say one site goes down, so there should be no need for SRM.

    Here’s some more information:

    Planning a Management Group Design

    Assigning Gateways and Agents to Management Servers using PowerShell

    Best regards,

    Leon


    Blog: https://thesystemcenterblog.com LinkedIn:

    Tuesday, June 25, 2019 5:59 AM
  • Hi Leon,.. Thanks for your quick response and very much appreciated. I understand a SCOM management group provides High availability if one of the management server goes down. If I have a single management group (MG) in the primary DC at the DC goes down I am concerned I will have no monitoring for servers in the second datacentre. How does the automatic failover or is configured if DC1 goes now and I want DC" to take over. Will this require a second MG in DC2 with multihoming. Also if I do create a MG in DC2 can it write back to the clustered DB in DC1?
    Tuesday, June 25, 2019 11:30 AM
  • Failover/High availability is automatic in the Active Directory (AD) as well, so if you have more than one Domain Controller (DC), everything will failover automatically.

    I suggest you have a SCOM management server in each datacenter, this way you will avoid unecessary latency issues and your servers will likely communicate with the closest DC.

    In case of a disaster, and let’s say the primary datacenter goes down, everything will automatically failover and work on the secondary datacenter, your monitoring will still be up and running.

    If you want to do multihoming, then you will need to create a second management group (one in the primary datacenter and another one in the secondary datacenter).



    Blog: https://thesystemcenterblog.com LinkedIn:

    Tuesday, June 25, 2019 11:55 AM
  • Thanks Leon.. Can both management groups in the two DCs (DC1 and DC2) use the same database in DC1 or it needs to be separate
    Tuesday, June 25, 2019 1:25 PM
  • Separate databases.

    Blog: https://thesystemcenterblog.com LinkedIn:

    Tuesday, June 25, 2019 1:28 PM
  • Unless the latency is low between the datacenters (<10ms), having management servers in DC other than the one where the DBs reside isn't a good idea.

    &quot;Fear disturbs your concentration&quot;

    Tuesday, June 25, 2019 3:58 PM