locked
SCCM 2012 CAS RRS feed

  • Question

  • I have few questions on CM12 Central Administration Server (CAS)

    1.Why many of them are recommending not to have CAS in the hierarchy in SCCM 2012?

    2.What is the Complexity,If we have CAS.even for the mid range hierarchy (30K - 60K End Users)?

    3.Any Maintenance/Cost burden on the organization?

    4.If not CAS,Can one Primary Site server will take care of the workload.If Yes.How about the BCP and Failover?

     

    Tuesday, May 14, 2013 8:35 AM

Answers

  • Multiple primary sites have *nothing* to do with business continuity, high availability, site resiliency, redundancy, or disaster recovery. The documentation never even mentions this when discussing primary sites -- just because I have two of something doesn't mean I have any of those things.

    Primary sites are for one primary thing -- client scalability in terms of clients supported; i.e., each primary site can support up to 100,000 clients. There are a couple of other scenarios that primary sites address, but these are generally not used but nowhere is HA, SR, DR mentioned or eluded to when it comes to primary sites.

    To directly address these concerns in ConfigMgr, you add multiple site systems hosting duplicate roles that are critical to client operation like DPs, MPs, and SUPs -- all other roles and the site server itself are not critical to client operations.


    Jason | http://blog.configmgrftw.com

    Tuesday, May 14, 2013 12:42 PM
  • I have few questions on CM12 Central Administration Server (CAS)

    1.Why many of them are recommending not to have CAS in the hierarchy in SCCM 2012?

    2.What is the Complexity,If we have CAS.even for the mid range hierarchy (30K - 60K End Users)?

    3.Any Maintenance/Cost burden on the organization?

    4.If not CAS,Can one Primary Site server will take care of the workload.If Yes.How about the BCP and Failover?

     

    1. 100K+ of clients, this is really the only reason for having a CAS - Scaling. There are some minor technical reasons why you may go for a CAS such as MP isolation (clients will only use the MP's at their given site and not all MP's across the entire hierarchy). I always recommend people use a stand-alone primary hierarchy and not a CAS, there are only a handful of companies that need a CAS in the "entire world". CAS introduces complexity when it is not needed. It's like giving yourself a self-induced headache.

    2. Because it's not a good design practice to ignore new design elements in favor of producing a CM07 design for a CM12 hierarchy. A lot of people are not adjusting to the new products design principles and instead are forcing a CM07 mindset into their decisions for a CM12 hierarchy. This i bad. Complexity, replication, object locking and many other items are introduced when adopting a CAS. At the end of the day everyone see's everything as there is no such thing as a tiered primary site, they are in a ring of replication now, administer in one place and appear everywhere, and RBA is used to create the partitioning that folks used CM07 tiered primaries for.

    3. licensing costs go up - SQL, OS, hardware (unless it's VM's). A CAS introduces multiple moving parts. There is additional support overhead for fixing replication issues, bugs in that space, usual stuff.

    4. There is no such thing as a multi-site DR solution, I think Jason below points this out, you can certainly attempt to keep multiple sites and when DR comes you could in theory reassign ALL your clients to the new site using Login script or whatever means available to you, it is not a part of the DR strategy from Microsoft. I have always said that assign to a new site is the closest you'll get to a DR solution that doesn't involve rebuilding the failed machine but it is clunky. Microsoft DR for this product is simple "recover your site", and since now in CM12 we can get away with recovering just the database we can return a server to operational status far faster than every before.


    Rob Marshall | UK | My Blog | WMUG | File CM12 Feedback | CM12 Docs | CM12 Release Notes

    Saturday, May 18, 2013 10:47 AM

All replies

  • I CAS is (in most cases) simply not needed if you have less then 100k clients. Why making it more complex (replication) as it should be (plus expensive: more hardware)? A CAS does not provide fault-tolerance and clients will never perform any automatic failover.

    Torsten Meringer | http://www.mssccmfaq.de

    Tuesday, May 14, 2013 8:57 AM
  • Hi Torsten,Thanks for the response

    Then how about Business continuity and Fail over,If i am going for stand-alone primary site server?


    srkr

    Tuesday, May 14, 2013 12:22 PM
  • Tuesday, May 14, 2013 12:28 PM
  • Multiple primary sites have *nothing* to do with business continuity, high availability, site resiliency, redundancy, or disaster recovery. The documentation never even mentions this when discussing primary sites -- just because I have two of something doesn't mean I have any of those things.

    Primary sites are for one primary thing -- client scalability in terms of clients supported; i.e., each primary site can support up to 100,000 clients. There are a couple of other scenarios that primary sites address, but these are generally not used but nowhere is HA, SR, DR mentioned or eluded to when it comes to primary sites.

    To directly address these concerns in ConfigMgr, you add multiple site systems hosting duplicate roles that are critical to client operation like DPs, MPs, and SUPs -- all other roles and the site server itself are not critical to client operations.


    Jason | http://blog.configmgrftw.com

    Tuesday, May 14, 2013 12:42 PM
  • I have few questions on CM12 Central Administration Server (CAS)

    1.Why many of them are recommending not to have CAS in the hierarchy in SCCM 2012?

    2.What is the Complexity,If we have CAS.even for the mid range hierarchy (30K - 60K End Users)?

    3.Any Maintenance/Cost burden on the organization?

    4.If not CAS,Can one Primary Site server will take care of the workload.If Yes.How about the BCP and Failover?

     

    1. 100K+ of clients, this is really the only reason for having a CAS - Scaling. There are some minor technical reasons why you may go for a CAS such as MP isolation (clients will only use the MP's at their given site and not all MP's across the entire hierarchy). I always recommend people use a stand-alone primary hierarchy and not a CAS, there are only a handful of companies that need a CAS in the "entire world". CAS introduces complexity when it is not needed. It's like giving yourself a self-induced headache.

    2. Because it's not a good design practice to ignore new design elements in favor of producing a CM07 design for a CM12 hierarchy. A lot of people are not adjusting to the new products design principles and instead are forcing a CM07 mindset into their decisions for a CM12 hierarchy. This i bad. Complexity, replication, object locking and many other items are introduced when adopting a CAS. At the end of the day everyone see's everything as there is no such thing as a tiered primary site, they are in a ring of replication now, administer in one place and appear everywhere, and RBA is used to create the partitioning that folks used CM07 tiered primaries for.

    3. licensing costs go up - SQL, OS, hardware (unless it's VM's). A CAS introduces multiple moving parts. There is additional support overhead for fixing replication issues, bugs in that space, usual stuff.

    4. There is no such thing as a multi-site DR solution, I think Jason below points this out, you can certainly attempt to keep multiple sites and when DR comes you could in theory reassign ALL your clients to the new site using Login script or whatever means available to you, it is not a part of the DR strategy from Microsoft. I have always said that assign to a new site is the closest you'll get to a DR solution that doesn't involve rebuilding the failed machine but it is clunky. Microsoft DR for this product is simple "recover your site", and since now in CM12 we can get away with recovering just the database we can return a server to operational status far faster than every before.


    Rob Marshall | UK | My Blog | WMUG | File CM12 Feedback | CM12 Docs | CM12 Release Notes

    Saturday, May 18, 2013 10:47 AM