Forum Discussion

EVStrategist's avatar
14 years ago
Solved

Preferred HA / DR solution - VCS or MCS

Just looking for Symantec's preferred solution they would recommend for HA and DR solutions -  Veritas Cluster Server or Microsoft Cluster Service.  Both seem to be offered as possible solutions and the whitepapers say both work but don't clearly state which one is better Symantec's mind.  I would be sure they support both if you had to make any kind of support call.

  • In terms of a preferred platform, we would certainly recommend the complete SFWHA solution.  This includes Volume Manager, DMP and Cluster Server for Windows.  In the documentation there will be reference to what we call our "MSCS Cluster Option" which is designed to extended support for Dynamic Disks and Volumes to MSCS/WFOC.  Natively Windows does not support this capability within their cluster framework without introducing Storage Foundation for Windows Enterprise Edition. 

    VCS supports a substantially larger number of applications for both HA and DR configurations.  MSCS/WFOC is very limited in how it extends cluster topologies such as Campus/Metro or Global/DR.

    If you are looking for justification for one versus the other,  I would say that you would want to consider the following:

     A. Flexibly - VCS offers an easily customizable framework for all manner of off-the-shelf or home-grown applications

    B. VCS is not tied to the OS itself and as such can be upgraded independently of the OS version.

    C. VCS supports instantaneous failure detection and "Cluster Volume Manager" for substantially reduced failover times from that of native MSCS

    D. VCS supports all major storage array replication tools (SRDF, TrueCopy, SnapMIrror, Global Mirror etc).

    E. SFWHA offers host based, hw independent replication with the Replicator Option (formerly VVR).  Any to Any storage replication.

    Hope this gives you the background you are looking for.

    Joe D 

6 Replies

  • Well I think Symantec would prefer VCS over Microsofts Cluster service since they own VCS.  wink

  • Obviously, we'd prefer you use VCS, but we support Storage Foundation for Windows in a Microsoft cluster environment, as well. :-)

  • In terms of a preferred platform, we would certainly recommend the complete SFWHA solution.  This includes Volume Manager, DMP and Cluster Server for Windows.  In the documentation there will be reference to what we call our "MSCS Cluster Option" which is designed to extended support for Dynamic Disks and Volumes to MSCS/WFOC.  Natively Windows does not support this capability within their cluster framework without introducing Storage Foundation for Windows Enterprise Edition. 

    VCS supports a substantially larger number of applications for both HA and DR configurations.  MSCS/WFOC is very limited in how it extends cluster topologies such as Campus/Metro or Global/DR.

    If you are looking for justification for one versus the other,  I would say that you would want to consider the following:

     A. Flexibly - VCS offers an easily customizable framework for all manner of off-the-shelf or home-grown applications

    B. VCS is not tied to the OS itself and as such can be upgraded independently of the OS version.

    C. VCS supports instantaneous failure detection and "Cluster Volume Manager" for substantially reduced failover times from that of native MSCS

    D. VCS supports all major storage array replication tools (SRDF, TrueCopy, SnapMIrror, Global Mirror etc).

    E. SFWHA offers host based, hw independent replication with the Replicator Option (formerly VVR).  Any to Any storage replication.

    Hope this gives you the background you are looking for.

    Joe D 

  • excellent info....this does firm up which solution is Symantec's preferred, as I suspected.

    Thanks,

    G.

  • what about scalability and performance, would VCS alone still winout over MCS or even Load Balance configurations. ?

  • Not really sure what you mean by "performance" with regards to an HA clustering solution as that's not a metric often thought of as relevant to HA. Generally speaking, one cluster product isn't going to failover an application any faster than another since HA clusters are basically at the mercy of the app and how fast it can shut down and start up. VCS with asynchronous monitoring enabled may edge out a competing product, though, since faults would be detected much sooner (as in nearly instantaneously) after they occur.

    As for scalability, VCS in a single cluster scales to 32 nodes, and with Veritas Operations Manager (VOM)  controlling multiple clusters, scaling to thousands of nodes is possible. Nothing else comes close.

    Why waltz when you can rock 'n roll? ;-)